diff --git a/.codecov.yml b/.codecov.yml
index 5fb5a76bb5..e2676824a2 100755
--- a/.codecov.yml
+++ b/.codecov.yml
@@ -1,42 +1,46 @@
#see https://github.com/codecov/support/wiki/Codecov-Yaml
codecov:
- notify:
- require_ci_to_pass: yes
+ require_ci_to_pass: yes
coverage:
- precision: 2 # 2 = xx.xx%, 0 = xx%
- round: nearest # how coverage is rounded: down/up/nearest
- range: 10...90 # custom range of coverage colors from red -> yellow -> green
+
+ # 2 = xx.xx%, 0 = xx%
+ precision: 2
+
+ # https://docs.codecov.com/docs/commit-status
status:
- # https://codecov.readme.io/v1.0/docs/commit-status
+
+ # We want our total main project to always remain above 84% coverage, a
+ # drop of 0.20% is allowed. It should fail if coverage couldn't be uploaded
+ # of the CI fails otherwise
project:
default:
- against: auto
- target: 70% # specify the target coverage for each commit status
- threshold: 50% # allow this little decrease on project
- # https://github.com/codecov/support/wiki/Filtering-Branches
- # branches: master
+ target: 84%
+ threshold: 0.20%
+ if_not_found: failure
if_ci_failed: error
- # https://github.com/codecov/support/wiki/Patch-Status
+
+ # The code changed by a PR should have 90% coverage. This is different from the
+ # overall number shown above.
+ # This encourages small PR's as they are easier to test.
patch:
default:
- against: auto
- target: 30% # specify the target "X%" coverage to hit
- threshold: 50% # allow this much decrease on patch
- changes: false
+ target: 90%
+ if_not_found: failure
+ if_ci_failed: error
+# We upload additional information on branching with pytest-cov `--cov-branch`
+# This information can be used by codecov.com to increase analysis of code
parsers:
gcov:
branch_detection:
conditional: true
loop: true
+ method: true
macro: false
- method: false
- javascript:
- enable_partials: false
+
comment:
- layout: header, diff
+ layout: diff, reach
+ behavior: default
require_changes: false
- behavior: default # update if exists else create new
- branches: *
\ No newline at end of file
diff --git a/.flake8 b/.flake8
index 704b67af15..d01cdb93b6 100644
--- a/.flake8
+++ b/.flake8
@@ -1,7 +1,13 @@
[flake8]
-max-line-length = 100
show-source = True
-application-import-names = autosklearn
-exclude =
+max-line-length = 88
+extend-exclude =
venv
+ .venv
build
+ autosklearn/automl_common
+extend-ignore =
+ # No whitespace before ':' in [x : y]
+ E203
+ # No lambdas — too strict
+ E731
diff --git a/.github/dependabot.yml b/.github/dependabot.yml
new file mode 100644
index 0000000000..e5e5092a20
--- /dev/null
+++ b/.github/dependabot.yml
@@ -0,0 +1,9 @@
+version: 2
+
+updates:
+ # This will check for updates to github actions every day
+ # https://docs.github.com/en/enterprise-server@3.4/code-security/dependabot/working-with-dependabot/keeping-your-actions-up-to-date-with-dependabot
+ - package-ecosystem: "github-actions"
+ directory: "/"
+ schedule:
+ interval: "daily"
diff --git a/.github/workflows/benchmarking-files/regressions-util.py b/.github/workflows/benchmarking-files/regressions-util.py
index c18635ca43..5128d0bbbe 100644
--- a/.github/workflows/benchmarking-files/regressions-util.py
+++ b/.github/workflows/benchmarking-files/regressions-util.py
@@ -7,34 +7,36 @@
import numpy as np
import pandas as pd
-CLASSIFICATION_METRICS = ['acc', 'auc', 'balacc', 'logloss']
-REGRESSION_METRICS = ['mae', 'r2', 'rmse']
+CLASSIFICATION_METRICS = ["acc", "auc", "balacc", "logloss"]
+REGRESSION_METRICS = ["mae", "r2", "rmse"]
METRICS = CLASSIFICATION_METRICS + REGRESSION_METRICS
+
def _get_mean_results_across_folds(df) -> pd.DataFrame:
- """ Returns a dataframe with the task, id, metric and the mean values
- across folds
+ """Returns a dataframe with the task, id, metric and the mean values
+ across folds
- [idx: 'task', 'id', 'metric', ... mean metrics across folds ...]
+ [idx: 'task', 'id', 'metric', ... mean metrics across folds ...]
"""
# Get the information about id and metric, only need info from first fold
# [idx: task, id, metric]
- df_info = df[df['fold'] == 0][['task', 'id', 'metric']].set_index('task')
+ df_info = df[df["fold"] == 0][["task", "id", "metric"]].set_index("task")
# [idx: task, ... mean metrics across folds ...]
available_metrics = list(set(METRICS).intersection(set(df.columns)))
- df_means = df[['task'] + available_metrics].groupby(['task']).mean()
+ df_means = df[["task"] + available_metrics].groupby(["task"]).mean()
return df_info.join(df_means)
+
def generate_framework_def(
user_dir: str,
username: str,
branch: str,
commit: str, # Not used in this setup but perhaps in a different one
):
- """ Creates a framework definition to run an autosklearn repo.
+ """Creates a framework definition to run an autosklearn repo.
Technically we only use the commit to pull the targeted version but for
naming consistency, we need to know the branch too.
@@ -61,40 +63,44 @@ def generate_framework_def(
or
#8b474a437ce980bd0909db59141b40d56f6d5688
"""
- assert len(commit) == 41 and commit[0] == '#' or len(commit) == 40, \
- "Not a commit hash"
+ assert (
+ len(commit) == 41 and commit[0] == "#" or len(commit) == 40
+ ), "Not a commit hash"
# automlbenchmark requires the '#' to identify it's a commit rather than
# a branch being targeted
- if commit[0] != '#':
- commit = '#' + commit
+ if commit[0] != "#":
+ commit = "#" + commit
# Tried commit and ssh repo but was getting errors with ssh
# Tried commit and https but getting issues with commit ref
# Using branch and https
version = branch
- repo = f'https://github.com/{username}/auto-sklearn.git'
+ repo = f"https://github.com/{username}/auto-sklearn.git"
# Create the framework file
- lines = '\n'.join([
- f"---",
- f"autosklearn_targeted:",
- f" extends: autosklearn",
- f" version: '{version}'",
- f" repo: '{repo}'"
- ])
-
- filepath = os.path.join(user_dir, 'frameworks.yaml')
- with open(filepath, 'w') as f:
+ lines = "\n".join(
+ [
+ f"---",
+ f"autosklearn_targeted:",
+ f" extends: autosklearn",
+ f" version: '{version}'",
+ f" repo: '{repo}'",
+ ]
+ )
+
+ filepath = os.path.join(user_dir, "frameworks.yaml")
+ with open(filepath, "w") as f:
f.writelines(lines)
+
def create_comparison(
baseline_csv_classification: str,
baseline_csv_regression: str,
targeted_csv_classification: str,
targeted_csv_regression: str,
) -> Tuple[pd.DataFrame, pd.DataFrame, pd.DataFrame]:
- """ Creates a csv with comparisons between the baseline and results.
+ """Creates a csv with comparisons between the baseline and results.
Scores are averaged across folds for a given task.
@@ -143,28 +149,33 @@ def create_comparison(
df_targeted_means = _get_mean_results_across_folds(df_targeted)
# Find the set intersection of tasks they have in common
- common_tasks = set(df_baseline_means.index).intersection(set(df_targeted_means.index))
+ common_tasks = set(df_baseline_means.index).intersection(
+ set(df_targeted_means.index)
+ )
# Find the set of metrics that are comparable
baseline_metrics = set(METRICS).intersection(set(df_baseline_means.columns))
common_metrics = baseline_metrics.intersection(set(df_targeted_means.columns))
# Calculate the differences for in common tasks, across all available metrics
- df_differences = df_targeted_means.loc[common_tasks][common_metrics] \
+ df_differences = (
+ df_targeted_means.loc[common_tasks][common_metrics]
- df_baseline_means.loc[common_tasks][common_metrics]
+ )
# Get the metric used for training and the dataset id of common tasks
- df_info = df_baseline_means.loc[common_tasks][['id', 'metric']]
+ df_info = df_baseline_means.loc[common_tasks][["id", "metric"]]
# Join together the info and the differences
return df_baseline_means, df_targeted_means, df_info.join(df_differences)
+
def create_comparisons_markdown(
baseline_means_csv: str,
targeted_means_csv: str,
compared_means_csv: str,
) -> str:
- """ Creates markdown that can be posted to Github that shows
+ """Creates markdown that can be posted to Github that shows
a comparison between the baseline and the targeted branch.
Parameters
@@ -186,36 +197,40 @@ def create_comparisons_markdown(
"""
# Create colours and func to create the markdown for it
colours = {
- 'Worse': ['353536', '800000', 'bd0000', 'ff0000'],
- 'Better': ['353536', '306300', '51a800', '6fe600'],
- 'Good': '6fe600',
- 'Bad': 'ff0000',
- 'Neutral': '353536',
- 'NaN': '52544f',
+ "Worse": ["353536", "800000", "bd0000", "ff0000"],
+ "Better": ["353536", "306300", "51a800", "6fe600"],
+ "Good": "6fe600",
+ "Bad": "ff0000",
+ "Neutral": "353536",
+ "NaN": "52544f",
}
+
def colour(kind, scale=None):
c = colours[kind] if scale is None else colours[kind][scale]
- return f''
+ return f""
# Metrics, whether positive is better and the tolerances between each
# Neutral, kind of good/bad, very good/bad etc...
metric_tolerances = {
- 'acc': { 'positive_is_better': True ,'tol': [0.001, 0.01, 0.2] },
- 'auc':{ 'positive_is_better': True ,'tol': [0.001, 0.01, 0.2] },
- 'balacc': { 'positive_is_better': True ,'tol': [0.001, 0.01, 0.2] },
- 'logloss': { 'positive_is_better': False ,'tol': [0.009, 0.01, 0.2] },
- 'mae': { 'positive_is_better': False ,'tol': [0.001, 0.01, 0.2] },
- 'r2': { 'positive_is_better': True ,'tol': [0.001, 0.01, 0.2] },
- 'rmse': { 'positive_is_better': False ,'tol': [0.001, 0.01, 0.2] },
+ "acc": {"positive_is_better": True, "tol": [0.001, 0.01, 0.2]},
+ "auc": {"positive_is_better": True, "tol": [0.001, 0.01, 0.2]},
+ "balacc": {"positive_is_better": True, "tol": [0.001, 0.01, 0.2]},
+ "logloss": {"positive_is_better": False, "tol": [0.009, 0.01, 0.2]},
+ "mae": {"positive_is_better": False, "tol": [0.001, 0.01, 0.2]},
+ "r2": {"positive_is_better": True, "tol": [0.001, 0.01, 0.2]},
+ "rmse": {"positive_is_better": False, "tol": [0.001, 0.01, 0.2]},
}
+
def is_good(score, metric):
return (
- score > 0 and metric_tolerances[metric]['positive_is_better']
- or score < 0 and not metric_tolerances[metric]['positive_is_better']
+ score > 0
+ and metric_tolerances[metric]["positive_is_better"]
+ or score < 0
+ and not metric_tolerances[metric]["positive_is_better"]
)
def is_neutral(diff, baseline_score, metric):
- tolerance = metric_tolerances[metric]['tol'][0]
+ tolerance = metric_tolerances[metric]["tol"][0]
if baseline_score == 0:
baseline_score = 1e-10
prc_diff = diff / baseline_score
@@ -223,18 +238,18 @@ def is_neutral(diff, baseline_score, metric):
def tolerance_colour(baseline_value, comparison_value, metric):
if np.isnan(baseline_value) or np.isnan(comparison_value):
- return colour('NaN')
+ return colour("NaN")
if baseline_value == 0:
baseline_value = 1e-10
prc_diff = comparison_value / baseline_value
- tolerances = metric_tolerances[metric]['tol']
- if metric_tolerances[metric]['positive_is_better']:
- diff_color = 'Better' if prc_diff > 0 else 'Worse'
+ tolerances = metric_tolerances[metric]["tol"]
+ if metric_tolerances[metric]["positive_is_better"]:
+ diff_color = "Better" if prc_diff > 0 else "Worse"
else:
- diff_color = 'Better' if prc_diff < 0 else 'Worse'
+ diff_color = "Better" if prc_diff < 0 else "Worse"
if abs(prc_diff) < tolerances[0]:
return colour(diff_color, 0)
@@ -245,23 +260,24 @@ def tolerance_colour(baseline_value, comparison_value, metric):
else:
return colour(diff_color, 3)
-
legend = {
- 'B': 'Baseline',
- 'T': 'Target Version',
- '**Bold**': 'Training Metric',
- '/': 'Missing Value',
- '---': 'Missing Task'
+ "B": "Baseline",
+ "T": "Target Version",
+ "**Bold**": "Training Metric",
+ "/": "Missing Value",
+ "---": "Missing Task",
}
- legend.update({
- key: colour(key)
- for key in set(colours.keys()) - set(['Worse', 'Better', 'Good', 'Bad'])
- })
+ legend.update(
+ {
+ key: colour(key)
+ for key in set(colours.keys()) - set(["Worse", "Better", "Good", "Bad"])
+ }
+ )
# Worse and better are handled seperatly
- compared = pd.read_csv(compared_means_csv, index_col='task')
- baseline = pd.read_csv(baseline_means_csv, index_col='task')
- targeted = pd.read_csv(targeted_means_csv, index_col='task')
+ compared = pd.read_csv(compared_means_csv, index_col="task")
+ baseline = pd.read_csv(baseline_means_csv, index_col="task")
+ targeted = pd.read_csv(targeted_means_csv, index_col="task")
# Some things to keep track of for the textual summary
n_performed_equally = 0
@@ -269,9 +285,9 @@ def tolerance_colour(baseline_value, comparison_value, metric):
n_performed_worse = 0
n_could_not_compare = 0
- headers = ['task', 'metric'] + METRICS
- table_header = '|'.join(headers)
- seperator = '|'.join(len(headers) * ['---'])
+ headers = ["task", "metric"] + METRICS
+ table_header = "|".join(headers)
+ seperator = "|".join(len(headers) * ["---"])
lines = [table_header, seperator]
@@ -279,13 +295,13 @@ def tolerance_colour(baseline_value, comparison_value, metric):
# The chosen metric name and the csv column differ with neg_logloss and
# logloss
- training_metric = baseline.loc[task]['metric']
+ training_metric = baseline.loc[task]["metric"]
if training_metric == "neg_logloss":
training_metric = "logloss"
# The baseline has tasks we can't compare with
if task not in compared.index:
- line = '|'.join([task, training_metric] + len(METRICS) * ['---'])
+ line = "|".join([task, training_metric] + len(METRICS) * ["---"])
lines.append(line)
# We can compare for a given tasks
@@ -299,34 +315,23 @@ def tolerance_colour(baseline_value, comparison_value, metric):
# If the metric does not exist for either, do fill it in as
# missing
- if (
- metric not in baseline.columns
- and metric not in compared.columns
- ):
+ if metric not in baseline.columns and metric not in compared.columns:
n_could_not_compare += 1
- entry = '/'
+ entry = "/"
# If the metric exists in the baseline but not in the comparison
- elif (
- metric in baseline.columns
- and not metric in compared.columns
- ):
+ elif metric in baseline.columns and not metric in compared.columns:
n_could_not_compare += 1
- entry = '
'.join([
- f' B : {baseline.loc[task][metric]:.3f}',
- f' T : /'
- ])
+ entry = "
".join(
+ [f" B : {baseline.loc[task][metric]:.3f}", f" T : /"]
+ )
# If the metric exists in the comparison but not in the baseline
- elif (
- metric in compared.columns
- and not metric in baseline.columns
- ):
+ elif metric in compared.columns and not metric in baseline.columns:
n_could_not_compare += 1
- entry = '
'.join([
- f' B : /',
- f' T : {targeted.loc[task][metric]:.3f}'
- ])
+ entry = "
".join(
+ [f" B : /", f" T : {targeted.loc[task][metric]:.3f}"]
+ )
# The metric must exist in both
else:
@@ -339,37 +344,43 @@ def tolerance_colour(baseline_value, comparison_value, metric):
else:
n_performed_worse += 1
- diff_colour = tolerance_colour(baseline_score,
- compared_score,
- metric)
- entry = '
'.join([
- f' B : {baseline.loc[task][metric]:.3f}',
- f' T : {targeted.loc[task][metric]:.3f}',
- f'{diff_colour}: {compared.loc[task][metric]:.3f}'
- ])
+ diff_colour = tolerance_colour(
+ baseline_score, compared_score, metric
+ )
+ entry = "
".join(
+ [
+ f" B : {baseline.loc[task][metric]:.3f}",
+ f" T : {targeted.loc[task][metric]:.3f}",
+ f"{diff_colour}: {compared.loc[task][metric]:.3f}",
+ ]
+ )
# Make the training metric entry bold
if metric == training_metric:
- entry = '' + entry + ''
+ entry = "" + entry + ""
entries.append(entry)
# Create the line
- line = '|'.join([task, training_metric] + entries)
+ line = "|".join([task, training_metric] + entries)
lines.append(line)
# Create the legend line
score_scale = {
- 'worse': "".join(colour('Worse', scale) for scale in range(len(colours['Worse']) - 1, 0, -1)),
- 'better': "".join(colour('Better', scale) for scale in range(len(colours['Better'])))
+ "worse": "".join(
+ colour("Worse", scale) for scale in range(len(colours["Worse"]) - 1, 0, -1)
+ ),
+ "better": "".join(
+ colour("Better", scale) for scale in range(len(colours["Better"]))
+ ),
}
score_scale = f'worse {score_scale["worse"] + score_scale["better"]} better'
- legend_str = '\t\t\t||\t\t'.join([score_scale] + [
- f'{key} - {text}' for key, text in legend.items()
- ])
+ legend_str = "\t\t\t||\t\t".join(
+ [score_scale] + [f"{key} - {text}" for key, text in legend.items()]
+ )
- lines.append('')
+ lines.append("")
lines.append(legend_str)
# Create a textual summary to go at the top
@@ -377,7 +388,6 @@ def tolerance_colour(baseline_value, comparison_value, metric):
compared_tasks = list(compared.index)
non_compared_tasks = list(set(baseline.index) - set(compared_tasks))
-
# Populate info about each metric
per_metric_info = {}
for metric in compared_metrics:
@@ -387,36 +397,37 @@ def tolerance_colour(baseline_value, comparison_value, metric):
item_colour = ""
if is_neutral(compared_average, baseline_average, metric):
- item_colour = colour('Neutral')
+ item_colour = colour("Neutral")
elif is_good(compared_average, metric):
- item_colour = colour('Good')
+ item_colour = colour("Good")
else:
- item_colour = colour('Bad')
+ item_colour = colour("Bad")
per_metric_info[metric] = {
- 'average': compared_average,
- 'n_compared': n_compared,
- 'colour': item_colour
+ "average": compared_average,
+ "n_compared": n_compared,
+ "colour": item_colour,
}
- summary = '\n'.join([
- f"# Results",
- f"Overall the targeted versions performance across {len(compared_tasks)} task(s) and {len(compared_metrics)} metric(s)",
- f"",
- f"* Equally on {n_performed_equally} comparisons",
- f"* Better on {n_performed_better} comparisons",
- f"* Worse on {n_performed_worse} comparisons",
- f"",
- f"There were {len(non_compared_tasks)} task(s) that could not be compared.",
- f"",
- f"The average change for each metric is:"
- f""
- ] + [
+ summary = "\n".join(
+ [
+ f"# Results",
+ f"Overall the targeted versions performance across {len(compared_tasks)} task(s) and {len(compared_metrics)} metric(s)",
+ f"",
+ f"* Equally on {n_performed_equally} comparisons",
+ f"* Better on {n_performed_better} comparisons",
+ f"* Worse on {n_performed_worse} comparisons",
+ f"",
+ f"There were {len(non_compared_tasks)} task(s) that could not be compared.",
+ f"",
+ f"The average change for each metric is:" f"",
+ ]
+ + [
f"* {metric}: {info['colour']} {info['average']:.4f} across {info['n_compared']} task(s)"
for metric, info in per_metric_info.items()
]
)
- return '\n'.join([summary] + [""] + lines)
+ return "\n".join([summary] + [""] + lines)
if __name__ == "__main__":
@@ -424,30 +435,30 @@ def tolerance_colour(baseline_value, comparison_value, metric):
# Generates a framework definition for automlbenchmark so that we can target
# auto-sklearn versions that are not our own
- parser.add_argument('--generate-framework-def', action='store_true')
- parser.add_argument('--user-dir', type=str)
- parser.add_argument('--owner', type=str)
- parser.add_argument('--branch', type=str)
- parser.add_argument('--commit', type=str)
+ parser.add_argument("--generate-framework-def", action="store_true")
+ parser.add_argument("--user-dir", type=str)
+ parser.add_argument("--owner", type=str)
+ parser.add_argument("--branch", type=str)
+ parser.add_argument("--commit", type=str)
# For comparing results generated by automlbenchmark for:
# -> baseline results generated [--baseline-csv]
# -> targeted results generated [--target-csv]
- # by automlbenchmark and the target branch 'results' generated
- parser.add_argument('--compare-results', action='store_true')
- parser.add_argument('--baseline-csv-classification', type=str)
- parser.add_argument('--baseline-csv-regression', type=str)
- parser.add_argument('--targeted-csv-classification', type=str)
- parser.add_argument('--targeted-csv-regression', type=str)
- parser.add_argument('--baseline-means-to', type=str)
- parser.add_argument('--targeted-means-to', type=str)
- parser.add_argument('--compared-means-to', type=str)
+ # by automlbenchmark and the target branch 'results' generated
+ parser.add_argument("--compare-results", action="store_true")
+ parser.add_argument("--baseline-csv-classification", type=str)
+ parser.add_argument("--baseline-csv-regression", type=str)
+ parser.add_argument("--targeted-csv-classification", type=str)
+ parser.add_argument("--targeted-csv-regression", type=str)
+ parser.add_argument("--baseline-means-to", type=str)
+ parser.add_argument("--targeted-means-to", type=str)
+ parser.add_argument("--compared-means-to", type=str)
# For generating markdown that can be posted to github that shows the results
- parser.add_argument('--generate-markdown', action='store_true')
- parser.add_argument('--compared-means-csv', type=str)
- parser.add_argument('--baseline-means-csv', type=str)
- parser.add_argument('--targeted-means-csv', type=str)
+ parser.add_argument("--generate-markdown", action="store_true")
+ parser.add_argument("--compared-means-csv", type=str)
+ parser.add_argument("--baseline-means-csv", type=str)
+ parser.add_argument("--targeted-means-csv", type=str)
args = parser.parse_args()
@@ -459,11 +470,17 @@ def tolerance_colour(baseline_value, comparison_value, metric):
elif args.compare_results:
- assert all([
- args.baseline_csv_classification, args.baseline_csv_regression,
- args.targeted_csv_classification, args.baseline_csv_regression,
- args.baseline_means_to, args.targeted_means_to, args.compared_means_to
- ])
+ assert all(
+ [
+ args.baseline_csv_classification,
+ args.baseline_csv_regression,
+ args.targeted_csv_classification,
+ args.baseline_csv_regression,
+ args.baseline_means_to,
+ args.targeted_means_to,
+ args.compared_means_to,
+ ]
+ )
baseline_means, targeted_means, compared_means = create_comparison(
baseline_csv_classification=args.baseline_csv_classification,
@@ -480,9 +497,9 @@ def tolerance_colour(baseline_value, comparison_value, metric):
df.to_csv(path)
elif args.generate_markdown:
- assert all([
- args.baseline_means_csv, args.targeted_means_csv, args.compared_means_csv
- ])
+ assert all(
+ [args.baseline_means_csv, args.targeted_means_csv, args.compared_means_csv]
+ )
comparisons_table = create_comparisons_markdown(
baseline_means_csv=args.baseline_means_csv,
diff --git a/.github/workflows/citation_cff.yml b/.github/workflows/citation_cff.yml
new file mode 100644
index 0000000000..6851c52d38
--- /dev/null
+++ b/.github/workflows/citation_cff.yml
@@ -0,0 +1,19 @@
+name: cffconvert
+
+on:
+ push:
+ paths:
+ - CITATION.cff
+
+jobs:
+ validate:
+ name: "validate"
+ runs-on: ubuntu-latest
+ steps:
+ - name: Check out a copy of the repository
+ uses: actions/checkout@v3
+
+ - name: Check whether the citation metadata from CITATION.cff is valid
+ uses: citation-file-format/cffconvert-github-action@2.0.0
+ with:
+ args: "--validate"
diff --git a/.github/workflows/dist.yml b/.github/workflows/dist.yml
index ada0593183..07ad9366a2 100644
--- a/.github/workflows/dist.yml
+++ b/.github/workflows/dist.yml
@@ -1,31 +1,57 @@
name: dist-check
-on: [push, pull_request]
+on:
+ # Manually triggerable in github
+ workflow_dispatch:
+
+ # When a push occurs on either of these branches
+ push:
+ branches:
+ - master
+ - development
+
+ # When a push occurs on a PR that targets these branches
+ pull_request:
+ branches:
+ - master
+ - development
jobs:
dist:
runs-on: ubuntu-latest
+
steps:
- - uses: actions/checkout@v2
+ - name: Check out the repo
+ uses: actions/checkout@v3
+ with:
+ submodules: recursive
+
- name: Setup Python
- uses: actions/setup-python@v2
+ uses: actions/setup-python@v4
with:
python-version: 3.8
+
- name: Build dist
run: |
python setup.py sdist
+
- name: Twine check
run: |
pip install twine
last_dist=$(ls -t dist/auto-sklearn-*.tar.gz | head -n 1)
- twine_output=`twine check "$last_dist"`
- if [[ "$twine_output" != "Checking $last_dist: PASSED" ]]; then echo $twine_output && exit 1;fi
+ twine check "$last_dist" --strict
+
- name: Install dist
run: |
last_dist=$(ls -t dist/auto-sklearn-*.tar.gz | head -n 1)
pip install $last_dist
+
- name: PEP 561 Compliance
run: |
pip install mypy
+
cd .. # required to use the installed version of autosklearn
- if ! python -c "import autosklearn"; then exit 1; fi
+
+ # Note this doesnt perform mypy checks, only
+ # that the types are exported
+ if ! mypy -c "import autosklearn"; then exit 1; fi
diff --git a/.github/workflows/docker-publish.yml b/.github/workflows/docker-publish.yml
index fe8c7f154e..4c4773700e 100644
--- a/.github/workflows/docker-publish.yml
+++ b/.github/workflows/docker-publish.yml
@@ -1,24 +1,32 @@
#https://help.github.com/en/actions/language-and-framework-guides/publishing-docker-images#publishing-images-to-github-packages
name: Publish Docker image
+
on:
+
push:
- # Push to `master` or `development`
branches:
- master
- development
- docker_workflow
jobs:
+
push_to_registry:
name: Push Docker image to GitHub Packages
runs-on: ubuntu-latest
+
steps:
- name: Check out the repo
- uses: actions/checkout@v2
+ uses: actions/checkout@v3
+ with:
+ submodules: recursive
+
- name: Extract branch name
shell: bash
run: echo "##[set-output name=branch;]$(echo ${GITHUB_REF#refs/heads/})"
id: extract_branch
+
+ # Updating this to @v2 requires a github auth token
- name: Push to GitHub Packages
uses: docker/build-push-action@v1
with:
@@ -28,6 +36,8 @@ jobs:
repository: automl/auto-sklearn/auto-sklearn
tag_with_ref: true
tags: ${{ steps.extract_branch.outputs.branch }}
+
+ # Updating this to @v2 requires a github auth token
- name: Push to Docker Hub
uses: docker/build-push-action@v1
with:
@@ -35,19 +45,24 @@ jobs:
password: ${{ secrets.DOCKER_PASSWORD }}
repository: mfeurer/auto-sklearn
tags: ${{ steps.extract_branch.outputs.branch }}
+
- name: Docker Login
run: docker login docker.pkg.github.com -u $GITHUB_ACTOR -p $GITHUB_TOKEN
env:
GITHUB_TOKEN: ${{secrets.GITHUB_TOKEN}}
+
- name: Pull Docker image
run: docker pull docker.pkg.github.com/$GITHUB_REPOSITORY/auto-sklearn:$BRANCH
env:
BRANCH: ${{ steps.extract_branch.outputs.branch }}
+
- name: Run image
run: docker run -i -d --name unittester -v $GITHUB_WORKSPACE:/workspace -w /workspace docker.pkg.github.com/$GITHUB_REPOSITORY/auto-sklearn:$BRANCH
env:
BRANCH: ${{ steps.extract_branch.outputs.branch }}
+
- name: Auto-Sklearn loaded
run: docker exec -i unittester python3 -c 'import autosklearn; print(f"Auto-sklearn imported from {autosklearn.__file__}")'
+
- name: Run unit testing
run: docker exec -i unittester python3 -m pytest -v test
diff --git a/.github/workflows/docs.yml b/.github/workflows/docs.yml
index 12c1f9f390..83510c5483 100644
--- a/.github/workflows/docs.yml
+++ b/.github/workflows/docs.yml
@@ -1,31 +1,57 @@
name: Docs
-on: [pull_request, push]
+
+on:
+ # Manually triggerable in github
+ workflow_dispatch:
+
+ # When a push occurs on either of these branches
+ push:
+ branches:
+ - master
+ - development
+
+ # When a push occurs on a PR that targets these branches
+ pull_request:
+ branches:
+ - master
+ - development
jobs:
+
build-and-deploy:
runs-on: ubuntu-latest
steps:
- - uses: actions/checkout@v2
+
+ - name: Checkout
+ uses: actions/checkout@v3
+ with:
+ submodules: recursive
+
- name: Setup Python
- uses: actions/setup-python@v2
+ uses: actions/setup-python@v4
with:
python-version: 3.8
+
- name: Install dependencies
run: |
- pip install -e .[docs,examples,examples_unix]
+ pip install -e .[docs,examples]
+
- name: Make docs
run: |
cd doc
make html
+
- name: Check links
run: |
cd doc
make linkcheck
+
- name: Pull latest gh-pages
if: (contains(github.ref, 'develop') || contains(github.ref, 'master')) && github.event_name == 'push'
run: |
cd ..
git clone https://github.com/automl/auto-sklearn.git --branch gh-pages --single-branch gh-pages
+
- name: Copy new doc into gh-pages
if: (contains(github.ref, 'develop') || contains(github.ref, 'master')) && github.event_name == 'push'
run: |
@@ -33,6 +59,7 @@ jobs:
cd ../gh-pages
rm -rf $branch_name
cp -r ../auto-sklearn/doc/build/html $branch_name
+
- name: Push to gh-pages
if: (contains(github.ref, 'develop') || contains(github.ref, 'master')) && github.event_name == 'push'
run: |
diff --git a/.github/workflows/generate-baselines.yml b/.github/workflows/generate-baselines.yml
index f066d4150c..5149dd57d8 100644
--- a/.github/workflows/generate-baselines.yml
+++ b/.github/workflows/generate-baselines.yml
@@ -1,3 +1,4 @@
+# Not currently in use
name: Generate Baselines
on:
@@ -8,10 +9,6 @@ on:
required: false
default: 'Dummy argument'
- push:
- branches:
- - development
-
env:
AUTOMLBENCHMARK_REPO: 'openml/automlbenchmark'
AUTOMLBENCHMARK_REF: 'master'
@@ -62,7 +59,7 @@ jobs:
# value: The python version used by the installed system
- name: Setup Python
- uses: actions/setup-python@v2
+ uses: actions/setup-python@v4
with:
python-version: ${{ steps.python-version.outputs.value }}
@@ -112,7 +109,7 @@ jobs:
# results_path: path to the benchmark results
- name: Upload Results as Artifact
- uses: actions/upload-artifact@v2
+ uses: actions/upload-artifact@v3
with:
name: baselines
path: |
diff --git a/.github/workflows/pre-commit.yaml b/.github/workflows/pre-commit.yaml
index eabada7e8d..c7e5b94438 100644
--- a/.github/workflows/pre-commit.yaml
+++ b/.github/workflows/pre-commit.yaml
@@ -1,20 +1,39 @@
name: pre-commit
-on: [push, pull_request]
+on:
+ # Manually triggerable in github
+ workflow_dispatch:
+
+ # When a push occurs on either of these branches
+ push:
+ branches:
+ - master
+ - development
+
+ # When a push occurs on a PR that targets these branches
+ pull_request:
+ branches:
+ - master
+ - development
jobs:
run-all-files:
runs-on: ubuntu-latest
steps:
- - uses: actions/checkout@v2
+ - uses: actions/checkout@v3
+ with:
+ submodules: recursive
+
- name: Setup Python 3.7
- uses: actions/setup-python@v2
+ uses: actions/setup-python@v4
with:
python-version: 3.7
+
- name: Install pre-commit
run: |
pip install pre-commit
pre-commit install
+
- name: Run pre-commit
run: |
pre-commit run --all-files
diff --git a/.github/workflows/pytest.yml b/.github/workflows/pytest.yml
index 513d8ff07f..794157f602 100644
--- a/.github/workflows/pytest.yml
+++ b/.github/workflows/pytest.yml
@@ -1,106 +1,144 @@
name: Tests
on:
+ # Manually triggerable in github
+ workflow_dispatch:
+
+ # When a push occurs on either of these branches
push:
+ branches:
+ - master
+ - development
+
+ # When a push occurs on a PR that targets these branches
pull_request:
+ branches:
+ - master
+ - development
+
schedule:
- # Every Monday at 7AM UTC
- - cron: '0 07 * * 1'
+ # Every day at 7AM UTC
+ - cron: '0 07 * * *'
+
+env:
+
+ # Arguments used for pytest
+ pytest-args: >-
+ --forked
+ --durations=20
+ --timeout=600
+ --timeout-method=thread
+ -s
+
+ # Arguments used for code-cov which is later used to annotate PR's on github
+ code-cov-args: >-
+ --cov=autosklearn
+ --cov-report=xml
jobs:
+
ubuntu:
- runs-on: ubuntu-20.04
+
+ name: ${{ matrix.os }}-${{ matrix.python-version }}-${{ matrix.kind }}
+ runs-on: ${{ matrix.os }}
strategy:
+ fail-fast: false
matrix:
- python-version: [3.7, 3.8, 3.9]
- use-conda: [true, false]
- use-dist: [false]
+ os: [windows-latest, macos-latest, ubuntu-latest]
+ python-version: ['3.7', '3.8', '3.9', '3.10']
+ kind: ['conda', 'source', 'dist']
+
+ exclude:
+ # Exclude all configurations *-*-dist, include one later
+ - kind: 'dist'
+
+ # Exclude windows as bash commands wont work in windows runner
+ - os: windows-latest
+
+ # Exclude macos as there are permission errors using conda as we do
+ - os: macos-latest
+
include:
- - python-version: 3.8
+ # Add the tag code-cov to ubuntu-3.7-source
+ - os: ubuntu-latest
+ python-version: 3.7
+ kind: 'source'
code-cov: true
- - python-version: 3.7
- use-conda: false
- use-dist: true
- fail-fast: false
+
+ # Include one config with dist, ubuntu-3.7-dist
+ - os: ubuntu-latest
+ python-version: 3.7
+ kind: 'dist'
steps:
- - uses: actions/checkout@v2
+ - name: Checkout
+ uses: actions/checkout@v3
+ with:
+ submodules: recursive
+
- name: Setup Python ${{ matrix.python-version }}
- uses: actions/setup-python@v2
- # A note on checkout: When checking out the repository that
- # triggered a workflow, this defaults to the reference or SHA for that event.
- # Otherwise, uses the default branch (master) is used.
+ uses: actions/setup-python@v4
with:
python-version: ${{ matrix.python-version }}
- - name: Conda Install test dependencies
- if: matrix.use-conda == true
+ - name: Conda install
+ if: matrix.kind == 'conda'
run: |
# Miniconda is available in $CONDA env var
$CONDA/bin/conda create -n testenv --yes pip wheel gxx_linux-64 gcc_linux-64 swig python=${{ matrix.python-version }}
$CONDA/envs/testenv/bin/python3 -m pip install --upgrade pip
$CONDA/envs/testenv/bin/pip3 install -e .[test]
- - name: Install test dependencies
- if: matrix.use-conda == false && matrix.use-dist == false
+ - name: Source install
+ if: matrix.kind == 'source'
run: |
python -m pip install --upgrade pip
- if [[ `python -c 'import platform; print(platform.python_version())' | cut -d '.' -f 2` -eq 6 ]]; then
- # Numpy 1.20 dropped suppert for Python3.6
- pip install "numpy<=1.19"
- fi
pip install -e .[test]
- sudo apt-get update
- sudo apt-get remove swig
- sudo apt-get install swig3.0
- sudo ln -s /usr/bin/swig3.0 /usr/bin/swig
- - name: Dist Install test dependencies
- if: matrix.use-conda == false && matrix.use-dist == true
+ - name: Dist install
+ if: matrix.kind == 'dist'
run: |
python -m pip install --upgrade pip
- sudo apt-get update
- sudo apt-get remove swig
- sudo apt-get install swig3.0
- sudo ln -s /usr/bin/swig3.0 /usr/bin/swig
- # We need to install for the dependencies, like pytest
python setup.py sdist
last_dist=$(ls -t dist/auto-sklearn-*.tar.gz | head -n 1)
pip install $last_dist[test]
- - name: Store repository status
+ - name: Store git status
id: status-before
run: |
echo "::set-output name=BEFORE::$(git status --porcelain -b)"
- - name: Conda Run tests
- timeout-minutes: 60
- if: matrix.use-conda == true
- run: |
- export OPENBLAS_NUM_THREADS=1
- export OMP_NUM_THREADS=1
- export MKL_NUM_THREADS=1
- # We activate conda as metalearning uses python directly, so we need
- # to change the default python
- export PATH="$CONDA/envs/testenv/bin:$PATH"
- if [ ${{ matrix.code-cov }} ]; then codecov='--cov=autosklearn --cov-report=xml'; fi
- $CONDA/envs/testenv/bin/python3 -m pytest --durations=20 --timeout=300 --timeout-method=thread -v $codecov test
-
- - name: Run tests
- timeout-minutes: 60
- if: matrix.use-conda == false
+ - name: Tests
+ timeout-minutes: 120
run: |
export OPENBLAS_NUM_THREADS=1
export OMP_NUM_THREADS=1
export MKL_NUM_THREADS=1
- if [ ${{ matrix.code-cov }} ]; then codecov='--cov=autosklearn --cov-report=xml'; fi
- pytest --durations=20 --timeout=300 --timeout-method=thread -v $codecov test
+
+ if [[ ${{ matrix.kind }} == 'conda' ]]; then
+ PYTHON=$CONDA/envs/testenv/bin/python3
+
+ # As one of the tests runs a subprocess command and calls `python3`, we must
+ # explicitly add it to the path
+ export PATH="$CONDA/envs/testenv/bin:$PATH"
+
+ else
+ PYTHON=$(which python3)
+ fi
+
+ if [ ${{ matrix.code-cov }} ]; then
+ $PYTHON -m pytest ${{ env.pytest-args }} ${{ env.code-cov-args }} test
+ else
+ $PYTHON -m pytest ${{ env.pytest-args }} test
+ fi
- name: Check for files left behind by test
if: ${{ always() }}
run: |
+ # Deleting `.pytest_chache` as it's used during testing and not deleted
+ rm -rf ".pytest_cache"
before="${{ steps.status-before.outputs.BEFORE }}"
after="$(git status --porcelain -b)"
if [[ "$before" != "$after" ]]; then
@@ -112,7 +150,7 @@ jobs:
- name: Upload coverage
if: matrix.code-cov && always()
- uses: codecov/codecov-action@v1
+ uses: codecov/codecov-action@v3
with:
fail_ci_if_error: true
verbose: true
diff --git a/.github/workflows/regressions.yml b/.github/workflows/regressions.yml
index 8042afd666..8bb0addcf4 100644
--- a/.github/workflows/regressions.yml
+++ b/.github/workflows/regressions.yml
@@ -1,3 +1,4 @@
+# Not currently in use
name: Regression Tests
on:
@@ -33,7 +34,7 @@ jobs:
)
steps:
- name: Create comment
- uses: peter-evans/create-or-update-comment@v1
+ uses: peter-evans/create-or-update-comment@v2
with:
issue-number: ${{ github.event.pull_request.number }}
body: |
@@ -81,7 +82,7 @@ jobs:
# branch: the branch name
- name: Checkout Automlbenchmark
- uses: actions/checkout@v2
+ uses: actions/checkout@v3
with:
repository: ${{ env.AUTOMLBENCHMARK_REPO }}
ref: ${{ env.AUTOMLBENCHMARK_REF }}
@@ -141,7 +142,7 @@ jobs:
# value: The python version used by the installed system
- name: Setup Python
- uses: actions/setup-python@v2
+ uses: actions/setup-python@v4
with:
python-version: ${{ steps.python-version.outputs.value }}
@@ -184,7 +185,7 @@ jobs:
# filename: name of the results file
- name: Upload ${{matrix.task_type}} results as artifact
- uses: actions/upload-artifact@v2
+ uses: actions/upload-artifact@v3
with:
name: ${{ steps.benchmark.outputs.filename }}
path: |
@@ -205,7 +206,7 @@ jobs:
# value: The python version used by the installed system
- name: Setup Python
- uses: actions/setup-python@v2
+ uses: actions/setup-python@v4
with:
python-version: ${{ steps.python-version.outputs.value }}
@@ -249,7 +250,7 @@ jobs:
# - baseline_regression_x_x_x.csv
- name: Download workflow artifacts
- uses: actions/download-artifact@v2
+ uses: actions/download-artifact@v3
with:
path: artifacts
@@ -306,7 +307,7 @@ jobs:
# compared_means: path to the results of regression test vs baseline
- name: Upload all results together as an artifact
- uses: actions/upload-artifact@v2
+ uses: actions/upload-artifact@v3
with:
name: ${{ github.repository_owner }}_${{ steps.extract.outputs.branch }}_${{ github.sha }}
path: |
@@ -326,7 +327,7 @@ jobs:
&& github.event.action == 'labeled'
&& github.event.label.name == 'regression-tests'
)
- uses: peter-evans/find-comment@v1
+ uses: peter-evans/find-comment@v2
id: comment_finder
with:
issue-number: ${{ github.event.pull_request.number }}
diff --git a/.github/workflows/stale.yaml b/.github/workflows/stale.yaml
index 45422d04eb..5d24ae0627 100644
--- a/.github/workflows/stale.yaml
+++ b/.github/workflows/stale.yaml
@@ -1,22 +1,26 @@
name: 'Close stale issues'
+
on:
schedule:
- - cron: '30 1 * * *'
+ - cron: '0 7 * * *'
jobs:
+
stale:
runs-on: ubuntu-latest
steps:
- - uses: actions/stale@v3
+ - uses: actions/stale@v5
with:
days-before-stale: 60
days-before-close: 7
+ stale-issue-label: 'stale'
+ only-issue-labels: 'Answered,Feedback-Required,invalid,wontfix'
+ exempt-all-milestones: true
+
stale-issue-message: >
This issue has been automatically marked as stale because it has not had
recent activity. It will be closed if no further activity occurs for the
next 7 days. Thank you for your contributions.
+
close-issue-message: >
This issue has been automatically closed due to inactivity.
- stale-issue-label: 'stale'
- only-issue-labels: 'Answered,Feedback-Required,invalid,wontfix'
- exempt-all-milestones: true
diff --git a/.gitignore b/.gitignore
index 9d2b72965a..03cbf93339 100755
--- a/.gitignore
+++ b/.gitignore
@@ -1,6 +1,6 @@
# Documentation
-docs/build/*
-docs/examples
+doc/build/*
+doc/examples
*.py[cod]
@@ -38,6 +38,9 @@ htmlcov
.coverage
.tox
+# Pytest cache
+.pytest_cache
+
# pycharm
.idea
diff --git a/.gitmodules b/.gitmodules
new file mode 100644
index 0000000000..28a5492b66
--- /dev/null
+++ b/.gitmodules
@@ -0,0 +1,3 @@
+[submodule "autosklearn/automl_common"]
+ path = autosklearn/automl_common
+ url = https://github.com/automl/automl_common
diff --git a/.pre-commit-config.yaml b/.pre-commit-config.yaml
index 91b5bbdf54..ab3472ddc6 100644
--- a/.pre-commit-config.yaml
+++ b/.pre-commit-config.yaml
@@ -1,57 +1,57 @@
+# If you see me, please update my `rev` field using the provided links
+# Click the repo and update to latest tags.
+# If things break on update, raise an issue
repos:
+
+ - repo: https://github.com/pycqa/isort
+ rev: 5.10.1
+ hooks:
+ - id: isort
+ name: isort imports autosklearn
+ files: autosklearn/.*
+
+ - id: isort
+ name: isort imports test
+ files: test/.*
+
+ - repo: https://github.com/psf/black
+ rev: 22.3.0
+ hooks:
+ - id: black
+ name: black formatter autosklearn
+ files: autosklearn/.*
+
+ - id: black
+ name: black formatter test
+ files: test/.*
+
+ - id: black
+ name: black formatter examples
+ files: examples/.*
+
+ # This is disabled as most modules fail this
+ - repo: https://github.com/pycqa/pydocstyle
+ rev: 6.1.1
+ hooks:
+ - id: pydocstyle
+ files: DISABLED # autosklearn/.*
+ always_run: false
+ additional_dependencies: ["toml"] # Needed to parse pyproject.toml
+
- repo: https://github.com/pre-commit/mirrors-mypy
- rev: v0.761
+ rev: v0.942
hooks:
- id: mypy
- args: [--show-error-codes]
- name: mypy auto-sklearn-ensembles
- files: autosklearn/ensembles
- - id: mypy
- args: [--show-error-codes]
- name: mypy auto-sklearn-metrics
- files: autosklearn/metrics
- - id: mypy
- args: [--show-error-codes]
- name: mypy auto-sklearn-data
- files: autosklearn/data
- - id: mypy
- args: [--show-error-codes]
- name: mypy auto-sklearn-util
- files: autosklearn/util
- - id: mypy
- args: [--show-error-codes]
- name: mypy auto-sklearn-evaluation
- files: autosklearn/evaluation
- - id: mypy
- args: [--show-error-codes]
- name: mypy auto-sklearn-datapreprocessing
- files: autosklearn/pipeline/components/data_preprocessing/
+ name: mypy auto-sklearn
+ files: autosklearn/.*
+
- repo: https://gitlab.com/pycqa/flake8
- rev: 3.8.3
+ rev: 4.0.1
hooks:
- id: flake8
- name: flake8 auto-sklearn
+ name: flake8 autosklearn
files: autosklearn/.*
+
- id: flake8
- name: flake8 file-order-data
- files: autosklearn/data
- additional_dependencies:
- - flake8-import-order
- - id: flake8
- name: flake8 file-order-ensemble
- files: autosklearn/ensembles
- additional_dependencies:
- - flake8-import-order
- - id: flake8
- name: flake8 file-order-metrics
- files: autosklearn/metrics
- additional_dependencies:
- - flake8-import-order
- - id: flake8
- name: flake8 file-order-util
- files: autosklearn/util
- additional_dependencies:
- - flake8-import-order
- - id: flake8
- name: flake8 autosklearn-test
+ name: flake8 test
files: test/.*
diff --git a/CITATION.cff b/CITATION.cff
index a4e1395c80..38777995f0 100644
--- a/CITATION.cff
+++ b/CITATION.cff
@@ -25,15 +25,15 @@ authors:
affiliation: "University of Freiburg, Germany"
- family-names: "Eggensperger"
- given-names: "Katharina"
+ given-names: "Katharina"
orcid: "https://orcid.org/0000-0002-0309-401X"
affiliation: "University of Freiburg, Germany"
- - family-name: "Bergman"
+ - family-names: "Bergman"
given-names: "Edward"
affiliation: "University of Freiburg, Germany"
- - family-name: "Rivera Valverde"
+ - family-names: "Rivera Valverde"
given-names: "Francisco"
affiliation: "University of Freiburg, Germany"
@@ -81,7 +81,7 @@ preferred-citation:
authors:
- family-names: "Feurer"
- given-names: "Matthias"
+ given-names: "Matthias"
orcid: "https://orcid.org/0000-0001-9611-8588"
affiliation: "University of Freiburg, Germany"
@@ -90,7 +90,7 @@ preferred-citation:
affiliation: "University of Freiburg, Germany"
- family-names: "Eggensperger"
- given-names: "Katharina"
+ given-names: "Katharina"
orcid: "https://orcid.org/0000-0002-0309-401X"
affiliation: "University of Freiburg, Germany"
diff --git a/CONTRIBUTING.md b/CONTRIBUTING.md
index 38f0280a32..73ce781618 100644
--- a/CONTRIBUTING.md
+++ b/CONTRIBUTING.md
@@ -30,10 +30,12 @@ Following that we'll tell you about how you can test your changes locally and th
It's important to work off the latest changes on the **development** branch.
```bash
# With https
- git clone https://github.com/your-username/auto-sklearn
+ # Note the --recurse-submodules args, we use a submodule autosklearn/automl_common
+ # so it needs to be downloaded too
+ git clone --recurse-submodules https://github.com/your-username/auto-sklearn
# ... or with ssh
- git clone git@github.com:your-username/auto-sklearn.git
+ git clone --recurse-submodules git@github.com:your-username/auto-sklearn.git
# Navigate into the cloned repo
cd auto-sklearn
@@ -41,6 +43,11 @@ Following that we'll tell you about how you can test your changes locally and th
# Create a new branch based off the development one
git checkout -b my_new_branch development
+ # If you missed the --recurse-submodules arg during clone or need to install the
+ # submodule manually, then execute the following line:
+ #
+ # git submodule udate --init --recursive
+
# ... Alternatively, if you would prefer a more manual method
# Show all the available branches with a * beside your current one
git branch
@@ -50,6 +57,11 @@ Following that we'll tell you about how you can test your changes locally and th
# Create a new branch based on the currently active branch
git checkout -b my_new_branch
+
+ # If you missed the --recurse-submodules arg during clone or need to install the
+ # submodule manually, then execute the following line:
+ #
+ # git submodule update --init --recursive
```
The reason to create a new branch is two fold:
@@ -75,13 +87,17 @@ Following that we'll tell you about how you can test your changes locally and th
Fortunately, you can check out [pyenv](https://github.com/pyenv/pyenv) which lets you switch between Python versions on the fly!
* Now that we have a virtual environment, it's time to install all the dependencies into it.
+ We've provided a simple `make` command to help do this.
```bash
+ make install-dev
+
+ # Manually
pip install -e .[test,examples,doc]
# If you're using shells other than bash you'll need to use
pip install -e ".[test,examples,doc]"
```
- * If you're only exposure to using pip is `pip install package_name` then this might be a bit confusing.
+ * If your only exposure to using pip is `pip install package_name` then this might be a bit confusing.
* If we type `pip install -e .` (notice the 'dot'), this tells `pip` to install a package located here, in this directory, `.`.
The `-e` flag indicates that it should be editable, meaning you will not have to run `pip install .` every time you make a change and want to try it.
* Finally the `[test,examples,doc]` tells `pip` that there's some extra optional dependencies that we want to install.
@@ -89,6 +105,8 @@ Following that we'll tell you about how you can test your changes locally and th
You can check out what these are in the `setup.py` file.
* If you're new to virtual environments, after performing all this, it's a great time to check out what actually exists in the `my-virtual-env` folder.
+* You can check out some functionality we have captured in a `Makefile` by running `make help`
+
* Now it's time to make some changes, whether it be for [documentation](#documentation), a [bug fix](#bug-fixes) or a new [features](#features).
## Making Changes
@@ -189,16 +207,16 @@ Sometimes, the new functionality isn't so clear from a simple parameter descript
Lastly, if the feature really is a game changer or you're very proud of it, consider making an `example_*.py` that will be run and rendered in the online docs!
## Testing
-* Let's assume you've made some changes, now we have to make sure they work.
+* Let's assume you've made some changes, now we have to make sure they work.
Begin by simply running all the tests.
If there's any errors, they'll pop up once it's complete.
```bash
pytest
```
- * Note that these may take a while so check out `pytest --help` to see how you can run tests so that only previous failures run or only certain tests are run.
+ * Note that these may take a while so check out `pytest --help` to see how you can run tests so that only previous failures run or only certain tests are run.
This can help you try changes and get results faster.
Do however run one last full `pytest` once you are finished and happy!
- * Here are some we find particularly useful
+ * Here are some we find particularly useful
```
# Run tests in specific file like 'test_estimators.py'
pytest "test/test_automl/test_estimators.py"
@@ -218,27 +236,31 @@ Lastly, if the feature really is a game changer or you're very proud of it, cons
# Exit on the first test failure
pytest -x
```
- * More advanced editors like PyCharm may have built in integrations which could be good to check out!
+ * More advanced editors like PyCharm may have built in integrations which could be good to check out!
+ * Running all unittests will take a while, here's how you can run them in parallel
+ ```
+ export OPENBLAS_NUM_THREADS=1
+ export MKL_NUM_THREADS=1
+ export OMP_NUM_THREADS=1
+
+ pytest -n 4
+ ```
-* Now we are going to use [sphinx](https://www.sphinx-doc.org/en/master/) to generate all the documentation and make sure there are no issues.
+
+* Now we are going to use [sphinx](https://www.sphinx-doc.org/en/master/) to generate all the documentation and make sure there are no issues.
```bash
- cd doc
- make html
+ make doc
```
* If you're unfamiliar with sphinx, it's a documentation generator which can read comments and docstrings from within the code and generate html documentation.
- * We also use sphinx-gallery which can take python files (such as those in the `examples` folder) and run them, creating html which shows the code and the output it generates.
- Unfortunately this can take quite some time but you should only have to run this once.
- * If you need to quickly check something, you can run `make html-noexamples` to prevent sphinx from running the examples.
- ```bash
- cd doc
- make html-noexamples
- ```
- * Sphinx also has a command `linkcheck` for making sure all the links correctly go to some destination.
+ * If you've added documentation, we also has a command `linkcheck` for making sure all the links correctly go to some destination.
This helps tests for dead links or accidental typos.
```bash
- cd doc
make linkcheck
```
+ * We also use sphinx-gallery which can take python files (such as those in the `examples` folder) and run them, creating html which shows the code and the output it generates.
+ ```bash
+ make examples
+ ```
* To view the documentation itself, make sure it is built with the above commands and then open `doc/build/html/index.html` with your favourite browser:
```bash
# Firefox
@@ -248,20 +270,34 @@ Lastly, if the feature really is a game changer or you're very proud of it, cons
xdg-open ./doc/build/html/index.html
```
-* Once you've made all your changes and all the tests pass successfully, we need to make sure that the code fits a certain format and that the [typing](https://docs.python.org/3/library/typing.html) is correct.
- To do this, we use a tool call `pre-commit` which runs `flake8`, a code checker and `mypy`, a static type checker against the code.
+* Once you've made all your changes and all the tests pass successfully, we need to make sure that the code fits a certain format and that the [typing](https://docs.python.org/3/library/typing.html) is correct.
+ * Formatting and import sorting can helps keep things uniform across all coding styles. We use [`black`](https://black.readthedocs.io/en/stable/) and [`isort`](https://isort.readthedocs.io/en/latest/) to do this for us. To automatically run these formatters across the code base, just run the following command:
+ ```bash
+ make format
+ ```
+ * To then check for issues using [`black`](https://black.readthedocs.io/en/stable/), [`isort`](https://isort.readthedocs.io/en/latest/), [`mypy`](http://mypy-lang.org/), [`flake8`](https://flake8.pycqa.org/en/latest/) and [`pydocstyle`](http://www.pydocstyle.org/en/stable/), run
+ ```bash
+ make check
+ ```
+ * To do this checking automatically, we use `pre-commit` which if you already installed everything with `make install-dev` then this has been done for you.
+ This will happen every time you make a commit and warn you of any issues.
+ Otherwise you can run the following to install pre-commit.
+ ```bash
+ pre-commit install
+ ```
+ * To run `pre-commit` manually:
```bash
- pip install pre-commit
pre-commit run --all-files
```
- * The reason we use a code standard like `flake8` is to make sure that when we review code:
- * There are no extra blank spaces and blank lines.
- * Lines don't end up too long
- * Code from multiple source keeps a similar appearance.
- * We perform static type checking with `mypy` as this can remove a majority of bugs, before a test is even run.
- It points out programmer errors and is what makes compiled languages so safe.
+ * The reason we use tools like [`flake8`](https://flake8.pycqa.org/en/latest/), [`mypy`](http://mypy-lang.org/), [`black`](https://black.readthedocs.io/en/stable/), [`isort`](https://isort.readthedocs.io/en/latest/) and [`pydocstyle`](http://www.pydocstyle.org/en/stable/) is to make sure that when we review code:
+ * There are no extra blank spaces and blank lines. (`flake8`, `black`)
+ * Lines don't end up too long. (`flake8`, `black`)
+ * Code from multiple source keeps a similar appearance. (`black`)
+ * Importing things is consistently ordered. (`isort`)
+ * Functions are type annotated and correct with static type checking. (`mypy`)
+ * Function and classes have docstrings. (`pydocstyle`)
If you are new to Python types, or stuck with how something should be 'typed', please feel free to push the pull request in the following steps and we should be able to help you out.
- * If interested, the configuration for `pre-commit` can be found in `.pre-commit-config.yaml`
+ * If interested, the configuration for `pre-commit` can be found in `.pre-commit-config.yaml` with the other tools mainly being configured in `pyproject.toml` and `.flake8`.
## Creating the PR
@@ -329,26 +365,39 @@ Lastly, if the feature really is a game changer or you're very proud of it, cons
# Pull Request Overview
* Create a [fork](https://docs.github.com/en/get-started/quickstart/fork-a-repo) of the [automl/auto-sklearn](https://github.com/automl/auto-sklearn) git repo
+* Check out what's available by running `make help`.
* Clone your own fork and create a new branch from the branch to work on
```bash
git clone git@github.com:your-username/auto-sklearn.git
cd auto-sklearn
git checkout -b my_new_branch development
+ # Initialize autosklearn/automl_common submodule
+ git submodule update --init --recursive
+
# Create a virtual environment and activate it so there are no package
# conflicts
python -m venv my-virtual-env
source my-virtual-env/bin/activate
- pip install -e .[test,docs,examples] # zsh users need quotes ".[test,...]"
+ make install-dev
+ # pip install -e ".[test,docs,examples]" # To manually install things
# Edit files...
+ # Format code
+ make format
+
+ # Check for any issues
+ make check
+
+ # ... fix any issues
+
# If you changed documentation:
- # This will generate all documentation, run examples and check links
- cd doc
- make html
+ # This will generate all documentation and check links
+ make doc
make linkcheck
+ make examples # mainly needed if you modified some examples
# ... fix any issues
@@ -358,8 +407,8 @@ Lastly, if the feature really is a game changer or you're very proud of it, cons
# ... fix any issues
- # Use pre-commit for style and typing checks
- pip install pre-commit
+ # If you want to run pre-commit, the formatting checks we run on github
+ pre-commit install
pre-commit run --all-files
# ... fix any issues
diff --git a/Dockerfile b/Dockerfile
index d9f73b2c83..cb71f2f6ff 100644
--- a/Dockerfile
+++ b/Dockerfile
@@ -32,9 +32,6 @@ ADD . /auto-sklearn/
# Upgrade pip then install dependencies
RUN pip3 install --upgrade pip
-RUN pip3 install pytest==4.6.* pep8 codecov pytest-cov flake8 flaky openml
-RUN cat /auto-sklearn/requirements.txt | xargs -n 1 -L 1 pip3 install
-RUN pip3 install jupyter
# Install
-RUN pip3 install /auto-sklearn/
+RUN pip3 install "/auto-sklearn[test, examples]"
diff --git a/MANIFEST.in b/MANIFEST.in
index e76cdbb0ea..dffd0c7283 100644
--- a/MANIFEST.in
+++ b/MANIFEST.in
@@ -1,9 +1,18 @@
-recursive-include autosklearn/metalearning/files *.arff
-recursive-include autosklearn/metalearning/files *.csv
-recursive-include autosklearn/metalearning/files *.txt
-include autosklearn/util/logging.yaml
+include LICENSE.txt
include requirements.txt
+include autosklearn/util/logging.yaml
include autosklearn/requirements.txt
-recursive-include autosklearn/experimental/ *.json
-include autosklearn/experimental/askl2_training_data.json
-include LICENSE.txt
+include autosklearn/py.typed
+
+# Meta-data
+recursive-include autosklearn/metalearning/files *.arff *.csv *.txt
+recursive-include autosklearn/experimental *.json
+
+# Remove tests from automl_common
+prune autosklearn/automl_common/test
+exclude autosklearn/automl_common/setup.py
+
+# Include automl_common LICENSE and README
+include autosklearn/automl_common/LICENSE
+include autosklearn/automl_common/README.md
+
diff --git a/Makefile b/Makefile
index b00a2392ec..2841ea9da0 100644
--- a/Makefile
+++ b/Makefile
@@ -1,32 +1,130 @@
-# simple makefile to simplify repetitive build env management tasks under posix
+# NOTE: Used on linux, limited support outside of Linux
+#
+# A simple makefile to help with small tasks related to development of autosklearn
+# These have been configured to only really run short tasks. Longer form tasks
+# are usually completed in github actions.
+
+.PHONY: help install-dev check format pre-commit clean clean-doc clean-build build doc links examples publish test
+
+help:
+ @echo "Makefile autosklearn"
+ @echo "* install-dev to install all dev requirements and install pre-commit"
+ @echo "* check to check the source code for issues"
+ @echo "* format to format the code with black and isort"
+ @echo "* pre-commit to run the pre-commit check"
+ @echo "* clean to clean the dist and doc build files"
+ @echo "* build to build a dist"
+ @echo "* doc to generate and view the html files"
+ @echo "* linkcheck to check the documentation links"
+ @echo "* examples to run and generate the examples"
+ @echo "* publish to help publish the current branch to pypi"
+ @echo "* test to run the tests"
PYTHON ?= python
CYTHON ?= cython
PYTEST ?= python -m pytest
CTAGS ?= ctags
+PIP ?= python -m pip
+MAKE ?= make
+BLACK ?= black
+ISORT ?= isort
+PYDOCSTYLE ?= pydocstyle
+MYPY ?= mypy
+PRECOMMIT ?= pre-commit
+FLAKE8 ?= flake8
+
+DIR := "${CURDIR}"
+DIST := "${CURDIR}/dist""
+DOCDIR := "${DIR}/doc"
+INDEX_HTML := "file://${DOCDIR}/build/html/index.html"
+
+install-dev:
+ $(PIP) install -e ".[test,examples,docs]"
+ pre-commit install
+
+check-black:
+ $(BLACK) autosklearn examples test --check || :
+
+check-isort:
+ $(ISORT) autosklearn test --check || :
+
+check-pydocstyle:
+ $(PYDOCSTYLE) autosklearn || :
+
+check-mypy:
+ $(MYPY) autosklearn || :
+
+check-flake8:
+ $(FLAKE8) autosklearn || :
+ $(FLAKE8) test || :
-all: clean inplace test
+# pydocstyle does not have easy ignore rules, instead, we include as they are covered
+check: check-black check-isort check-mypy check-flake8 # check-pydocstyle
-clean:
+pre-commit:
+ $(PRECOMMIT) run --all-files
+
+format-black:
+ $(BLACK) autosklearn/.*
+ $(BLACK) test/.*
+ $(BLACK) examples/.*
+
+format-isort:
+ $(ISORT) autosklearn
+ $(ISORT) test
+
+
+format: format-black format-isort
+
+clean-doc:
+ $(MAKE) -C ${DOCDIR} clean
+
+clean-build:
$(PYTHON) setup.py clean
- rm -rf dist
+ rm -rf ${DIST}
-in: inplace # just a shortcut
-inplace:
- $(PYTHON) setup.py build_ext -i
+# Clean up any builds in ./dist as well as doc
+clean: clean-doc clean-build
+
+# Build a distribution in ./dist
+build:
+ $(PYTHON) setup.py sdist
doc:
- cd ./doc
- make html
- cd ..
+ $(MAKE) -C ${DOCDIR} html-noexamples
+ @echo
+ @echo "View docs at:"
+ @echo ${INDEX_HTML}
+
+links:
+ $(MAKE) -C ${DOCDIR} linkcheck
-test-code: in
- $(PYTEST) -s -v test
-test-doc:
- $(PYTEST) -s -v doc/*.rst
+examples:
+ $(MAKE) -C ${DOCDIR} html
+ @echo
+ @echo "View docs at:"
+ @echo ${INDEX_HTML}
-test-coverage:
- rm -rf coverage .coverage
- $(PYTEST) -s -v --with-coverage test
+# Publish to testpypi
+# Will echo the commands to actually publish to be run to publish to actual PyPi
+# This is done to prevent accidental publishing but provide the same conveniences
+publish: clean build
+ $(PIP) install twine
+ $(PYTHON) -m twine upload --repository testpypi ${DIST}/*
+ @echo
+ @echo "Test with the following:"
+ @echo "* Create a new virtual environment to install the uplaoded distribution into"
+ @echo "* Run the following:"
+ @echo
+ @echo " pip install --index-url https://test.pypi.org/simple/ --extra-index-url https://pypi.org/simple/ autosklearn"
+ @echo
+ @echo "* Run this to make sure it can import correctly, plus whatever else you'd like to test:"
+ @echo
+ @echo " python -c 'import autosklearn'"
+ @echo
+ @echo "Once you have decided it works, publish to actual pypi with"
+ @echo
+ @echo " python -m twine upload dist/*"
-test: test-code test-sphinxext test-doc
+test:
+ $(PYTEST) test
diff --git a/README.md b/README.md
index 2a7908a4af..b506960d70 100644
--- a/README.md
+++ b/README.md
@@ -26,15 +26,15 @@ predictions = cls.predict(X_test)
If you use auto-sklearn in scientific publications, we would appreciate citations.
-**Efficient and Robust Automated Machine Learning**
-*Matthias Feurer, Aaron Klein, Katharina Eggensperger, Jost Springenberg, Manuel Blum and Frank Hutter*
-Advances in Neural Information Processing Systems 28 (2015)
+**Efficient and Robust Automated Machine Learning**
+*Matthias Feurer, Aaron Klein, Katharina Eggensperger, Jost Springenberg, Manuel Blum and Frank Hutter*
+Advances in Neural Information Processing Systems 28 (2015)
[Link](https://papers.neurips.cc/paper/5872-efficient-and-robust-automated-machine-learning.pdf) to publication.
```
@inproceedings{feurer-neurips15a,
title = {Efficient and Robust Automated Machine Learning},
- author = {Feurer, Matthias and Klein, Aaron and Eggensperger, Katharina Springenberg, Jost and Blum, Manuel and Hutter, Frank},
+ author = {Feurer, Matthias and Klein, Aaron and Eggensperger, Katharina and Springenberg, Jost and Blum, Manuel and Hutter, Frank},
booktitle = {Advances in Neural Information Processing Systems 28 (2015)},
pages = {2962--2970},
year = {2015}
@@ -43,8 +43,8 @@ Advances in Neural Information Processing Systems 28 (2015)
----------------------------------------
-**Auto-Sklearn 2.0: The Next Generation**
-*Matthias Feurer, Katharina Eggensperger, Stefan Falkner, Marius Lindauer and Frank Hutter**
+**Auto-Sklearn 2.0: The Next Generation**
+*Matthias Feurer, Katharina Eggensperger, Stefan Falkner, Marius Lindauer and Frank Hutter**
arXiv:2007.04074 [cs.LG], 2020
[Link](https://arxiv.org/abs/2007.04074) to publication.
diff --git a/autosklearn/__init__.py b/autosklearn/__init__.py
index dae47a1089..2bff637af8 100644
--- a/autosklearn/__init__.py
+++ b/autosklearn/__init__.py
@@ -1,27 +1,26 @@
# -*- encoding: utf-8 -*-
import os
-import pkg_resources
import sys
-from autosklearn.util import dependencies
-from autosklearn.__version__ import __version__ # noqa (imported but unused)
+import pkg_resources
+from autosklearn.__version__ import __version__ # noqa (imported but unused)
+from autosklearn.util import dependencies
-requirements = pkg_resources.resource_string('autosklearn', 'requirements.txt')
-requirements = requirements.decode('utf-8')
+requirements = pkg_resources.resource_string("autosklearn", "requirements.txt")
+requirements = requirements.decode("utf-8")
dependencies.verify_packages(requirements)
-if os.name != 'posix':
+if os.name != "posix":
raise ValueError(
- 'Detected unsupported operating system: %s. Please check '
- 'the compability information of auto-sklearn: https://automl.github.io'
- '/auto-sklearn/stable/installation.html#windows-osx-compability' %
- sys.platform
+ "Detected unsupported operating system: %s. Please check "
+ "the compability information of auto-sklearn: https://automl.github.io"
+ "/auto-sklearn/stable/installation.html#windows-osx-compability" % sys.platform
)
if sys.version_info < (3, 6):
raise ValueError(
- 'Unsupported python version %s found. Auto-sklearn requires Python '
- '3.6 or higher.' % sys.version_info
+ "Unsupported python version %s found. Auto-sklearn requires Python "
+ "3.6 or higher." % sys.version_info
)
diff --git a/autosklearn/__version__.py b/autosklearn/__version__.py
index d33bd90441..6928aee41e 100644
--- a/autosklearn/__version__.py
+++ b/autosklearn/__version__.py
@@ -1,4 +1,4 @@
"""Version information."""
# The following line *must* be the last in the module, exactly as formatted:
-__version__ = "0.14.3"
+__version__ = "0.14.6"
diff --git a/autosklearn/automl.py b/autosklearn/automl.py
index 064a887a4a..278cd5c146 100644
--- a/autosklearn/automl.py
+++ b/autosklearn/automl.py
@@ -1,81 +1,130 @@
-# -*- encoding: utf-8 -*-
+from __future__ import annotations
+
+from typing import (
+ Any,
+ Callable,
+ Dict,
+ Iterable,
+ Mapping,
+ Optional,
+ Sequence,
+ Tuple,
+ Type,
+)
+
import copy
import io
+import itertools
import json
-import platform
import logging.handlers
import multiprocessing
import os
+import platform
import sys
+import tempfile
import time
-from typing import Any, Dict, Optional, List, Tuple, Union
+import types
import uuid
-import unittest.mock
-import tempfile
+import warnings
-from ConfigSpace.configuration_space import Configuration
-from ConfigSpace.read_and_write import json as cs_json
-import dask
-import dask.distributed
+import distro
+import joblib
import numpy as np
import numpy.ma as ma
import pandas as pd
import pkg_resources
import scipy.stats
-from sklearn.base import BaseEstimator
-from sklearn.ensemble import VotingRegressor
-from sklearn.model_selection._split import _RepeatedSplits, \
- BaseShuffleSplit, BaseCrossValidator
-from smac.runhistory.runhistory import RunInfo, RunValue
-from smac.tae import StatusType
-from smac.stats.stats import Stats
-import joblib
import sklearn.utils
+from ConfigSpace.configuration_space import Configuration, ConfigurationSpace
+from ConfigSpace.read_and_write import json as cs_json
+from dask.distributed import Client, LocalCluster
from scipy.sparse import spmatrix
+from sklearn.base import BaseEstimator
+from sklearn.dummy import DummyClassifier, DummyRegressor
+from sklearn.ensemble import VotingClassifier, VotingRegressor
+from sklearn.metrics._classification import type_of_target
+from sklearn.model_selection._split import (
+ BaseCrossValidator,
+ BaseShuffleSplit,
+ _RepeatedSplits,
+)
+from sklearn.pipeline import Pipeline
from sklearn.utils import check_random_state
from sklearn.utils.validation import check_is_fitted
-from sklearn.metrics._classification import type_of_target
-from sklearn.dummy import DummyClassifier, DummyRegressor
-
-from autosklearn.metrics import Scorer, default_metric_for_task
-from autosklearn.data.xy_data_manager import XYDataManager
+from smac.callbacks import IncorporateRunResultCallback
+from smac.runhistory.runhistory import RunInfo, RunValue
+from smac.stats.stats import Stats
+from smac.tae import StatusType
+from typing_extensions import Literal
+
+from autosklearn.automl_common.common.utils.backend import Backend, create
+from autosklearn.constants import (
+ BINARY_CLASSIFICATION,
+ CLASSIFICATION_TASKS,
+ MULTICLASS_CLASSIFICATION,
+ MULTILABEL_CLASSIFICATION,
+ MULTIOUTPUT_REGRESSION,
+ REGRESSION,
+ REGRESSION_TASKS,
+)
from autosklearn.data.validation import (
- convert_if_sparse,
- InputValidator,
SUPPORTED_FEAT_TYPES,
SUPPORTED_TARGET_TYPES,
+ InputValidator,
+ convert_if_sparse,
)
+from autosklearn.data.xy_data_manager import XYDataManager
+from autosklearn.ensemble_building import EnsembleBuilderManager
+from autosklearn.ensembles.abstract_ensemble import (
+ AbstractEnsemble,
+ AbstractMultiObjectiveEnsemble,
+)
+from autosklearn.ensembles.ensemble_selection import EnsembleSelection
+from autosklearn.ensembles.singlebest_ensemble import SingleBest
from autosklearn.evaluation import ExecuteTaFuncWithQueue, get_cost_of_crash
from autosklearn.evaluation.abstract_evaluator import _fit_and_suppress_warnings
from autosklearn.evaluation.train_evaluator import TrainEvaluator, _fit_with_budget
-from autosklearn.metrics import calculate_metric
-from autosklearn.util.backend import Backend, create
-from autosklearn.util.stopwatch import StopWatch
-from autosklearn.util.logging_ import (
- setup_logger,
- start_log_server,
- get_named_client_logger,
- warnings_to,
- PicklableClientLogger,
+from autosklearn.metrics import (
+ Scorer,
+ _validate_metrics,
+ compute_single_metric,
+ default_metric_for_task,
)
-from autosklearn.util import pipeline, RE_PATTERN
-from autosklearn.util.parallel import preload_modules
-from autosklearn.ensemble_builder import EnsembleBuilderManager
-from autosklearn.ensembles.singlebest_ensemble import SingleBest
-from autosklearn.smbo import AutoMLSMBO
-from autosklearn.constants import MULTILABEL_CLASSIFICATION, MULTICLASS_CLASSIFICATION, \
- REGRESSION_TASKS, REGRESSION, BINARY_CLASSIFICATION, MULTIOUTPUT_REGRESSION, \
- CLASSIFICATION_TASKS
from autosklearn.pipeline.base import BasePipeline
from autosklearn.pipeline.components.classification import ClassifierChoice
-from autosklearn.pipeline.components.regression import RegressorChoice
-from autosklearn.pipeline.components.feature_preprocessing import FeaturePreprocessorChoice
-from autosklearn.pipeline.components.data_preprocessing.categorical_encoding import OHEChoice
+from autosklearn.pipeline.components.data_preprocessing.categorical_encoding import (
+ OHEChoice,
+)
from autosklearn.pipeline.components.data_preprocessing.minority_coalescense import (
- CoalescenseChoice
+ CoalescenseChoice,
)
from autosklearn.pipeline.components.data_preprocessing.rescaling import RescalingChoice
+from autosklearn.pipeline.components.feature_preprocessing import (
+ FeaturePreprocessorChoice,
+)
+from autosklearn.pipeline.components.regression import RegressorChoice
+from autosklearn.smbo import AutoMLSMBO
+from autosklearn.util import RE_PATTERN, pipeline
+from autosklearn.util.data import (
+ DatasetCompressionSpec,
+ default_dataset_compression_arg,
+ reduce_dataset_size_if_too_large,
+ supported_precision_reductions,
+ validate_dataset_compression_arg,
+)
+from autosklearn.util.logging_ import (
+ PicklableClientLogger,
+ get_named_client_logger,
+ setup_logger,
+ start_log_server,
+ warnings_to,
+)
+from autosklearn.util.parallel import preload_modules
from autosklearn.util.single_thread_client import SingleThreadedClient
+from autosklearn.util.smac_wrap import SMACCallback, SmacRunCallback
+from autosklearn.util.stopwatch import StopWatch
+
+import unittest.mock
def _model_predict(
@@ -85,7 +134,7 @@ def _model_predict(
batch_size: Optional[int] = None,
logger: Optional[PicklableClientLogger] = None,
) -> np.ndarray:
- """ Generates the predictions from a model.
+ """Generates the predictions from a model.
This is seperated out into a seperate function to allow for multiprocessing
and perform parallel predictions.
@@ -140,158 +189,176 @@ def _model_predict(
else:
predict_func = model.predict
- if batch_size is not None and hasattr(model, 'batch_size'):
+ if batch_size is not None and hasattr(model, "batch_size"):
prediction = predict_func(X_, batch_size=batch_size)
else:
prediction = predict_func(X_)
# Check that probability values lie between 0 and 1.
if task in CLASSIFICATION_TASKS:
- assert (prediction >= 0).all() and (prediction <= 1).all(), \
- f"For {model}, prediction probability not within [0, 1]!"
+ assert (prediction >= 0).all() and (
+ prediction <= 1
+ ).all(), f"For {model}, prediction probability not within [0, 1]!"
- assert prediction.shape[0] == X_.shape[0], \
- f"Prediction shape {model} is {prediction.shape} while X_.shape is {X_.shape}"
+ assert (
+ prediction.shape[0] == X_.shape[0]
+ ), f"Prediction shape {model} is {prediction.shape} while X_.shape is {X_.shape}"
return prediction
class AutoML(BaseEstimator):
+ """Base class for handling the AutoML procedure"""
+
+ def __init__(
+ self,
+ time_left_for_this_task: int,
+ per_run_time_limit: int,
+ temporary_directory: Optional[str] = None,
+ delete_tmp_folder_after_terminate: bool = True,
+ initial_configurations_via_metalearning: int = 25,
+ ensemble_class: Type[AbstractEnsemble] | None = EnsembleSelection,
+ ensemble_kwargs: Dict[str, Any] | None = None,
+ ensemble_nbest: int = 1,
+ max_models_on_disc: int = 1,
+ seed: int = 1,
+ memory_limit: int = 3072,
+ metadata_directory: Optional[str] = None,
+ include: Optional[dict[str, list[str]]] = None,
+ exclude: Optional[dict[str, list[str]]] = None,
+ resampling_strategy: str | Any = "holdout-iterative-fit",
+ resampling_strategy_arguments: Mapping[str, Any] = None,
+ n_jobs: Optional[int] = None,
+ dask_client: Optional[Client] = None,
+ precision: Literal[16, 32, 64] = 32,
+ disable_evaluator_output: bool | Iterable[str] = False,
+ get_smac_object_callback: Optional[Callable] = None,
+ smac_scenario_args: Optional[Mapping] = None,
+ logging_config: Optional[Mapping] = None,
+ metrics: Sequence[Scorer] | None = None,
+ scoring_functions: Optional[list[Scorer]] = None,
+ get_trials_callback: SMACCallback | None = None,
+ dataset_compression: bool | Mapping[str, Any] = True,
+ allow_string_features: bool = True,
+ ):
+ super().__init__()
+
+ if isinstance(disable_evaluator_output, Iterable):
+ disable_evaluator_output = list(disable_evaluator_output) # Incase iterator
+ allowed = set(["model", "cv_model", "y_optimization", "y_test"])
+ unknown = allowed - set(disable_evaluator_output)
+ if any(unknown):
+ raise ValueError(
+ f"Unknown arg {unknown} for '_disable_evaluator_output',"
+ f" must be one of {allowed}"
+ )
+
+ # Validate dataset_compression and set its values
+ self._dataset_compression: Optional[DatasetCompressionSpec]
+ if isinstance(dataset_compression, bool):
+ if dataset_compression is True:
+ self._dataset_compression = default_dataset_compression_arg
+ else:
+ self._dataset_compression = None
+ else:
+ self._dataset_compression = validate_dataset_compression_arg(
+ dataset_compression,
+ memory_limit=memory_limit,
+ )
+
+ # If we got something callable for `get_trials_callback`, wrap it so SMAC
+ # will accept it.
+ if (
+ get_trials_callback is not None
+ and callable(get_trials_callback)
+ and not isinstance(get_trials_callback, IncorporateRunResultCallback)
+ ):
+ get_trials_callback = SmacRunCallback(get_trials_callback)
- def __init__(self,
- time_left_for_this_task,
- per_run_time_limit,
- temporary_directory: Optional[str] = None,
- delete_tmp_folder_after_terminate: bool = True,
- initial_configurations_via_metalearning=25,
- ensemble_size=1,
- ensemble_nbest=1,
- max_models_on_disc=1,
- seed=1,
- memory_limit=3072,
- metadata_directory=None,
- debug_mode=False,
- include=None,
- exclude=None,
- resampling_strategy='holdout-iterative-fit',
- resampling_strategy_arguments=None,
- n_jobs=None,
- dask_client: Optional[dask.distributed.Client] = None,
- precision=32,
- disable_evaluator_output=False,
- get_smac_object_callback=None,
- smac_scenario_args=None,
- logging_config=None,
- metric=None,
- scoring_functions=None,
- get_trials_callback=None
- ):
- super(AutoML, self).__init__()
- self.configuration_space = None
- self._backend: Optional[Backend] = None
- self._temporary_directory = temporary_directory
self._delete_tmp_folder_after_terminate = delete_tmp_folder_after_terminate
- # self._tmp_dir = tmp_dir
self._time_for_task = time_left_for_this_task
self._per_run_time_limit = per_run_time_limit
- self._initial_configurations_via_metalearning = \
- initial_configurations_via_metalearning
- self._ensemble_size = ensemble_size
+ self._metrics = metrics
+ self._ensemble_class = ensemble_class
+ self._ensemble_kwargs = ensemble_kwargs
self._ensemble_nbest = ensemble_nbest
self._max_models_on_disc = max_models_on_disc
self._seed = seed
self._memory_limit = memory_limit
- self._data_memory_limit = None
self._metadata_directory = metadata_directory
self._include = include
self._exclude = exclude
self._resampling_strategy = resampling_strategy
- self._scoring_functions = scoring_functions if scoring_functions is not None else []
- self._resampling_strategy_arguments = resampling_strategy_arguments \
- if resampling_strategy_arguments is not None else {}
- self._n_jobs = n_jobs
- self._dask_client = dask_client
-
- self.precision = precision
self._disable_evaluator_output = disable_evaluator_output
- # Check arguments prior to doing anything!
- if not isinstance(self._disable_evaluator_output, (bool, list)):
- raise ValueError('disable_evaluator_output must be of type bool '
- 'or list.')
- if isinstance(self._disable_evaluator_output, list):
- allowed_elements = ['model', 'cv_model', 'y_optimization', 'y_test', 'y_valid']
- for element in self._disable_evaluator_output:
- if element not in allowed_elements:
- raise ValueError("List member '%s' for argument "
- "'disable_evaluator_output' must be one "
- "of " + str(allowed_elements))
self._get_smac_object_callback = get_smac_object_callback
self._get_trials_callback = get_trials_callback
self._smac_scenario_args = smac_scenario_args
self.logging_config = logging_config
+ self.precision = precision
+ self.allow_string_features = allow_string_features
+ self._initial_configurations_via_metalearning = (
+ initial_configurations_via_metalearning
+ )
+
+ self._scoring_functions = scoring_functions or []
+ self._resampling_strategy_arguments = resampling_strategy_arguments or {}
+ # Single core, local runs should use fork to prevent the __main__ requirements
+ # in examples. Nevertheless, multi-process runs have spawn as requirement to
+ # reduce the possibility of a deadlock
+ if n_jobs == 1 and dask_client is None:
+ self._multiprocessing_context = "fork"
+ self._dask_client = SingleThreadedClient()
+ self._n_jobs = 1
+ else:
+ self._multiprocessing_context = "forkserver"
+ self._dask_client = dask_client
+ self._n_jobs = n_jobs
+
+ # Create the backend
+ self._backend: Backend = create(
+ temporary_directory=temporary_directory,
+ output_directory=None,
+ prefix="auto-sklearn",
+ delete_output_folder_after_terminate=delete_tmp_folder_after_terminate,
+ )
+
+ self._data_memory_limit = None # TODO: dead variable? Always None
self._datamanager = None
self._dataset_name = None
self._feat_type = None
- self._stopwatch = StopWatch()
- self._logger = None
+ self._logger: Optional[PicklableClientLogger] = None
self._task = None
-
- self._metric = metric
-
self._label_num = None
self._parser = None
- self.models_ = None
- self.cv_models_ = None
- self.ensemble_ = None
self._can_predict = False
- self._debug_mode = debug_mode
-
- self.InputValidator = None # type: Optional[InputValidator]
+ self._read_at_most = None
+ self._max_ensemble_build_iterations = None
+ self.models_: Optional[dict] = None
+ self.cv_models_: Optional[dict] = None
+ self.ensemble_ = None
+ self.InputValidator: Optional[InputValidator] = None
+ self.configuration_space = None
# The ensemble performance history through time
- self.ensemble_performance_history = []
-
- # Single core, local runs should use fork
- # to prevent the __main__ requirements in
- # examples. Nevertheless, multi-process runs
- # have spawn as requirement to reduce the
- # possibility of a deadlock
- self._multiprocessing_context = 'forkserver'
- if self._n_jobs == 1 and self._dask_client is None:
- self._multiprocessing_context = 'fork'
- self._dask_client = SingleThreadedClient()
-
- if not isinstance(self._time_for_task, int):
- raise ValueError("time_left_for_this_task not of type integer, "
- "but %s" % str(type(self._time_for_task)))
- if not isinstance(self._per_run_time_limit, int):
- raise ValueError("per_run_time_limit not of type integer, but %s" %
- str(type(self._per_run_time_limit)))
-
- # By default try to use the TCP logging port or get a new port
+ self._stopwatch = StopWatch()
self._logger_port = logging.handlers.DEFAULT_TCP_LOGGING_PORT
+ self.ensemble_performance_history = []
- # Num_run tell us how many runs have been launched
- # It can be seen as an identifier for each configuration
- # saved to disk
+ # Num_run tell us how many runs have been launched. It can be seen as an
+ # identifier for each configuration saved to disk
self.num_run = 0
+ self.fitted = False
- def _create_backend(self) -> Backend:
- return create(
- temporary_directory=self._temporary_directory,
- delete_tmp_folder_after_terminate=self._delete_tmp_folder_after_terminate,
- )
-
- def _create_dask_client(self):
+ def _create_dask_client(self) -> None:
self._is_dask_client_internally_created = True
- self._dask_client = dask.distributed.Client(
- dask.distributed.LocalCluster(
+ self._dask_client = Client(
+ LocalCluster(
n_workers=self._n_jobs,
processes=False,
threads_per_worker=1,
# We use the temporal directory to save the
- # dask workers, because deleting workers
+ # dask workers, because deleting workers takes
# more time than deleting backend directories
# This prevent an error saying that the worker
# file was deleted, so the client could not close
@@ -304,11 +371,9 @@ def _create_dask_client(self):
heartbeat_interval=10000,
)
- def _close_dask_client(self):
- if (
- hasattr(self, '_is_dask_client_internally_created')
- and self._is_dask_client_internally_created
- and self._dask_client
+ def _close_dask_client(self, force: bool = False) -> None:
+ if getattr(self, "_dask_client", None) is not None and (
+ force or getattr(self, "_is_dask_client_internally_created", False)
):
self._dask_client.shutdown()
self._dask_client.close()
@@ -317,14 +382,14 @@ def _close_dask_client(self):
self._is_dask_client_internally_created = False
del self._is_dask_client_internally_created
- def _get_logger(self, name):
- logger_name = 'AutoML(%d):%s' % (self._seed, name)
+ def _get_logger(self, name: str) -> PicklableClientLogger:
+ logger_name = "AutoML(%d):%s" % (self._seed, name)
# Setup the configuration for the logger
# This is gonna be honored by the server
# Which is created below
setup_logger(
- filename='%s.log' % str(logger_name),
+ filename="%s.log" % str(logger_name),
logging_config=self.logging_config,
output_dir=self._backend.temporary_directory,
)
@@ -338,17 +403,17 @@ def _get_logger(self, name):
context = multiprocessing.get_context(self._multiprocessing_context)
preload_modules(context)
self.stop_logging_server = context.Event()
- port = context.Value('l') # be safe by using a long
+ port = context.Value("l") # be safe by using a long
port.value = -1
self.logging_server = context.Process(
target=start_log_server,
kwargs=dict(
- host='localhost',
+ host="localhost",
logname=logger_name,
event=self.stop_logging_server,
port=port,
- filename='%s.log' % str(logger_name),
+ filename="%s.log" % str(logger_name),
logging_config=self.logging_config,
output_dir=self._backend.temporary_directory,
),
@@ -367,12 +432,12 @@ def _get_logger(self, name):
return get_named_client_logger(
name=logger_name,
- host='localhost',
+ host="localhost",
port=self._logger_port,
)
- def _clean_logger(self):
- if not hasattr(self, 'stop_logging_server') or self.stop_logging_server is None:
+ def _clean_logger(self) -> None:
+ if not hasattr(self, "stop_logging_server") or self.stop_logging_server is None:
return
# Clean up the logger
@@ -389,30 +454,16 @@ def _clean_logger(self):
self.logging_server.terminate()
del self.stop_logging_server
- @staticmethod
- def _start_task(watcher, task_name):
- watcher.start_task(task_name)
-
- @staticmethod
- def _stop_task(watcher, task_name):
- watcher.stop_task(task_name)
-
- @staticmethod
- def _print_load_time(basename, time_left_for_this_task,
- time_for_load_data, logger):
-
- time_left_after_reading = max(
- 0, time_left_for_this_task - time_for_load_data)
- logger.info('Remaining time after reading %s %5.2f sec' %
- (basename, time_left_after_reading))
- return time_for_load_data
+ def _do_dummy_prediction(self) -> None:
+ # When using partial-cv it makes no sense to do dummy predictions
+ if self._resampling_strategy in ["partial-cv", "partial-cv-iterative-fit"]:
+ return
- def _do_dummy_prediction(self, datamanager: XYDataManager, num_run: int) -> int:
+ if self._metrics is None:
+ raise ValueError("Metric/Metrics was/were not set")
- # When using partial-cv it makes no sense to do dummy predictions
- if self._resampling_strategy in ['partial-cv',
- 'partial-cv-iterative-fit']:
- return num_run
+ # Dummy prediction always have num_run set to 1
+ dummy_run_num = 1
self._logger.info("Starting to create dummy predictions.")
@@ -426,54 +477,50 @@ def _do_dummy_prediction(self, datamanager: XYDataManager, num_run: int) -> int:
# already be generated here!
stats = Stats(scenario_mock)
stats.start_timing()
- ta = ExecuteTaFuncWithQueue(backend=self._backend,
- autosklearn_seed=self._seed,
- resampling_strategy=self._resampling_strategy,
- initial_num_run=num_run,
- stats=stats,
- metric=self._metric,
- memory_limit=memory_limit,
- disable_file_output=self._disable_evaluator_output,
- abort_on_first_run_crash=False,
- cost_for_crash=get_cost_of_crash(self._metric),
- port=self._logger_port,
- pynisher_context=self._multiprocessing_context,
- **self._resampling_strategy_arguments)
-
- status, cost, runtime, additional_info = ta.run(num_run, cutoff=self._time_for_task)
+ ta = ExecuteTaFuncWithQueue(
+ backend=self._backend,
+ autosklearn_seed=self._seed,
+ multi_objectives=[metric.name for metric in self._metrics],
+ resampling_strategy=self._resampling_strategy,
+ initial_num_run=dummy_run_num,
+ stats=stats,
+ metrics=self._metrics,
+ memory_limit=memory_limit,
+ disable_file_output=self._disable_evaluator_output,
+ abort_on_first_run_crash=False,
+ cost_for_crash=get_cost_of_crash(self._metrics),
+ port=self._logger_port,
+ pynisher_context=self._multiprocessing_context,
+ **self._resampling_strategy_arguments,
+ )
+
+ status, cost, runtime, additional_info = ta.run(
+ config=dummy_run_num,
+ cutoff=self._time_for_task,
+ )
if status == StatusType.SUCCESS:
self._logger.info("Finished creating dummy predictions.")
+
+ # Fail if dummy prediction fails.
else:
- if additional_info.get('exitcode') == -6:
- self._logger.error(
- "Dummy prediction failed with run state %s. "
- "The error suggests that the provided memory limits were too tight. Please "
- "increase the 'memory_limit' and try again. If this does not solve your "
- "problem, please open an issue and paste the additional output. "
- "Additional output: %s.",
- str(status), str(additional_info),
+ if additional_info.get("exitcode") == -6:
+ msg = (
+ f"Dummy prediction failed with run state {status}."
+ " The error suggests that the provided memory limits are too tight."
+ " Please increase the 'memory_limit' and try again. If this does"
+ " not solve your problem, please open an issue and paste the"
+ f" additional output. Additional output: {additional_info}"
)
- # Fail if dummy prediction fails.
- raise ValueError(
- "Dummy prediction failed with run state %s. "
- "The error suggests that the provided memory limits were too tight. Please "
- "increase the 'memory_limit' and try again. If this does not solve your "
- "problem, please open an issue and paste the additional output. "
- "Additional output: %s." %
- (str(status), str(additional_info)),
- )
-
else:
- self._logger.error(
- "Dummy prediction failed with run state %s and additional output: %s.",
- str(status), str(additional_info),
+ msg = (
+ f" Dummy prediction failed with run state {status} and"
+ f" additional output: {additional_info}.",
)
- # Fail if dummy prediction fails.
- raise ValueError(
- "Dummy prediction failed with run state %s and additional output: %s."
- % (str(status), str(additional_info))
- )
- return num_run
+
+ self._logger.error(msg)
+ raise ValueError(msg)
+
+ return
@classmethod
def _task_type_id(cls, task_type: str) -> int:
@@ -486,11 +533,11 @@ def _supports_task_type(cls, task_type: str) -> bool:
def fit(
self,
X: SUPPORTED_FEAT_TYPES,
- y: Union[SUPPORTED_TARGET_TYPES, spmatrix],
+ y: SUPPORTED_TARGET_TYPES,
task: Optional[int] = None,
X_test: Optional[SUPPORTED_FEAT_TYPES] = None,
- y_test: Optional[Union[SUPPORTED_TARGET_TYPES, spmatrix]] = None,
- feat_type: Optional[List[str]] = None,
+ y_test: Optional[SUPPORTED_TARGET_TYPES] = None,
+ feat_type: Optional[list[str]] = None,
dataset_name: Optional[str] = None,
only_return_configuration_space: bool = False,
load_models: bool = True,
@@ -499,20 +546,20 @@ def fit(
"""Fit AutoML to given training set (X, y).
Fit both optimizes the machine learning models and builds an ensemble
- out of them. To disable ensembling, set ``ensemble_size==0``.
+ out of them.
# TODO PR1213
#
# `task: Optional[int]` and `is_classification`
#
- # `AutoML` tries to identify the task itself with
- # `sklearn.type_of_target`, leaving little for the subclasses to do.
+ # `AutoML` tries to identify the task itself with `sklearn.type_of_target`,
+ # leaving little for the subclasses to do.
# Except this failes when type_of_target(y) == "multiclass".
#
# "multiclass" be mean either REGRESSION or MULTICLASS_CLASSIFICATION,
# and so this is where the subclasses are used to determine which.
# However, this could also be deduced from the `is_classification`
- # paramaeter.
+ # parameter.
#
# In the future, there is little need for the subclasses of `AutoML`
# and no need for the `task` parameter. The extra functionality
@@ -535,7 +582,6 @@ def fit(
Parameters
----------
-
X : {array-like, sparse matrix}, shape (n_samples, n_features)
The training input samples.
@@ -555,7 +601,7 @@ def fit(
of all models. This allows to evaluate the performance of
Auto-sklearn over time.
- feat_type : Optional[List],
+ feat_type : Optional[list],
List of str of `len(X.shape[1])` describing the attribute type.
Possible types are `Categorical` and `Numerical`. `Categorical`
attributes will be automatically One-Hot encoded. The values
@@ -582,8 +628,9 @@ def fit(
Returns
-------
self
-
"""
+ if (X_test is not None) ^ (y_test is not None):
+ raise ValueError("Must provide both X_test and y_test together")
# AutoSklearn does not handle sparse y for now
y = convert_if_sparse(y)
@@ -593,382 +640,455 @@ def fit(
if task is None:
y_task = type_of_target(y)
if not self._supports_task_type(y_task):
- raise ValueError(f"{self.__class__.__name__} does not support"
- f" task {y_task}")
+ raise ValueError(
+ f"{self.__class__.__name__} does not support" f" task {y_task}"
+ )
self._task = self._task_type_id(y_task)
else:
self._task = task
# Assign a metric if it doesnt exist
- if self._metric is None:
- self._metric = default_metric_for_task[self._task]
+ if self._metrics is None:
+ self._metrics = [default_metric_for_task[self._task]]
+ _validate_metrics(self._metrics, self._scoring_functions)
if dataset_name is None:
dataset_name = str(uuid.uuid1(clock_seq=os.getpid()))
- # Create the backend
- self._backend = self._create_backend()
-
# By default try to use the TCP logging port or get a new port
self._logger_port = logging.handlers.DEFAULT_TCP_LOGGING_PORT
- self._logger = self._get_logger(dataset_name)
- # The first thing we have to do is create the logger to update the backend
- self._backend.setup_logger(self._logger_port)
+ # Once we start the logging server, it starts in a new process
+ # If an error occurs then we want to make sure that we exit cleanly
+ # and shut it down, else it might hang
+ # https://github.com/automl/auto-sklearn/issues/1480
+ try:
+ self._logger = self._get_logger(dataset_name)
- if not only_return_configuration_space:
- # If only querying the configuration space, we do not save the start time
- # The start time internally checks for the fit() method to execute only once
- # But this does not apply when only querying the configuration space
- self._backend.save_start_time(self._seed)
+ # The first thing we have to do is create the logger to update the backend
+ self._backend.setup_logger(self._logger_port)
- self._stopwatch = StopWatch()
+ if not only_return_configuration_space:
+ # If only querying the configuration space, we do not save the start
+ # time The start time internally checks for the fit() method to execute
+ # only once but this does not apply when only querying the configuration
+ # space
+ self._backend.save_start_time(self._seed)
- # Make sure that input is valid
- # Performs Ordinal one hot encoding to the target
- # both for train and test data
- self.InputValidator = InputValidator(
- is_classification=is_classification,
- feat_type=feat_type,
- logger_port=self._logger_port,
- )
- self.InputValidator.fit(X_train=X, y_train=y, X_test=X_test, y_test=y_test)
- X, y = self.InputValidator.transform(X, y)
+ self._stopwatch = StopWatch()
- if X_test is not None:
- X_test, y_test = self.InputValidator.transform(X_test, y_test)
+ # Make sure that input is valid
+ # Performs Ordinal one hot encoding to the target
+ # both for train and test data
+ self.InputValidator = InputValidator(
+ is_classification=is_classification,
+ feat_type=feat_type,
+ logger_port=self._logger_port,
+ allow_string_features=self.allow_string_features,
+ )
+ self.InputValidator.fit(X_train=X, y_train=y, X_test=X_test, y_test=y_test)
+ X, y = self.InputValidator.transform(X, y)
- X, y = self.subsample_if_too_large(
- X=X,
- y=y,
- logger=self._logger,
- seed=self._seed,
- memory_limit=self._memory_limit,
- task=self._task,
- )
+ if X_test is not None and y_test is not None:
+ X_test, y_test = self.InputValidator.transform(X_test, y_test)
- # Check the re-sampling strategy
- try:
+ # We don't support size reduction on pandas type object yet
+ if (
+ self._dataset_compression is not None
+ and not isinstance(X, pd.DataFrame)
+ and not (isinstance(y, pd.Series) or isinstance(y, pd.DataFrame))
+ ):
+ methods = self._dataset_compression["methods"]
+ memory_allocation = self._dataset_compression["memory_allocation"]
+
+ # Remove precision reduction if we can't perform it
+ if (
+ "precision" in methods
+ and X.dtype not in supported_precision_reductions
+ ):
+ methods = [method for method in methods if method != "precision"]
+
+ with warnings_to(self._logger):
+ X, y = reduce_dataset_size_if_too_large(
+ X=X,
+ y=y,
+ memory_limit=self._memory_limit,
+ is_classification=is_classification,
+ random_state=self._seed,
+ operations=methods,
+ memory_allocation=memory_allocation,
+ )
+
+ # Check the re-sampling strategy
self._check_resampling_strategy(
- X=X, y=y, task=self._task,
+ X=X,
+ y=y,
+ task=self._task,
)
- except Exception as e:
- self._fit_cleanup()
- raise e
- # Reset learnt stuff
- self.models_ = None
- self.cv_models_ = None
- self.ensemble_ = None
+ # Reset learnt stuff
+ self.models_ = None
+ self.cv_models_ = None
+ self.ensemble_ = None
+
+ # The metric must exist as of this point
+ # It can be provided in the constructor, or automatically
+ # defined in the estimator fit call
+ if isinstance(self._metrics, Sequence):
+ for entry in self._metrics:
+ if not isinstance(entry, Scorer):
+ raise ValueError(
+ f"Metric {entry} must be instance of"
+ " autosklearn.metrics.Scorer."
+ )
+ else:
+ raise ValueError(
+ "Metric must be a sequence of instances of "
+ "autosklearn.metrics.Scorer."
+ )
- # The metric must exist as of this point
- # It can be provided in the constructor, or automatically
- # defined in the estimator fit call
- if self._metric is None:
- raise ValueError('No metric given.')
- if not isinstance(self._metric, Scorer):
- raise ValueError('Metric must be instance of '
- 'autosklearn.metrics.Scorer.')
-
- # If no dask client was provided, we create one, so that we can
- # start a ensemble process in parallel to smbo optimize
- if (
- self._dask_client is None and
- (self._ensemble_size > 0 or self._n_jobs is not None and self._n_jobs > 1)
- ):
- self._create_dask_client()
- else:
- self._is_dask_client_internally_created = False
+ # If no dask client was provided, we create one, so that we can
+ # start a ensemble process in parallel to smbo optimize
+ if self._dask_client is None and (
+ self._ensemble_class is not None
+ or self._n_jobs is not None
+ and self._n_jobs > 1
+ ):
+ self._create_dask_client()
+ else:
+ self._is_dask_client_internally_created = False
+
+ self._dataset_name = dataset_name
+ self._stopwatch.start(self._dataset_name)
+
+ # Take the feature types from the validator
+ self._feat_type = self.InputValidator.feature_validator.feat_type
+
+ self._log_fit_setup()
+
+ # == Pickle the data manager to speed up loading
+ with self._stopwatch.time("Save Datamanager"):
+ datamanager = XYDataManager(
+ X,
+ y,
+ X_test=X_test,
+ y_test=y_test,
+ task=self._task,
+ feat_type=self._feat_type,
+ dataset_name=dataset_name,
+ )
+
+ self._backend._make_internals_directory()
+ self._label_num = datamanager.info["label_num"]
+
+ self._backend.save_datamanager(datamanager)
+
+ # = Create a searchspace
+ # Do this before One Hot Encoding to make sure that it creates a
+ # search space for a dense classifier even if one hot encoding would
+ # make it sparse (tradeoff; if one hot encoding would make it sparse,
+ # densifier and truncatedSVD would probably lead to a MemoryError,
+ # like this we can't use some of the preprocessing methods in case
+ # the data became sparse)
+ with self._stopwatch.time("Create Search space"):
+ self.configuration_space, configspace_path = self._create_search_space(
+ self._backend.temporary_directory,
+ self._backend,
+ datamanager,
+ include=self._include,
+ exclude=self._exclude,
+ )
+
+ if only_return_configuration_space:
+ return self.configuration_space
+
+ # == Perform dummy predictions
+ with self._stopwatch.time("Dummy predictions"):
+ self.num_run += 1
+ self._do_dummy_prediction()
+
+ # == RUN ensemble builder
+ # Do this before calculating the meta-features to make sure that the
+ # dummy predictions are actually included in the ensemble even if
+ # calculating the meta-features takes very long
+ with self._stopwatch.time("Run Ensemble Builder"):
+
+ elapsed_time = self._stopwatch.time_since(self._dataset_name, "start")
+
+ time_left_for_ensembles = max(0, self._time_for_task - elapsed_time)
+ proc_ensemble = None
+ if time_left_for_ensembles <= 0:
+ # Fit only raises error when an ensemble class is given but
+ # time_left_for_ensembles is zero.
+ if self._ensemble_class is not None:
+ raise ValueError(
+ "Not starting ensemble builder because there "
+ "is no time left. Try increasing the value "
+ "of time_left_for_this_task."
+ )
+ elif self._ensemble_class is None:
+ self._logger.info(
+ "No ensemble buildin because no ensemble class was given."
+ )
+ else:
+ self._logger.info(
+ "Start Ensemble with %5.2fsec time left"
+ % time_left_for_ensembles
+ )
+
+ proc_ensemble = EnsembleBuilderManager(
+ start_time=time.time(),
+ time_left_for_ensembles=time_left_for_ensembles,
+ backend=copy.deepcopy(self._backend),
+ dataset_name=dataset_name,
+ task=self._task,
+ metrics=self._metrics,
+ ensemble_class=self._ensemble_class,
+ ensemble_kwargs=self._ensemble_kwargs,
+ ensemble_nbest=self._ensemble_nbest,
+ max_models_on_disc=self._max_models_on_disc,
+ seed=self._seed,
+ precision=self.precision,
+ max_iterations=self._max_ensemble_build_iterations,
+ read_at_most=self._read_at_most,
+ memory_limit=self._memory_limit,
+ random_state=self._seed,
+ logger_port=self._logger_port,
+ pynisher_context=self._multiprocessing_context,
+ )
+
+ # kill the datamanager as it will be re-loaded anyways from sub processes
+ try:
+ del self._datamanager
+ except Exception:
+ pass
+
+ # => RUN SMAC
+ with self._stopwatch.time("Run SMAC"):
+ elapsed_time = self._stopwatch.time_since(self._dataset_name, "start")
+ time_left = self._time_for_task - elapsed_time
+
+ if self._logger:
+ self._logger.info("Start SMAC with %5.2fsec time left" % time_left)
+ if time_left <= 0:
+ self._logger.warning(
+ "Not starting SMAC because there is no time left."
+ )
+ _proc_smac = None
+ self._budget_type = None
+ else:
+ if (
+ self._per_run_time_limit is None
+ or self._per_run_time_limit > time_left
+ ):
+ self._logger.warning(
+ "Time limit for a single run is higher than total time "
+ "limit. Capping the limit for a single run to the total "
+ "time given to SMAC (%f)" % time_left
+ )
+ per_run_time_limit = time_left
+ else:
+ per_run_time_limit = self._per_run_time_limit
+
+ # At least 2 models are created for the ensemble process
+ num_models = time_left // per_run_time_limit
+ if num_models < 2:
+ per_run_time_limit = time_left // 2
+ self._logger.warning(
+ "Capping the per_run_time_limit to {} to have "
+ "time for a least 2 models in each process.".format(
+ per_run_time_limit
+ )
+ )
+
+ n_meta_configs = self._initial_configurations_via_metalearning
+ _proc_smac = AutoMLSMBO(
+ config_space=self.configuration_space,
+ dataset_name=self._dataset_name,
+ backend=self._backend,
+ total_walltime_limit=time_left,
+ func_eval_time_limit=per_run_time_limit,
+ memory_limit=self._memory_limit,
+ data_memory_limit=self._data_memory_limit,
+ stopwatch=self._stopwatch,
+ n_jobs=self._n_jobs,
+ dask_client=self._dask_client,
+ start_num_run=self.num_run,
+ num_metalearning_cfgs=n_meta_configs,
+ config_file=configspace_path,
+ seed=self._seed,
+ metadata_directory=self._metadata_directory,
+ metrics=self._metrics,
+ resampling_strategy=self._resampling_strategy,
+ resampling_strategy_args=self._resampling_strategy_arguments,
+ include=self._include,
+ exclude=self._exclude,
+ disable_file_output=self._disable_evaluator_output,
+ get_smac_object_callback=self._get_smac_object_callback,
+ smac_scenario_args=self._smac_scenario_args,
+ scoring_functions=self._scoring_functions,
+ port=self._logger_port,
+ pynisher_context=self._multiprocessing_context,
+ ensemble_callback=proc_ensemble,
+ trials_callback=self._get_trials_callback,
+ )
- self._dataset_name = dataset_name
- self._stopwatch.start_task(self._dataset_name)
+ (
+ self.runhistory_,
+ self.trajectory_,
+ self._budget_type,
+ ) = _proc_smac.run_smbo()
+ trajectory_filename = os.path.join(
+ self._backend.get_smac_output_directory_for_run(self._seed),
+ "trajectory.json",
+ )
+ saveable_trajectory = [
+ list(entry[:2]) + [entry[2].get_dictionary()] + list(entry[3:])
+ for entry in self.trajectory_
+ ]
+ with open(trajectory_filename, "w") as fh:
+ json.dump(saveable_trajectory, fh)
+
+ self._logger.info("Starting shutdown...")
+ # Wait until the ensemble process is finished to avoid shutting down
+ # while the ensemble builder tries to access the data
+ if proc_ensemble is not None:
+ self.ensemble_performance_history = list(proc_ensemble.history)
+
+ if len(proc_ensemble.futures) > 0:
+ # Now we wait for the future to return as it cannot be cancelled
+ # while it is running: https://stackoverflow.com/a/49203129
+ self._logger.info(
+ "Ensemble script still running, waiting for it to finish."
+ )
+ result = proc_ensemble.futures.pop().result()
+ if result:
+ ensemble_history, _ = result
+ self.ensemble_performance_history.extend(ensemble_history)
+ self._logger.info("Ensemble script finished, continue shutdown.")
+
+ # save the ensemble performance history file
+ if len(self.ensemble_performance_history) > 0:
+ pd.DataFrame(self.ensemble_performance_history).to_json(
+ os.path.join(
+ self._backend.internals_directory, "ensemble_history.json"
+ )
+ )
- # Take the feature types from the validator
- self._feat_type = self.InputValidator.feature_validator.feat_type
+ if load_models:
+ self._logger.info("Loading models...")
+ self._load_models()
+ self._logger.info("Finished loading models...")
+ # The whole logic above from where we begin the logging server is capture
+ # in a try: finally: so that if something goes wrong, we at least close
+ # down the logging server, preventing it from hanging and not closing
+ # until ctrl+c is pressed
+ except Exception as e:
+ # This will be called before the _fit_cleanup
+ self._logger.exception(e)
+ raise e
+ finally:
+ self._fit_cleanup()
+
+ self.fitted = True
+
+ return self
+
+ def _log_fit_setup(self) -> None:
# Produce debug information to the logfile
- self._logger.debug('Starting to print environment information')
- self._logger.debug(' Python version: %s', sys.version.split('\n'))
+ self._logger.debug("Starting to print environment information")
+ self._logger.debug(" Python version: %s", sys.version.split("\n"))
try:
- self._logger.debug(' Distribution: %s', platform.linux_distribution())
+ self._logger.debug(
+ f"\tDistribution: {distro.id()}-{distro.version()}-{distro.name()}"
+ )
except AttributeError:
- # platform.linux_distribution() was removed in Python3.8
- # We should move to the distro package as soon as it supports Windows and OSX
pass
- self._logger.debug(' System: %s', platform.system())
- self._logger.debug(' Machine: %s', platform.machine())
- self._logger.debug(' Platform: %s', platform.platform())
+
+ self._logger.debug(" System: %s", platform.system())
+ self._logger.debug(" Machine: %s", platform.machine())
+ self._logger.debug(" Platform: %s", platform.platform())
# UNAME appears to leak sensible information
# self._logger.debug(' uname: %s', platform.uname())
- self._logger.debug(' Version: %s', platform.version())
- self._logger.debug(' Mac version: %s', platform.mac_ver())
- requirements = pkg_resources.resource_string('autosklearn', 'requirements.txt')
- requirements = requirements.decode('utf-8')
- requirements = [requirement for requirement in requirements.split('\n')]
+ self._logger.debug(" Version: %s", platform.version())
+ self._logger.debug(" Mac version: %s", platform.mac_ver())
+ requirements = pkg_resources.resource_string("autosklearn", "requirements.txt")
+ requirements = requirements.decode("utf-8")
+ requirements = [requirement for requirement in requirements.split("\n")]
for requirement in requirements:
if not requirement:
continue
match = RE_PATTERN.match(requirement)
if match:
- name = match.group('name')
+ name = match.group("name")
module_dist = pkg_resources.get_distribution(name)
- self._logger.debug(' %s', module_dist)
+ self._logger.debug(" %s", module_dist)
else:
- raise ValueError('Unable to read requirement: %s' % requirement)
- self._logger.debug('Done printing environment information')
- self._logger.debug('Starting to print arguments to auto-sklearn')
- self._logger.debug(' tmp_folder: %s', self._backend.context._temporary_directory)
- self._logger.debug(' time_left_for_this_task: %f', self._time_for_task)
- self._logger.debug(' per_run_time_limit: %f', self._per_run_time_limit)
+ raise ValueError("Unable to read requirement: %s" % requirement)
+
+ self._logger.debug("Done printing environment information")
+ self._logger.debug("Starting to print arguments to auto-sklearn")
+ self._logger.debug(" tmp_folder: %s", self._backend.temporary_directory)
+ self._logger.debug(" time_left_for_this_task: %f", self._time_for_task)
+ self._logger.debug(" per_run_time_limit: %f", self._per_run_time_limit)
self._logger.debug(
- ' initial_configurations_via_metalearning: %d',
+ " initial_configurations_via_metalearning: %d",
self._initial_configurations_via_metalearning,
)
- self._logger.debug(' ensemble_size: %d', self._ensemble_size)
- self._logger.debug(' ensemble_nbest: %f', self._ensemble_nbest)
- self._logger.debug(' max_models_on_disc: %s', str(self._max_models_on_disc))
- self._logger.debug(' seed: %d', self._seed)
- self._logger.debug(' memory_limit: %s', str(self._memory_limit))
- self._logger.debug(' metadata_directory: %s', self._metadata_directory)
- self._logger.debug(' debug_mode: %s', self._debug_mode)
- self._logger.debug(' include: %s', str(self._include))
- self._logger.debug(' exclude: %s', str(self._exclude))
- self._logger.debug(' resampling_strategy: %s', str(self._resampling_strategy))
- self._logger.debug(' resampling_strategy_arguments: %s',
- str(self._resampling_strategy_arguments))
- self._logger.debug(' n_jobs: %s', str(self._n_jobs))
- self._logger.debug(' multiprocessing_context: %s', str(self._multiprocessing_context))
- self._logger.debug(' dask_client: %s', str(self._dask_client))
- self._logger.debug(' precision: %s', str(self.precision))
- self._logger.debug(' disable_evaluator_output: %s', str(self._disable_evaluator_output))
- self._logger.debug(' get_smac_objective_callback: %s', str(self._get_smac_object_callback))
- self._logger.debug(' smac_scenario_args: %s', str(self._smac_scenario_args))
- self._logger.debug(' logging_config: %s', str(self.logging_config))
- self._logger.debug(' metric: %s', str(self._metric))
- self._logger.debug('Done printing arguments to auto-sklearn')
- self._logger.debug('Starting to print available components')
+ self._logger.debug(" ensemble_class: %s", self._ensemble_class)
+ self._logger.debug(" ensemble_kwargs: %s", self._ensemble_kwargs)
+ self._logger.debug(" ensemble_nbest: %f", self._ensemble_nbest)
+ self._logger.debug(" max_models_on_disc: %s", str(self._max_models_on_disc))
+ self._logger.debug(" seed: %d", self._seed)
+ self._logger.debug(" memory_limit: %s", str(self._memory_limit))
+ self._logger.debug(" metadata_directory: %s", self._metadata_directory)
+ self._logger.debug(" include: %s", str(self._include))
+ self._logger.debug(" exclude: %s", str(self._exclude))
+ self._logger.debug(" resampling_strategy: %s", str(self._resampling_strategy))
+ self._logger.debug(
+ " resampling_strategy_arguments: %s",
+ str(self._resampling_strategy_arguments),
+ )
+ self._logger.debug(" n_jobs: %s", str(self._n_jobs))
+ self._logger.debug(
+ " multiprocessing_context: %s", str(self._multiprocessing_context)
+ )
+ self._logger.debug(" dask_client: %s", str(self._dask_client))
+ self._logger.debug(" precision: %s", str(self.precision))
+ self._logger.debug(
+ " disable_evaluator_output: %s", str(self._disable_evaluator_output)
+ )
+ self._logger.debug(
+ " get_smac_objective_callback: %s", str(self._get_smac_object_callback)
+ )
+ self._logger.debug(" smac_scenario_args: %s", str(self._smac_scenario_args))
+ self._logger.debug(" logging_config: %s", str(self.logging_config))
+ if len(self._metrics) == 1:
+ self._logger.debug(" metric: %s", str(self._metrics[0]))
+ else:
+ self._logger.debug(" metrics: %s", str(self._metrics))
+ self._logger.debug("Done printing arguments to auto-sklearn")
+ self._logger.debug("Starting to print available components")
for choice in (
- ClassifierChoice, RegressorChoice, FeaturePreprocessorChoice,
- OHEChoice, RescalingChoice, CoalescenseChoice,
+ ClassifierChoice,
+ RegressorChoice,
+ FeaturePreprocessorChoice,
+ OHEChoice,
+ RescalingChoice,
+ CoalescenseChoice,
):
self._logger.debug(
- '%s: %s',
+ "%s: %s",
choice.__name__,
choice.get_components(),
)
- self._logger.debug('Done printing available components')
-
- datamanager = XYDataManager(
- X, y,
- X_test=X_test,
- y_test=y_test,
- task=self._task,
- feat_type=self._feat_type,
- dataset_name=dataset_name,
- )
-
- self._backend._make_internals_directory()
- self._label_num = datamanager.info['label_num']
-
- # == Pickle the data manager to speed up loading
- self._backend.save_datamanager(datamanager)
-
- time_for_load_data = self._stopwatch.wall_elapsed(self._dataset_name)
-
- if self._debug_mode:
- self._print_load_time(
- self._dataset_name,
- self._time_for_task,
- time_for_load_data,
- self._logger)
-
- # = Create a searchspace
- # Do this before One Hot Encoding to make sure that it creates a
- # search space for a dense classifier even if one hot encoding would
- # make it sparse (tradeoff; if one hot encoding would make it sparse,
- # densifier and truncatedSVD would probably lead to a MemoryError,
- # like this we can't use some of the preprocessing methods in case
- # the data became sparse)
- self.configuration_space, configspace_path = self._create_search_space(
- self._backend.temporary_directory,
- self._backend,
- datamanager,
- include=self._include,
- exclude=self._exclude,
- )
- if only_return_configuration_space:
- self._fit_cleanup()
- return self.configuration_space
-
- # == Perform dummy predictions
- # Dummy prediction always have num_run set to 1
- self.num_run += self._do_dummy_prediction(datamanager, num_run=1)
-
- # == RUN ensemble builder
- # Do this before calculating the meta-features to make sure that the
- # dummy predictions are actually included in the ensemble even if
- # calculating the meta-features takes very long
- ensemble_task_name = 'runEnsemble'
- self._stopwatch.start_task(ensemble_task_name)
- elapsed_time = self._stopwatch.wall_elapsed(self._dataset_name)
- time_left_for_ensembles = max(0, self._time_for_task - elapsed_time)
- proc_ensemble = None
- if time_left_for_ensembles <= 0:
- # Fit only raises error when ensemble_size is not zero but
- # time_left_for_ensembles is zero.
- if self._ensemble_size > 0:
- raise ValueError("Not starting ensemble builder because there "
- "is no time left. Try increasing the value "
- "of time_left_for_this_task.")
- elif self._ensemble_size <= 0:
- self._logger.info('Not starting ensemble builder because '
- 'ensemble size is <= 0.')
- else:
- self._logger.info(
- 'Start Ensemble with %5.2fsec time left' % time_left_for_ensembles)
-
- proc_ensemble = EnsembleBuilderManager(
- start_time=time.time(),
- time_left_for_ensembles=time_left_for_ensembles,
- backend=copy.deepcopy(self._backend),
- dataset_name=dataset_name,
- task=self._task,
- metric=self._metric,
- ensemble_size=self._ensemble_size,
- ensemble_nbest=self._ensemble_nbest,
- max_models_on_disc=self._max_models_on_disc,
- seed=self._seed,
- precision=self.precision,
- max_iterations=None,
- read_at_most=np.inf,
- ensemble_memory_limit=self._memory_limit,
- random_state=self._seed,
- logger_port=self._logger_port,
- pynisher_context=self._multiprocessing_context,
- )
-
- self._stopwatch.stop_task(ensemble_task_name)
-
- # kill the datamanager as it will be re-loaded anyways from sub processes
- try:
- del self._datamanager
- except Exception:
- pass
+ self._logger.debug("Done printing available components")
- # => RUN SMAC
- smac_task_name = 'runSMAC'
- self._stopwatch.start_task(smac_task_name)
- elapsed_time = self._stopwatch.wall_elapsed(self._dataset_name)
- time_left_for_smac = max(0, self._time_for_task - elapsed_time)
-
- if self._logger:
- self._logger.info(
- 'Start SMAC with %5.2fsec time left' % time_left_for_smac)
- if time_left_for_smac <= 0:
- self._logger.warning("Not starting SMAC because there is no time "
- "left.")
- _proc_smac = None
- self._budget_type = None
- else:
- if self._per_run_time_limit is None or \
- self._per_run_time_limit > time_left_for_smac:
- self._logger.warning(
- 'Time limit for a single run is higher than total time '
- 'limit. Capping the limit for a single run to the total '
- 'time given to SMAC (%f)' % time_left_for_smac
- )
- per_run_time_limit = time_left_for_smac
- else:
- per_run_time_limit = self._per_run_time_limit
-
- # Make sure that at least 2 models are created for the ensemble process
- num_models = time_left_for_smac // per_run_time_limit
- if num_models < 2:
- per_run_time_limit = time_left_for_smac//2
- self._logger.warning(
- "Capping the per_run_time_limit to {} to have "
- "time for a least 2 models in each process.".format(
- per_run_time_limit
- )
- )
+ def __sklearn_is_fitted__(self) -> bool:
+ return self.fitted
- _proc_smac = AutoMLSMBO(
- config_space=self.configuration_space,
- dataset_name=self._dataset_name,
- backend=self._backend,
- total_walltime_limit=time_left_for_smac,
- func_eval_time_limit=per_run_time_limit,
- memory_limit=self._memory_limit,
- data_memory_limit=self._data_memory_limit,
- watcher=self._stopwatch,
- n_jobs=self._n_jobs,
- dask_client=self._dask_client,
- start_num_run=self.num_run,
- num_metalearning_cfgs=self._initial_configurations_via_metalearning,
- config_file=configspace_path,
- seed=self._seed,
- metadata_directory=self._metadata_directory,
- metric=self._metric,
- resampling_strategy=self._resampling_strategy,
- resampling_strategy_args=self._resampling_strategy_arguments,
- include=self._include,
- exclude=self._exclude,
- disable_file_output=self._disable_evaluator_output,
- get_smac_object_callback=self._get_smac_object_callback,
- smac_scenario_args=self._smac_scenario_args,
- scoring_functions=self._scoring_functions,
- port=self._logger_port,
- pynisher_context=self._multiprocessing_context,
- ensemble_callback=proc_ensemble,
- trials_callback=self._get_trials_callback
- )
-
- try:
- self.runhistory_, self.trajectory_, self._budget_type = \
- _proc_smac.run_smbo()
- trajectory_filename = os.path.join(
- self._backend.get_smac_output_directory_for_run(self._seed),
- 'trajectory.json')
- saveable_trajectory = \
- [list(entry[:2]) + [entry[2].get_dictionary()] + list(entry[3:])
- for entry in self.trajectory_]
- with open(trajectory_filename, 'w') as fh:
- json.dump(saveable_trajectory, fh)
- except Exception as e:
- self._logger.exception(e)
- raise
-
- self._logger.info("Starting shutdown...")
- # Wait until the ensemble process is finished to avoid shutting down
- # while the ensemble builder tries to access the data
- if proc_ensemble is not None:
- self.ensemble_performance_history = list(proc_ensemble.history)
-
- if len(proc_ensemble.futures) > 0:
- # Now we need to wait for the future to return as it cannot be cancelled while it
- # is running: https://stackoverflow.com/a/49203129
- self._logger.info("Ensemble script still running, waiting for it to finish.")
- result = proc_ensemble.futures.pop().result()
- if result:
- ensemble_history, _, _, _, _ = result
- self.ensemble_performance_history.extend(ensemble_history)
- self._logger.info("Ensemble script finished, continue shutdown.")
-
- # save the ensemble performance history file
- if len(self.ensemble_performance_history) > 0:
- pd.DataFrame(self.ensemble_performance_history).to_json(
- os.path.join(self._backend.internals_directory, 'ensemble_history.json'))
-
- if load_models:
- self._logger.info("Loading models...")
- self._load_models()
- self._logger.info("Finished loading models...")
-
- self._fit_cleanup()
-
- return self
-
- def _fit_cleanup(self):
+ def _fit_cleanup(self) -> None:
self._logger.info("Closing the dask infrastructure")
self._close_dask_client()
self._logger.info("Finished closing the dask infrastructure")
@@ -1002,181 +1122,104 @@ def _check_resampling_strategy(
"""
is_split_object = isinstance(
self._resampling_strategy,
- (BaseCrossValidator, _RepeatedSplits, BaseShuffleSplit)
+ (BaseCrossValidator, _RepeatedSplits, BaseShuffleSplit),
)
- if self._resampling_strategy not in [
- 'holdout',
- 'holdout-iterative-fit',
- 'cv',
- 'cv-iterative-fit',
- 'partial-cv',
- 'partial-cv-iterative-fit',
- ] and not is_split_object:
- raise ValueError('Illegal resampling strategy: %s' % self._resampling_strategy)
+ if (
+ self._resampling_strategy
+ not in [
+ "holdout",
+ "holdout-iterative-fit",
+ "cv",
+ "cv-iterative-fit",
+ "partial-cv",
+ "partial-cv-iterative-fit",
+ ]
+ and not is_split_object
+ ):
+ raise ValueError(
+ "Illegal resampling strategy: %s" % self._resampling_strategy
+ )
elif is_split_object:
TrainEvaluator.check_splitter_resampling_strategy(
- X=X, y=y, task=task,
- groups=self._resampling_strategy_arguments.get('groups', None),
+ X=X,
+ y=y,
+ task=task,
+ groups=self._resampling_strategy_arguments.get("groups", None),
resampling_strategy=self._resampling_strategy,
)
- elif self._resampling_strategy in [
- 'partial-cv',
- 'partial-cv-iterative-fit',
- ] and self._ensemble_size != 0:
- raise ValueError("Resampling strategy %s cannot be used "
- "together with ensembles." % self._resampling_strategy)
-
- elif self._resampling_strategy in [
- 'partial-cv',
- 'cv',
- 'cv-iterative-fit',
- 'partial-cv-iterative-fit',
- ] and 'folds' not in self._resampling_strategy_arguments:
- self._resampling_strategy_arguments['folds'] = 5
+ elif (
+ self._resampling_strategy
+ in [
+ "partial-cv",
+ "partial-cv-iterative-fit",
+ ]
+ and self._ensemble_class is not None
+ ):
+ raise ValueError(
+ "Resampling strategy %s cannot be used "
+ "together with ensembles." % self._resampling_strategy
+ )
+
+ elif (
+ self._resampling_strategy
+ in [
+ "partial-cv",
+ "cv",
+ "cv-iterative-fit",
+ "partial-cv-iterative-fit",
+ ]
+ and "folds" not in self._resampling_strategy_arguments
+ ):
+ self._resampling_strategy_arguments["folds"] = 5
return
- @staticmethod
- def subsample_if_too_large(
+ def refit(
+ self,
X: SUPPORTED_FEAT_TYPES,
y: SUPPORTED_TARGET_TYPES,
- logger,
- seed: int,
- memory_limit: int,
- task: int,
- ):
- if memory_limit and isinstance(X, np.ndarray):
-
- if X.dtype == np.float32:
- multiplier = 4
- elif X.dtype in (np.float64, float):
- multiplier = 8
- elif (
- # In spite of the names, np.float96 and np.float128
- # provide only as much precision as np.longdouble,
- # that is, 80 bits on most x86 machines and 64 bits
- # in standard Windows builds.
- (hasattr(np, 'float128') and X.dtype == np.float128)
- or (hasattr(np, 'float96') and X.dtype == np.float96)
- ):
- multiplier = 16
- else:
- # Just assuming some value - very unlikely
- multiplier = 8
- logger.warning('Unknown dtype for X: %s, assuming it takes 8 bit/number',
- str(X.dtype))
-
- megabytes = X.shape[0] * X.shape[1] * multiplier / 1024 / 1024
- if memory_limit <= megabytes * 10 and X.dtype != np.float32:
- cast_to = {
- 8: np.float32,
- 16: np.float64,
- }.get(multiplier, np.float32)
- logger.warning(
- 'Dataset too large for memory limit %dMB, reducing the precision from %s to %s',
- memory_limit,
- X.dtype,
- cast_to,
- )
- X = X.astype(cast_to)
+ max_reshuffles: int = 10,
+ ) -> AutoML:
+ """Refit the models to a new given set of data
- megabytes = X.shape[0] * X.shape[1] * multiplier / 1024 / 1024
- if memory_limit <= megabytes * 10:
- new_num_samples = int(
- memory_limit / (10 * X.shape[1] * multiplier / 1024 / 1024)
- )
- logger.warning(
- 'Dataset too large for memory limit %dMB, reducing number of samples from '
- '%d to %d.',
- memory_limit,
- X.shape[0],
- new_num_samples,
- )
- if task in CLASSIFICATION_TASKS:
- # Identify if it has unique labels and allow for
- # stratification, with unique labels in training set
- values, idxs, counts = np.unique(y, axis=0,
- return_index=True,
- return_counts=True)
- unique_labels = {
- idx: value
- for value, idx, count in zip(values, idxs, counts)
- if count == 1
- }
-
- # If there are unique labeled elements, remove them and
- # place them back in later
- if len(unique_labels) > 0:
- idxs_of_unique = np.asarray(list(unique_labels.keys()))
- unique_X = X[idxs_of_unique]
- unique_y = y[idxs_of_unique]
-
- # NOTE optimization
- # If this ever turns out to be slow, this actually
- # copies the entire array. There might be a better
- # solution but it will probably require a lot more
- # manual work in how splitting is done.
- X = np.delete(X, idxs_of_unique, axis=0)
- y = np.delete(y, idxs_of_unique, axis=0)
-
- X, _, y, _ = sklearn.model_selection.train_test_split(
- X, y,
- train_size=new_num_samples - len(unique_y),
- random_state=seed,
- stratify=y,
- )
+ Parameters
+ ----------
+ X : SUPPORTED_FEAT_TYPES
+ The data to dit to
- X = np.append(X, unique_X, axis=0)
- y = np.append(y, unique_y, axis=0)
+ y : SUPPORTED_TARGET_TYPES
+ The targets to fit to
- # Otherwise we should be able to stratify as normal
- else:
- X, _, y, _ = sklearn.model_selection.train_test_split(
- X, y,
- train_size=new_num_samples,
- random_state=seed,
- stratify=y,
- )
- elif task in REGRESSION_TASKS:
- X, _, y, _ = sklearn.model_selection.train_test_split(
- X, y,
- train_size=new_num_samples,
- random_state=seed,
- )
- else:
- raise ValueError(task)
- return X, y
+ max_reshuffles : int = 10
+ How many times to try reshuffle the data. If fitting fails, shuffle the
+ data. This can alleviate the problem in algorithms that depend on the
+ ordering of the data.
- def refit(self, X, y):
- # AutoSklearn does not handle sparse y for now
- y = convert_if_sparse(y)
+ Returns
+ -------
+ AutoML
+ Self
+ """
+ check_is_fitted(self)
+ y = convert_if_sparse(y) # AutoSklearn does not handle sparse y for now
# Make sure input data is valid
- if self.InputValidator is None or not self.InputValidator._is_fitted:
- raise ValueError("refit() is only supported after calling fit. Kindly call first "
- "the estimator fit() method.")
X, y = self.InputValidator.transform(X, y)
if self.models_ is None or len(self.models_) == 0 or self.ensemble_ is None:
self._load_models()
- # Refit is not applicable when ensemble_size is set to zero.
+ # Refit is not applicable when no ensemble class is provided
if self.ensemble_ is None:
- raise ValueError("Refit can only be called if 'ensemble_size != 0'")
+ raise ValueError("Refit can only be called if an ensemble class is given")
random_state = check_random_state(self._seed)
- for identifier in self.models_:
- model = self.models_[identifier]
- # this updates the model inplace, it can then later be used in
- # predict method
-
- # try to fit the model. If it fails, shuffle the data. This
- # could alleviate the problem in algorithms that depend on
- # the ordering of the data.
- for i in range(10):
+
+ for identifier, model in self.models_.items():
+ for i in range(max_reshuffles):
try:
if self._budget_type is None:
_fit_and_suppress_warnings(self._logger, model, X, y)
@@ -1198,7 +1241,7 @@ def refit(self, X, y):
X = X[indices]
y = y[indices]
- if i == 9:
+ if i == (max_reshuffles - 1):
raise e
self._can_predict = True
@@ -1207,17 +1250,17 @@ def refit(self, X, y):
def fit_pipeline(
self,
X: SUPPORTED_FEAT_TYPES,
- y: Union[SUPPORTED_TARGET_TYPES, spmatrix],
+ y: SUPPORTED_TARGET_TYPES | spmatrix,
is_classification: bool,
- config: Union[Configuration, Dict[str, Union[str, float, int]]],
+ config: Configuration | dict[str, str | float | int],
task: Optional[int] = None,
dataset_name: Optional[str] = None,
X_test: Optional[SUPPORTED_FEAT_TYPES] = None,
- y_test: Optional[Union[SUPPORTED_TARGET_TYPES, spmatrix]] = None,
- feat_type: Optional[List[str]] = None,
- **kwargs: Dict,
+ y_test: Optional[SUPPORTED_TARGET_TYPES | spmatrix] = None,
+ feat_type: Optional[list[str]] = None,
+ **kwargs: dict,
) -> Tuple[Optional[BasePipeline], RunInfo, RunValue]:
- """ Fits and individual pipeline configuration and returns
+ """Fits and individual pipeline configuration and returns
the result to the user.
The Estimator constraints are honored, for example the resampling
@@ -1236,15 +1279,15 @@ def fit_pipeline(
If provided, the testing performance will be tracked on this features.
y_test: array-like
If provided, the testing performance will be tracked on this labels
- config: Union[Configuration, Dict[str, Union[str, float, int]]]
- A configuration object used to define the pipeline steps. If a dictionary is passed,
- a configuration is created based on this dictionary.
+ config: Configuration | dict[str, str | float | int]
+ A configuration object used to define the pipeline steps.
+ If a dict is passed, a configuration is created based on this dict.
dataset_name: Optional[str]
A string to tag and identify the Auto-Sklearn run
is_classification: bool
Whether the task is for classification or regression. This affects
how the targets are treated
- feat_type : list, optional (default=None)
+ feat_type : List, optional (default=None)
List of str of `len(X.shape[1])` describing the attribute type.
Possible types are `Categorical` and `Numerical`. `Categorical`
attributes will be automatically One-Hot encoded. The values
@@ -1269,27 +1312,32 @@ def fit_pipeline(
if task is None:
y_task = type_of_target(y)
if not self._supports_task_type(y_task):
- raise ValueError(f"{self.__class__.__name__} does not support"
- f" task {y_task}")
+ raise ValueError(
+ f"{self.__class__.__name__} does not support" f" task {y_task}"
+ )
self._task = self._task_type_id(y_task)
else:
self._task = task
# Assign a metric if it doesnt exist
- if self._metric is None:
- self._metric = default_metric_for_task[self._task]
+ if self._metrics is None:
+ self._metrics = [default_metric_for_task[self._task]]
# Get the configuration space
# This also ensures that the Backend has processed the
# dataset
if self.configuration_space is None:
self.configuration_space = self.fit(
- X=X, y=y,
- dataset_name=dataset_name if dataset_name is not None else self._dataset_name,
+ X=X,
+ y=y,
+ dataset_name=dataset_name
+ if dataset_name is not None
+ else self._dataset_name,
X_test=X_test,
y_test=y_test,
feat_type=feat_type,
- only_return_configuration_space=True)
+ only_return_configuration_space=True,
+ )
# We do not want to overwrite existing runs
self.num_run += 1
@@ -1298,25 +1346,35 @@ def fit_pipeline(
config.config_id = self.num_run
# Prepare missing components to the TAE function call
- if 'include' not in kwargs:
- kwargs['include'] = self._include
- if 'exclude' not in kwargs:
- kwargs['exclude'] = self._exclude
- if 'memory_limit' not in kwargs:
- kwargs['memory_limit'] = self._memory_limit
- if 'resampling_strategy' not in kwargs:
- kwargs['resampling_strategy'] = self._resampling_strategy
- if 'metric' not in kwargs:
- kwargs['metric'] = self._metric
- if 'disable_file_output' not in kwargs:
- kwargs['disable_file_output'] = self._disable_evaluator_output
- if 'pynisher_context' not in kwargs:
- kwargs['pynisher_context'] = self._multiprocessing_context
- if 'stats' not in kwargs:
+ if "include" not in kwargs:
+ kwargs["include"] = self._include
+ if "exclude" not in kwargs:
+ kwargs["exclude"] = self._exclude
+ if "memory_limit" not in kwargs:
+ kwargs["memory_limit"] = self._memory_limit
+ if "resampling_strategy" not in kwargs:
+ kwargs["resampling_strategy"] = self._resampling_strategy
+ if "metrics" not in kwargs:
+ if "metric" in kwargs:
+ kwargs["metrics"] = kwargs["metric"]
+ del kwargs["metric"]
+ else:
+ kwargs["metrics"] = self._metrics
+ if not isinstance(kwargs["metrics"], Sequence):
+ kwargs["metrics"] = [kwargs["metrics"]]
+ if "scoring_functions" not in kwargs:
+ kwargs["scoring_functions"] = self._scoring_functions
+ if "disable_file_output" not in kwargs:
+ kwargs["disable_file_output"] = self._disable_evaluator_output
+ if "pynisher_context" not in kwargs:
+ kwargs["pynisher_context"] = self._multiprocessing_context
+ if "stats" not in kwargs:
scenario_mock = unittest.mock.Mock()
scenario_mock.wallclock_limit = self._time_for_task
- kwargs['stats'] = Stats(scenario_mock)
- kwargs['stats'].start_timing()
+ kwargs["stats"] = Stats(scenario_mock)
+ kwargs["stats"].start_timing()
+
+ _validate_metrics(kwargs["metrics"], kwargs["scoring_functions"])
# Fit a pipeline, which will be stored on disk
# which we can later load via the backend
@@ -1324,10 +1382,11 @@ def fit_pipeline(
backend=self._backend,
autosklearn_seed=self._seed,
abort_on_first_run_crash=False,
- cost_for_crash=get_cost_of_crash(kwargs['metric']),
+ multi_objectives=["cost"],
+ cost_for_crash=get_cost_of_crash(kwargs["metrics"]),
port=self._logger_port,
**kwargs,
- **self._resampling_strategy_arguments
+ **self._resampling_strategy_arguments,
)
run_info, run_value = ta.run_wrapper(
@@ -1336,16 +1395,16 @@ def fit_pipeline(
instance=None,
instance_specific=None,
seed=self._seed,
- cutoff=kwargs.pop('cutoff', self._per_run_time_limit),
+ cutoff=kwargs.pop("cutoff", self._per_run_time_limit),
capped=False,
)
)
pipeline = None
- if kwargs['disable_file_output'] or kwargs['resampling_strategy'] == 'test':
+ if kwargs["disable_file_output"] or kwargs["resampling_strategy"] == "test":
self._logger.warning("File output is disabled. No pipeline can returned")
elif run_value.status == StatusType.SUCCESS:
- if kwargs['resampling_strategy'] in ('cv', 'cv-iterative-fit'):
+ if kwargs["resampling_strategy"] in ("cv", "cv-iterative-fit"):
load_function = self._backend.load_cv_model_by_seed_and_id_and_budget
else:
load_function = self._backend.load_model_by_seed_and_id_and_budget
@@ -1375,35 +1434,47 @@ def predict(self, X, batch_size=None, n_jobs=1):
Parallelize the predictions across the models with n_jobs
processes.
"""
+ check_is_fitted(self)
+
if (
- self._resampling_strategy not in (
- 'holdout', 'holdout-iterative-fit', 'cv', 'cv-iterative-fit')
+ self._resampling_strategy
+ not in ("holdout", "holdout-iterative-fit", "cv", "cv-iterative-fit")
and not self._can_predict
):
raise NotImplementedError(
- 'Predict is currently not implemented for resampling '
- 'strategy %s, please call refit().' % self._resampling_strategy)
+ "Predict is currently not implemented for resampling "
+ f"strategy {self._resampling_strategy}, please call refit()."
+ )
+ elif self._disable_evaluator_output is not False:
+ raise NotImplementedError(
+ "Predict cannot be called when evaluator output is disabled."
+ )
if self.models_ is None or len(self.models_) == 0 or self.ensemble_ is None:
self._load_models()
- # If self.ensemble_ is None, it means that ensemble_size is set to zero.
+ # If self.ensemble_ is None, it means that ensemble_class is None.
# In such cases, raise error because predict and predict_proba cannot
# be called.
if self.ensemble_ is None:
- raise ValueError("Predict and predict_proba can only be called "
- "if 'ensemble_size != 0'")
+ raise ValueError(
+ "Predict and predict_proba can only be called "
+ "if ensemble class is given."
+ )
# Make sure that input is valid
if self.InputValidator is None or not self.InputValidator._is_fitted:
- raise ValueError("predict() can only be called after performing fit(). Kindly call "
- "the estimator fit() method first.")
+ raise ValueError(
+ "predict() can only be called after performing fit(). Kindly call "
+ "the estimator fit() method first."
+ )
X = self.InputValidator.feature_validator.transform(X)
# Parallelize predictions across models with n_jobs processes.
# Each process computes predictions in chunks of batch_size rows.
try:
for i, tmp_model in enumerate(self.models_.values()):
+ # TODO, modify this
if isinstance(tmp_model, (DummyRegressor, DummyClassifier)):
check_is_fitted(tmp_model)
else:
@@ -1419,24 +1490,26 @@ def predict(self, X, batch_size=None, n_jobs=1):
check_is_fitted(list(self.cv_models_.values())[0])
models = self.cv_models_
except sklearn.exceptions.NotFittedError:
- raise ValueError('Found no fitted models!')
+ raise ValueError("Found no fitted models!")
all_predictions = joblib.Parallel(n_jobs=n_jobs)(
joblib.delayed(_model_predict)(
- model=models[identifier],
- X=X,
- task=self._task,
- batch_size=batch_size
+ model=models[identifier], X=X, task=self._task, batch_size=batch_size
)
for identifier in self.ensemble_.get_selected_model_identifiers()
)
if len(all_predictions) == 0:
- raise ValueError('Something went wrong generating the predictions. '
- 'The ensemble should consist of the following '
- 'models: %s, the following models were loaded: '
- '%s' % (str(list(self.ensemble_indices_.keys())),
- str(list(self.models_.keys()))))
+ raise ValueError(
+ "Something went wrong generating the predictions. "
+ "The ensemble should consist of the following "
+ "models: %s, the following models were loaded: "
+ "%s"
+ % (
+ str(list(self.ensemble_indices_.keys())),
+ str(list(self.models_.keys())),
+ )
+ )
predictions = self.ensemble_.predict(all_predictions)
@@ -1448,23 +1521,38 @@ def predict(self, X, batch_size=None, n_jobs=1):
return predictions
- def fit_ensemble(self, y, task=None, precision=32,
- dataset_name=None, ensemble_nbest=None,
- ensemble_size=None):
+ def fit_ensemble(
+ self,
+ y: SUPPORTED_TARGET_TYPES,
+ task: Optional[int] = None,
+ precision: Literal[16, 32, 64] = 32,
+ dataset_name: Optional[str] = None,
+ ensemble_nbest: Optional[int] = None,
+ ensemble_class: Optional[AbstractEnsemble] = EnsembleSelection,
+ ensemble_kwargs: Optional[Dict[str, Any]] = None,
+ metrics: Scorer | Sequence[Scorer] | None = None,
+ ):
+ check_is_fitted(self)
+
+ if ensemble_class is None and self._ensemble_class is None:
+ raise ValueError(
+ "Please pass `ensemble_class` either to `fit_ensemble()` "
+ "or the constructor."
+ )
+
# AutoSklearn does not handle sparse y for now
y = convert_if_sparse(y)
- if self._resampling_strategy in ['partial-cv', 'partial-cv-iterative-fit']:
- raise ValueError('Cannot call fit_ensemble with resampling '
- 'strategy %s.' % self._resampling_strategy)
+ if self._resampling_strategy in ["partial-cv", "partial-cv-iterative-fit"]:
+ raise ValueError(
+ "Cannot call fit_ensemble with resampling "
+ "strategy %s." % self._resampling_strategy
+ )
if self._logger is None:
self._logger = self._get_logger(dataset_name)
# Make sure that input is valid
- if self.InputValidator is None or not self.InputValidator._is_fitted:
- raise ValueError("fit_ensemble() can only be called after fit. Please call the "
- "estimator fit() method prior to fit_ensemble().")
y = self.InputValidator.target_validator.transform(y)
# Create a client if needed
@@ -1473,6 +1561,10 @@ def fit_ensemble(self, y, task=None, precision=32,
else:
self._is_dask_client_internally_created = False
+ metrics = metrics if metrics is not None else self._metrics
+ if not isinstance(metrics, Sequence):
+ metrics = [metrics]
+
# Use the current thread to start the ensemble builder process
# The function ensemble_builder_process will internally create a ensemble
# builder in the provide dask client
@@ -1482,15 +1574,22 @@ def fit_ensemble(self, y, task=None, precision=32,
backend=copy.deepcopy(self._backend),
dataset_name=dataset_name if dataset_name else self._dataset_name,
task=task if task else self._task,
- metric=self._metric,
- ensemble_size=ensemble_size if ensemble_size else self._ensemble_size,
+ metrics=metrics if metrics is not None else self._metrics,
+ ensemble_class=(
+ ensemble_class if ensemble_class is not None else self._ensemble_class
+ ),
+ ensemble_kwargs=(
+ ensemble_kwargs
+ if ensemble_kwargs is not None
+ else self._ensemble_kwargs
+ ),
ensemble_nbest=ensemble_nbest if ensemble_nbest else self._ensemble_nbest,
max_models_on_disc=self._max_models_on_disc,
seed=self._seed,
precision=precision if precision else self.precision,
max_iterations=1,
- read_at_most=np.inf,
- ensemble_memory_limit=self._memory_limit,
+ read_at_most=None,
+ memory_limit=self._memory_limit,
random_state=self._seed,
logger_port=self._logger_port,
pynisher_context=self._multiprocessing_context,
@@ -1499,52 +1598,60 @@ def fit_ensemble(self, y, task=None, precision=32,
future = manager.futures.pop()
result = future.result()
if result is None:
- raise ValueError("Error building the ensemble - please check the log file and command "
- "line output for error messages.")
- self.ensemble_performance_history, _, _, _, _ = result
+ raise ValueError(
+ "Error building the ensemble - please check the log file and command "
+ "line output for error messages."
+ )
+ self.ensemble_performance_history, _ = result
+ self._ensemble_class = ensemble_class
self._load_models()
self._close_dask_client()
return self
def _load_models(self):
- self.ensemble_ = self._backend.load_ensemble(self._seed)
-
- # If no ensemble is loaded, try to get the best performing model
- if not self.ensemble_:
+ if self._ensemble_class is not None:
+ self.ensemble_ = self._backend.load_ensemble(self._seed)
+ else:
+ self.ensemble_ = None
+
+ # If no ensemble is loaded, try to get the best performing model.
+ # This is triggered if
+ # 1. self._ensemble_class is None (see if-statement above)
+ # 2. if the ensemble builder crashed and no ensemble is available
+ # 3. if the ensemble cannot be built because of arguments passed
+ # by the user (disable_evaluator_output and
+ # resampling_strategy)
+ if (
+ not self.ensemble_
+ and not (
+ self._disable_evaluator_output is True
+ or (
+ isinstance(self._disable_evaluator_output, list)
+ and "model" in self._disable_evaluator_output
+ )
+ )
+ and self._resampling_strategy
+ not in (
+ "partial-cv",
+ "partial-cv-iterative-fit",
+ )
+ ):
self.ensemble_ = self._load_best_individual_model()
if self.ensemble_:
identifiers = self.ensemble_.get_selected_model_identifiers()
self.models_ = self._backend.load_models_by_identifiers(identifiers)
- if self._resampling_strategy in ('cv', 'cv-iterative-fit'):
- self.cv_models_ = self._backend.load_cv_models_by_identifiers(identifiers)
+
+ if self._resampling_strategy in ("cv", "cv-iterative-fit"):
+ self.cv_models_ = self._backend.load_cv_models_by_identifiers(
+ identifiers
+ )
else:
self.cv_models_ = None
- if (
- len(self.models_) == 0 and
- self._resampling_strategy not in ['partial-cv', 'partial-cv-iterative-fit']
- ):
- raise ValueError('No models fitted!')
- if (
- self._resampling_strategy in ['cv', 'cv-iterative-fit']
- and len(self.cv_models_) == 0
- ):
- raise ValueError('No models fitted!')
-
- elif self._disable_evaluator_output is False or \
- (isinstance(self._disable_evaluator_output, list) and
- 'model' not in self._disable_evaluator_output):
- model_names = self._backend.list_all_models(self._seed)
-
- if len(model_names) == 0 and self._resampling_strategy not in \
- ['partial-cv', 'partial-cv-iterative-fit']:
- raise ValueError('No models fitted!')
-
- self.models_ = []
-
else:
self.models_ = []
+ self.cv_models_ = []
def _load_best_individual_model(self):
"""
@@ -1553,29 +1660,134 @@ def _load_best_individual_model(self):
by AutoML.
This is a robust mechanism to be able to predict,
even though no ensemble was found by ensemble builder.
+ It is also used to load the single best model in case
+ the user does not want to build an ensemble.
"""
-
# We also require that the model is fit and a task is defined
- # The ensemble size must also be greater than 1, else it means
- # that the user intentionally does not want an ensemble
- if not self._task or self._ensemble_size < 1:
+ if not self._task:
return None
# SingleBest contains the best model found by AutoML
ensemble = SingleBest(
- metric=self._metric,
+ metrics=self._metrics,
+ task_type=self._task,
seed=self._seed,
run_history=self.runhistory_,
backend=self._backend,
+ random_state=self._seed,
)
self._logger.warning(
"No valid ensemble was created. Please check the log"
"file for errors. Default to the best individual estimator:{}".format(
- ensemble.identifiers_
+ ensemble.get_identifiers_with_weights()[0][0]
)
)
return ensemble
+ def _load_pareto_set(self) -> Sequence[VotingClassifier | VotingRegressor]:
+ if self.ensemble_ is None:
+ self.ensemble_ = self._backend.load_ensemble(self._seed)
+
+ # If no ensemble is loaded we cannot do anything
+ if not self.ensemble_:
+ raise ValueError("Pareto set only available if ensemble can be loaded.")
+
+ if isinstance(self.ensemble_, AbstractMultiObjectiveEnsemble):
+ pareto_set = self.ensemble_.get_pareto_set()
+ else:
+ self._logger.warning(
+ "Pareto set not available for single objective ensemble "
+ "method. The Pareto set will only include the single ensemble "
+ "constructed by %s",
+ type(self.ensemble_),
+ )
+ pareto_set = [self.ensemble_]
+
+ ensembles = []
+ for ensemble in pareto_set:
+ identifiers = ensemble.get_selected_model_identifiers()
+ weights = {
+ identifier: weight
+ for identifier, weight in ensemble.get_identifiers_with_weights()
+ }
+
+ if self._task in CLASSIFICATION_TASKS:
+ voter = VotingClassifier(
+ estimators=None,
+ voting="soft",
+ )
+ kind = "classifier"
+ else:
+ voter = VotingRegressor(estimators=None)
+ kind = "regeressor"
+
+ if self._resampling_strategy in ("cv", "cv-iterative-fit"):
+ models = self._backend.load_cv_models_by_identifiers(identifiers)
+ else:
+ models = self._backend.load_models_by_identifiers(identifiers)
+
+ if len(models) == 0:
+ raise ValueError("No models fitted!")
+
+ weight_vector = []
+ estimators = []
+ for identifier in identifiers:
+ estimator = models[identifier]
+ weight = weights[identifier]
+
+ # Kind of hacky, really the dummy models should
+ # act like everything else does. Doing this is
+ # required so that the VotingClassifier/Regressor
+ # can use it as intended
+ if not isinstance(estimator, Pipeline):
+ if kind == "classifier":
+ steps = [
+ ("data_preprocessor", None),
+ ("balancing", None),
+ ("feature_preprocessor", None),
+ (kind, estimator),
+ ]
+ else:
+ steps = [
+ ("data_preprocessor", None),
+ ("feature_preprocessor", None),
+ (kind, estimator),
+ ]
+
+ estimator = Pipeline(steps=steps)
+
+ weight_vector.append(weight)
+ estimators.append(estimator)
+
+ voter.estimators = estimators
+ voter.estimators_ = estimators
+ voter.weights = weight_vector
+
+ if self._task in CLASSIFICATION_TASKS:
+ # Scikit-learn would raise a shape error here which we
+ # have to work around.
+
+ def inverse_transform(self, y):
+ if len(y.shape) == 1:
+ y = y.reshape((-1, 1))
+ reshaped = True
+ else:
+ reshaped = False
+ y = self.old_inverse_transform(y)
+ if reshaped:
+ return y.flatten()
+ else:
+ return y
+
+ voter.le_ = copy.deepcopy(self.InputValidator.target_validator.encoder)
+ functype = types.MethodType
+ voter.le_.old_inverse_transform = voter.le_.inverse_transform
+ voter.le_.inverse_transform = functype(inverse_transform, voter.le_)
+
+ ensembles.append(voter)
+
+ return ensembles
+
def score(self, X, y):
# fix: Consider only index 1 of second dimension
# Don't know if the reshaping should be done there or in calculate_score
@@ -1585,12 +1797,8 @@ def score(self, X, y):
# The reason is we do not want to trigger the
# check for changing input types on successive
# input validator calls
+ check_is_fitted(self)
prediction = self.predict(X)
-
- # Make sure that input is valid
- if self.InputValidator is None or not self.InputValidator._is_fitted:
- raise ValueError("score() is only supported after calling fit. Kindly call first "
- "the estimator fit() method.")
y = self.InputValidator.target_validator.transform(y)
# Encode the prediction using the input validator
@@ -1601,13 +1809,15 @@ def score(self, X, y):
# same representation domain
prediction = self.InputValidator.target_validator.transform(prediction)
- return calculate_metric(solution=y,
- prediction=prediction,
- task_type=self._task,
- metric=self._metric, )
+ return compute_single_metric(
+ solution=y,
+ prediction=prediction,
+ task_type=self._task,
+ metric=self._metrics[0],
+ )
def _get_runhistory_models_performance(self):
- metric = self._metric
+ metric = self._metrics[0]
data = self.runhistory_.data
performance_list = []
for run_key, run_value in data.items():
@@ -1616,20 +1826,28 @@ def _get_runhistory_models_performance(self):
continue
# Alternatively, it is possible to also obtain the start time with
# ``run_value.starttime``
- endtime = pd.Timestamp(time.strftime('%Y-%m-%d %H:%M:%S',
- time.localtime(run_value.endtime)))
- val_score = metric._optimum - (metric._sign * run_value.cost)
- train_score = metric._optimum - (metric._sign * run_value.additional_info['train_loss'])
+ endtime = pd.Timestamp(
+ time.strftime("%Y-%m-%d %H:%M:%S", time.localtime(run_value.endtime))
+ )
+ cost = run_value.cost
+ train_loss = run_value.additional_info["train_loss"]
+ if len(self._metrics) > 1:
+ cost = cost[0]
+ train_loss = train_loss[0]
+ val_score = metric._optimum - (metric._sign * cost)
+ train_score = metric._optimum - (metric._sign * train_loss)
scores = {
- 'Timestamp': endtime,
- 'single_best_optimization_score': val_score,
- 'single_best_train_score': train_score,
+ "Timestamp": endtime,
+ "single_best_optimization_score": val_score,
+ "single_best_train_score": train_score,
}
# Append test-scores, if data for test_loss are available.
# This is the case, if X_test and y_test where provided.
- if 'test_loss' in run_value.additional_info:
- test_score = metric._optimum - (metric._sign *
- run_value.additional_info['test_loss'])
+ if "test_loss" in run_value.additional_info:
+ test_loss = run_value.additional_info["test_loss"]
+ if len(self._metrics) > 1:
+ test_loss = test_loss[0]
+ test_score = metric._optimum - (metric._sign * test_loss)
scores["single_best_test_score"] = test_score
performance_list.append(scores)
@@ -1637,42 +1855,54 @@ def _get_runhistory_models_performance(self):
@property
def performance_over_time_(self):
+ check_is_fitted(self)
individual_performance_frame = self._get_runhistory_models_performance()
- best_values = pd.Series({'single_best_optimization_score': -np.inf,
- 'single_best_test_score': -np.inf,
- 'single_best_train_score': -np.inf})
+ best_values = pd.Series(
+ {
+ "single_best_optimization_score": -np.inf,
+ "single_best_test_score": -np.inf,
+ "single_best_train_score": -np.inf,
+ }
+ )
for idx in individual_performance_frame.index:
if (
- individual_performance_frame.loc[idx, 'single_best_optimization_score']
- > best_values['single_best_optimization_score']
+ individual_performance_frame.loc[idx, "single_best_optimization_score"]
+ > best_values["single_best_optimization_score"]
):
best_values = individual_performance_frame.loc[idx]
individual_performance_frame.loc[idx] = best_values
performance_over_time = individual_performance_frame
- if self._ensemble_size != 0:
+ if self._ensemble_class is not None:
ensemble_performance_frame = pd.DataFrame(self.ensemble_performance_history)
- best_values = pd.Series({'ensemble_optimization_score': -np.inf,
- 'ensemble_test_score': -np.inf})
+ best_values = pd.Series(
+ {"ensemble_optimization_score": -np.inf, "ensemble_test_score": -np.inf}
+ )
for idx in ensemble_performance_frame.index:
if (
- ensemble_performance_frame.loc[idx, 'ensemble_optimization_score']
- > best_values['ensemble_optimization_score']
+ ensemble_performance_frame.loc[idx, "ensemble_optimization_score"]
+ > best_values["ensemble_optimization_score"]
):
best_values = ensemble_performance_frame.loc[idx]
ensemble_performance_frame.loc[idx] = best_values
- performance_over_time = pd.merge(
- ensemble_performance_frame,
- individual_performance_frame,
- on="Timestamp", how='outer'
- ).sort_values('Timestamp').fillna(method='ffill')
+ performance_over_time = (
+ pd.merge(
+ ensemble_performance_frame,
+ individual_performance_frame,
+ on="Timestamp",
+ how="outer",
+ )
+ .sort_values("Timestamp")
+ .fillna(method="ffill")
+ )
return performance_over_time
@property
def cv_results_(self):
+ check_is_fitted(self)
results = dict()
# Missing in contrast to scikit-learn
@@ -1690,8 +1920,8 @@ def cv_results_(self):
# TODO: add those arguments
# TODO remove this restriction!
- if self._resampling_strategy in ['partial-cv', 'partial-cv-iterative-fit']:
- raise ValueError('Cannot call cv_results when using partial-cv!')
+ if self._resampling_strategy in ["partial-cv", "partial-cv-iterative-fit"]:
+ raise ValueError("Cannot call cv_results when using partial-cv!")
parameter_dictionaries = dict()
masks = dict()
@@ -1706,14 +1936,11 @@ def cv_results_(self):
metric_mask = dict()
metric_dict = dict()
- metric_name = []
- for metric in self._scoring_functions:
- metric_name.append(metric.name)
+ for metric in itertools.chain(self._metrics, self._scoring_functions):
metric_dict[metric.name] = []
metric_mask[metric.name] = []
- mean_test_score = []
mean_fit_time = []
params = []
status = []
@@ -1726,19 +1953,19 @@ def cv_results_(self):
s = run_value.status
if s == StatusType.SUCCESS:
- status.append('Success')
+ status.append("Success")
elif s == StatusType.DONOTADVANCE:
- status.append('Success (but do not advance to higher budget)')
+ status.append("Success (but do not advance to higher budget)")
elif s == StatusType.TIMEOUT:
- status.append('Timeout')
+ status.append("Timeout")
elif s == StatusType.CRASHED:
- status.append('Crash')
+ status.append("Crash")
elif s == StatusType.ABORT:
- status.append('Abort')
+ status.append("Abort")
elif s == StatusType.MEMOUT:
- status.append('Memout')
- # TODO remove StatusType.RUNNING at some point in the future when the new SMAC 0.13.2
- # is the new minimum required version!
+ status.append("Memout")
+ # TODO remove StatusType.RUNNING at some point in the future when the new
+ # SMAC 0.13.2 is the new minimum required version!
elif s in (StatusType.STOP, StatusType.RUNNING):
continue
else:
@@ -1746,7 +1973,7 @@ def cv_results_(self):
param_dict = config.get_dictionary()
params.append(param_dict)
- mean_test_score.append(self._metric._optimum - (self._metric._sign * run_value.cost))
+
mean_fit_time.append(run_value.time)
budgets.append(run_key.budget)
@@ -1761,7 +1988,18 @@ def cv_results_(self):
parameter_dictionaries[hp_name].append(hp_value)
masks[hp_name].append(mask_value)
+ cost = [run_value.cost] if len(self._metrics) == 1 else run_value.cost
+ for metric_idx, metric in enumerate(self._metrics):
+ metric_cost = cost[metric_idx]
+ metric_value = metric._optimum - (metric._sign * metric_cost)
+ mask_value = False
+ metric_dict[metric.name].append(metric_value)
+ metric_mask[metric.name].append(mask_value)
+
+ optimization_metric_names = set(m.name for m in self._metrics)
for metric in self._scoring_functions:
+ if metric.name in optimization_metric_names:
+ continue
if metric.name in run_value.additional_info.keys():
metric_cost = run_value.additional_info[metric.name]
metric_value = metric._optimum - (metric._sign * metric_cost)
@@ -1772,113 +2010,297 @@ def cv_results_(self):
metric_dict[metric.name].append(metric_value)
metric_mask[metric.name].append(mask_value)
- results['mean_test_score'] = np.array(mean_test_score)
- for name in metric_name:
- masked_array = ma.MaskedArray(metric_dict[name],
- metric_mask[name])
- results['metric_%s' % name] = masked_array
+ if len(self._metrics) == 1:
+ results["mean_test_score"] = np.array(metric_dict[self._metrics[0].name])
+ rank_order = -1 * self._metrics[0]._sign * results["mean_test_score"]
+ results["rank_test_scores"] = scipy.stats.rankdata(rank_order, method="min")
+ else:
+ for metric in self._metrics:
+ key = f"mean_test_{metric.name}"
+ results[key] = np.array(metric_dict[metric.name])
+ rank_order = -1 * metric._sign * results[key]
+ results[f"rank_test_{metric.name}"] = scipy.stats.rankdata(
+ rank_order, method="min"
+ )
+ for metric in self._scoring_functions:
+ masked_array = ma.MaskedArray(
+ metric_dict[metric.name], metric_mask[metric.name]
+ )
+ results[f"metric_{metric.name}"] = masked_array
- results['mean_fit_time'] = np.array(mean_fit_time)
- results['params'] = params
- rank_order = -1 * self._metric._sign * results['mean_test_score']
- results['rank_test_scores'] = scipy.stats.rankdata(rank_order, method='min')
- results['status'] = status
- results['budgets'] = budgets
+ results["mean_fit_time"] = np.array(mean_fit_time)
+ results["params"] = params
+ results["status"] = status
+ results["budgets"] = budgets
for hp_name in hp_names:
- masked_array = ma.MaskedArray(parameter_dictionaries[hp_name],
- masks[hp_name])
- results['param_%s' % hp_name] = masked_array
+ masked_array = ma.MaskedArray(
+ parameter_dictionaries[hp_name], masks[hp_name]
+ )
+ results["param_%s" % hp_name] = masked_array
return results
- def sprint_statistics(self):
+ def sprint_statistics(self) -> str:
+ check_is_fitted(self)
cv_results = self.cv_results_
sio = io.StringIO()
- sio.write('auto-sklearn results:\n')
- sio.write(' Dataset name: %s\n' % self._dataset_name)
- sio.write(' Metric: %s\n' % self._metric)
- idx_success = np.where(np.array(
- [status in ['Success', 'Success (but do not advance to higher budget)']
- for status in cv_results['status']]
- ))[0]
+ sio.write("auto-sklearn results:\n")
+ sio.write(" Dataset name: %s\n" % self._dataset_name)
+ if len(self._metrics) == 1:
+ sio.write(" Metric: %s\n" % self._metrics[0])
+ else:
+ sio.write(" Metrics: %s\n" % self._metrics)
+ idx_success = np.where(
+ np.array(
+ [
+ status
+ in ["Success", "Success (but do not advance to higher budget)"]
+ for status in cv_results["status"]
+ ]
+ )
+ )[0]
if len(idx_success) > 0:
- if not self._metric._optimum:
- idx_best_run = np.argmin(cv_results['mean_test_score'][idx_success])
+ key = (
+ "mean_test_score"
+ if len(self._metrics) == 1
+ else f"mean_test_" f"{self._metrics[0].name}"
+ )
+
+ if not self._metrics[0]._optimum:
+ idx_best_run = np.argmin(cv_results[key][idx_success])
else:
- idx_best_run = np.argmax(cv_results['mean_test_score'][idx_success])
- best_score = cv_results['mean_test_score'][idx_success][idx_best_run]
- sio.write(' Best validation score: %f\n' % best_score)
- num_runs = len(cv_results['status'])
- sio.write(' Number of target algorithm runs: %d\n' % num_runs)
- num_success = sum([
- s in ['Success', 'Success (but do not advance to higher budget)']
- for s in cv_results['status']
- ])
- sio.write(' Number of successful target algorithm runs: %d\n' % num_success)
- num_crash = sum([s == 'Crash' for s in cv_results['status']])
- sio.write(' Number of crashed target algorithm runs: %d\n' % num_crash)
- num_timeout = sum([s == 'Timeout' for s in cv_results['status']])
- sio.write(' Number of target algorithms that exceeded the time '
- 'limit: %d\n' % num_timeout)
- num_memout = sum([s == 'Memout' for s in cv_results['status']])
- sio.write(' Number of target algorithms that exceeded the memory '
- 'limit: %d\n' % num_memout)
+ idx_best_run = np.argmax(cv_results[key][idx_success])
+ best_score = cv_results[key][idx_success][idx_best_run]
+ sio.write(" Best validation score: %f\n" % best_score)
+ num_runs = len(cv_results["status"])
+ sio.write(" Number of target algorithm runs: %d\n" % num_runs)
+ num_success = sum(
+ [
+ s in ["Success", "Success (but do not advance to higher budget)"]
+ for s in cv_results["status"]
+ ]
+ )
+ sio.write(" Number of successful target algorithm runs: %d\n" % num_success)
+ num_crash = sum([s == "Crash" for s in cv_results["status"]])
+ sio.write(" Number of crashed target algorithm runs: %d\n" % num_crash)
+ num_timeout = sum([s == "Timeout" for s in cv_results["status"]])
+ sio.write(
+ " Number of target algorithms that exceeded the time "
+ "limit: %d\n" % num_timeout
+ )
+ num_memout = sum([s == "Memout" for s in cv_results["status"]])
+ sio.write(
+ " Number of target algorithms that exceeded the memory "
+ "limit: %d\n" % num_memout
+ )
return sio.getvalue()
- def get_models_with_weights(self):
- if self.models_ is None or len(self.models_) == 0 or \
- self.ensemble_ is None:
+ def get_models_with_weights(self) -> list[Tuple[float, BasePipeline]]:
+ check_is_fitted(self)
+ if self.models_ is None or len(self.models_) == 0 or self.ensemble_ is None:
self._load_models()
return self.ensemble_.get_models_with_weights(self.models_)
- def show_models(self):
- models_with_weights = self.get_models_with_weights()
+ def show_models(self) -> dict[int, Any]:
+ """Returns a dictionary containing dictionaries of ensemble models.
+
+ Each model in the ensemble can be accessed by giving its ``model_id`` as key.
+
+ A model dictionary contains the following:
+
+ * ``"model_id"`` - The id given to a model by ``autosklearn``.
+ * ``"rank"`` - The rank of the model based on it's ``"cost"``.
+ * ``"cost"`` - The loss of the model on the validation set.
+ * ``"ensemble_weight"`` - The weight given to the model in the ensemble.
+ * ``"voting_model"`` - The ``cv_voting_ensemble`` model (for 'cv' resampling).
+ * ``"estimators"`` - List of models (dicts) in ``cv_voting_ensemble`` (for 'cv' resampling).
+ * ``"data_preprocessor"`` - The preprocessor used on the data.
+ * ``"balancing"`` - The balancing used on the data (for classification).
+ * ``"feature_preprocessor"`` - The preprocessor for features types.
+ * ``"classifier"`` or ``"regressor"`` - The autosklearn wrapped classifier or regressor.
+ * ``"sklearn_classifier"`` or ``"sklearn_regressor"`` - The sklearn classifier or regressor.
+
+ **Example**
+
+ .. code-block:: python
+
+ import sklearn.datasets
+ import autosklearn.regression
+
+ X, y = sklearn.datasets.load_diabetes(return_X_y=True)
+
+ automl = autosklearn.regression.AutoSklearnRegressor(
+ time_left_for_this_task=120
+ )
+ automl.fit(X_train, y_train, dataset_name='diabetes')
+
+ ensemble_dict = automl.show_models()
+ print(ensemble_dict)
+
+ Output:
+
+ .. code-block:: text
+
+ {
+ 25: {'model_id': 25.0,
+ 'rank': 1,
+ 'cost': 0.43667876507897496,
+ 'ensemble_weight': 0.38,
+ 'data_preprocessor': ,
+ 'feature_preprocessor': ,
+ 'regressor': ,
+ 'sklearn_regressor': SGDRegressor(alpha=0.0006517033225329654,...)
+ },
+ 6: {'model_id': 6.0,
+ 'rank': 2,
+ 'cost': 0.4550418898836528,
+ 'ensemble_weight': 0.3,
+ 'data_preprocessor': ,
+ 'feature_preprocessor': ,
+ 'regressor': ,
+ 'sklearn_regressor': ARDRegression(alpha_1=0.027,...)
+ },
+ ...
+ }
+
+ Returns
+ -------
+ dict[int, Any] : dictionary of length = number of models in the ensemble
+ A dictionary of models in the ensemble, where ``model_id`` is the key.
+ """ # noqa: E501
+ check_is_fitted(self)
+
+ ensemble_dict = {}
+
+ if self._ensemble_class is None:
+ warnings.warn(
+ "No models in the ensemble. Kindly provide an ensemble class."
+ )
+ return ensemble_dict
+
+ # check for condition when ensemble_size > 0 but there is no ensemble to load
+ if self.ensemble_ is None:
+ warnings.warn("No ensemble found. Returning empty dictionary.")
+ return ensemble_dict
+
+ def has_key(rv, key):
+ return rv.additional_info and key in rv.additional_info
+
+ table_dict = {}
+ for run_key, run_val in self.runhistory_.data.items():
+ if has_key(run_val, "num_run"):
+ model_id = run_val.additional_info["num_run"]
+ table_dict[model_id] = {"model_id": model_id, "cost": run_val.cost}
+
+ # Checking if the dictionary is empty
+ if not table_dict:
+ raise RuntimeError(
+ "No model found. Try increasing 'time_left_for_this_task'."
+ )
+
+ for (_, model_id, _), weight in self.ensemble_.get_identifiers_with_weights():
+ table_dict[model_id]["ensemble_weight"] = weight
+
+ table = pd.DataFrame.from_dict(table_dict, orient="index")
+ table.sort_values(by="cost", inplace=True)
+
+ # Check which resampling strategy is chosen and selecting the appropriate models
+ is_cv = self._resampling_strategy == "cv"
+ models = self.cv_models_ if is_cv else self.models_
+
+ rank = 1 # Initializing rank for the first model
+ for (_, model_id, _), model in models.items():
+ model_dict = {} # Declaring model dictionary
+
+ # Inserting model_id, rank, cost and ensemble weight
+ model_dict["model_id"] = table.loc[model_id]["model_id"].astype(int)
+ model_dict["rank"] = rank
+ model_dict["cost"] = table.loc[model_id]["cost"]
+ model_dict["ensemble_weight"] = table.loc[model_id]["ensemble_weight"]
+ rank += 1 # Incrementing rank by 1 for the next model
+
+ # The steps in the models pipeline are as follows:
+ # 'data_preprocessor': DataPreprocessor,
+ # 'balancing': Balancing,
+ # 'feature_preprocessor': FeaturePreprocessorChoice,
+ # 'classifier'/'regressor': ClassifierChoice/RegressorChoice (wrapped model)
+
+ # For 'cv' (cross validation) strategy
+ if is_cv:
+ # Voting model created by cross validation
+ cv_voting_ensemble = model
+ model_dict["voting_model"] = cv_voting_ensemble
+
+ # List of models, each trained on one cv fold
+ cv_models = []
+ for cv_model in cv_voting_ensemble.estimators_:
+ estimator = dict(cv_model.steps)
+
+ # Adding sklearn model to the model dictionary
+ model_type, autosklearn_wrapped_model = cv_model.steps[-1]
+ estimator[
+ f"sklearn_{model_type}"
+ ] = autosklearn_wrapped_model.choice.estimator
+ cv_models.append(estimator)
+ model_dict["estimators"] = cv_models
+
+ # For any other strategy
+ else:
+ steps = dict(model.steps)
+ model_dict.update(steps)
+
+ # Adding sklearn model to the model dictionary
+ model_type, autosklearn_wrapped_model = model.steps[-1]
+ model_dict[
+ f"sklearn_{model_type}"
+ ] = autosklearn_wrapped_model.choice.estimator
- with io.StringIO() as sio:
- sio.write("[")
- for weight, model in models_with_weights:
- sio.write("(%f, %s),\n" % (weight, model))
- sio.write("]")
+ # Insterting model_dict in the ensemble dictionary
+ ensemble_dict[model_id] = model_dict
- return sio.getvalue()
+ return ensemble_dict
- def _create_search_space(self, tmp_dir, backend, datamanager,
- include=None,
- exclude=None,
- ):
- task_name = 'CreateConfigSpace'
+ def has_ensemble(self) -> bool:
+ """
+ Returns
+ -------
+ bool
+ Whether this AutoML instance has an ensemble
+ """
+ return self.ensemble_ is not None
- self._stopwatch.start_task(task_name)
- configspace_path = os.path.join(tmp_dir, 'space.json')
+ def _create_search_space(
+ self,
+ tmp_dir: str,
+ backend: Backend,
+ datamanager: XYDataManager,
+ include: Optional[Mapping[str, list[str]]] = None,
+ exclude: Optional[Mapping[str, list[str]]] = None,
+ ) -> Tuple[ConfigurationSpace, str]:
+ configspace_path = os.path.join(tmp_dir, "space.json")
configuration_space = pipeline.get_configuration_space(
datamanager.info,
include=include,
exclude=exclude,
)
- configuration_space = self.configuration_space_created_hook(
- datamanager, configuration_space)
backend.write_txt_file(
configspace_path,
cs_json.write(configuration_space),
- 'Configuration space'
+ "Configuration space",
)
- self._stopwatch.stop_task(task_name)
return configuration_space, configspace_path
- def configuration_space_created_hook(self, datamanager, configuration_space):
- return configuration_space
-
- def __getstate__(self) -> Dict[str, Any]:
+ def __getstate__(self) -> dict[str, Any]:
# Cannot serialize a client!
self._dask_client = None
self.logging_server = None
self.stop_logging_server = None
return self.__dict__
- def __del__(self):
+ def __del__(self) -> None:
# Clean up the logger
self._clean_logger()
@@ -1888,9 +2310,9 @@ def __del__(self):
class AutoMLClassifier(AutoML):
_task_mapping = {
- 'multilabel-indicator': MULTILABEL_CLASSIFICATION,
- 'multiclass': MULTICLASS_CLASSIFICATION,
- 'binary': BINARY_CLASSIFICATION,
+ "multilabel-indicator": MULTILABEL_CLASSIFICATION,
+ "multiclass": MULTICLASS_CLASSIFICATION,
+ "binary": BINARY_CLASSIFICATION,
}
@classmethod
@@ -1904,16 +2326,17 @@ def _supports_task_type(cls, task_type: str) -> bool:
def fit(
self,
X: SUPPORTED_FEAT_TYPES,
- y: Union[SUPPORTED_TARGET_TYPES, spmatrix],
+ y: SUPPORTED_TARGET_TYPES | spmatrix,
X_test: Optional[SUPPORTED_FEAT_TYPES] = None,
- y_test: Optional[Union[SUPPORTED_TARGET_TYPES, spmatrix]] = None,
- feat_type: Optional[List[bool]] = None,
+ y_test: Optional[SUPPORTED_TARGET_TYPES | spmatrix] = None,
+ feat_type: Optional[list[str]] = None,
dataset_name: Optional[str] = None,
only_return_configuration_space: bool = False,
load_models: bool = True,
- ):
+ ) -> AutoMLClassifier:
return super().fit(
- X, y,
+ X,
+ y,
X_test=X_test,
y_test=y_test,
feat_type=feat_type,
@@ -1926,17 +2349,19 @@ def fit(
def fit_pipeline(
self,
X: SUPPORTED_FEAT_TYPES,
- y: Union[SUPPORTED_TARGET_TYPES, spmatrix],
- config: Union[Configuration, Dict[str, Union[str, float, int]]],
+ y: SUPPORTED_TARGET_TYPES | spmatrix,
+ config: Configuration | dict[str, str | float | int],
dataset_name: Optional[str] = None,
X_test: Optional[SUPPORTED_FEAT_TYPES] = None,
- y_test: Optional[Union[SUPPORTED_TARGET_TYPES, spmatrix]] = None,
- feat_type: Optional[List[str]] = None,
+ y_test: Optional[SUPPORTED_TARGET_TYPES | spmatrix] = None,
+ feat_type: Optional[list[str]] = None,
**kwargs,
) -> Tuple[Optional[BasePipeline], RunInfo, RunValue]:
return super().fit_pipeline(
- X=X, y=y,
- X_test=X_test, y_test=y_test,
+ X=X,
+ y=y,
+ X_test=X_test,
+ y_test=y_test,
dataset_name=dataset_name,
config=config,
is_classification=True,
@@ -1944,13 +2369,18 @@ def fit_pipeline(
**kwargs,
)
- def predict(self, X, batch_size=None, n_jobs=1):
- predicted_probabilities = super().predict(X, batch_size=batch_size,
- n_jobs=n_jobs)
+ def predict(
+ self,
+ X: SUPPORTED_FEAT_TYPES,
+ batch_size: Optional[int] = None,
+ n_jobs: int = 1,
+ ) -> np.ndarray:
+ check_is_fitted(self)
+
+ predicted_probabilities = super().predict(
+ X, batch_size=batch_size, n_jobs=n_jobs
+ )
- if self.InputValidator is None or not self.InputValidator._is_fitted:
- raise ValueError("predict() is only supported after calling fit. Kindly call first "
- "the estimator fit() method.")
if self.InputValidator.target_validator.is_single_column_target():
predicted_indexes = np.argmax(predicted_probabilities, axis=1)
else:
@@ -1958,16 +2388,21 @@ def predict(self, X, batch_size=None, n_jobs=1):
return self.InputValidator.target_validator.inverse_transform(predicted_indexes)
- def predict_proba(self, X, batch_size=None, n_jobs=1):
+ def predict_proba(
+ self,
+ X: SUPPORTED_FEAT_TYPES,
+ batch_size: Optional[int] = None,
+ n_jobs: int = 1,
+ ) -> np.ndarray:
return super().predict(X, batch_size=batch_size, n_jobs=n_jobs)
class AutoMLRegressor(AutoML):
_task_mapping = {
- 'continuous-multioutput': MULTIOUTPUT_REGRESSION,
- 'continuous': REGRESSION,
- 'multiclass': REGRESSION,
+ "continuous-multioutput": MULTIOUTPUT_REGRESSION,
+ "continuous": REGRESSION,
+ "multiclass": REGRESSION,
}
@classmethod
@@ -1981,16 +2416,17 @@ def _supports_task_type(cls, task_type: str) -> bool:
def fit(
self,
X: SUPPORTED_FEAT_TYPES,
- y: Union[SUPPORTED_TARGET_TYPES, spmatrix],
+ y: SUPPORTED_TARGET_TYPES | spmatrix,
X_test: Optional[SUPPORTED_FEAT_TYPES] = None,
- y_test: Optional[Union[SUPPORTED_TARGET_TYPES, spmatrix]] = None,
- feat_type: Optional[List[bool]] = None,
+ y_test: Optional[SUPPORTED_TARGET_TYPES | spmatrix] = None,
+ feat_type: Optional[list[str]] = None,
dataset_name: Optional[str] = None,
only_return_configuration_space: bool = False,
load_models: bool = True,
- ):
+ ) -> AutoMLRegressor:
return super().fit(
- X, y,
+ X,
+ y,
X_test=X_test,
y_test=y_test,
feat_type=feat_type,
@@ -2003,17 +2439,19 @@ def fit(
def fit_pipeline(
self,
X: SUPPORTED_FEAT_TYPES,
- y: Union[SUPPORTED_TARGET_TYPES, spmatrix],
- config: Union[Configuration, Dict[str, Union[str, float, int]]],
+ y: SUPPORTED_TARGET_TYPES | spmatrix,
+ config: Configuration | dict[str, str | float | int],
dataset_name: Optional[str] = None,
X_test: Optional[SUPPORTED_FEAT_TYPES] = None,
- y_test: Optional[Union[SUPPORTED_TARGET_TYPES, spmatrix]] = None,
- feat_type: Optional[List[str]] = None,
- **kwargs: Dict,
+ y_test: Optional[SUPPORTED_TARGET_TYPES | spmatrix] = None,
+ feat_type: Optional[list[str]] = None,
+ **kwargs: dict,
) -> Tuple[Optional[BasePipeline], RunInfo, RunValue]:
return super().fit_pipeline(
- X=X, y=y,
- X_test=X_test, y_test=y_test,
+ X=X,
+ y=y,
+ X_test=X_test,
+ y_test=y_test,
config=config,
feat_type=feat_type,
dataset_name=dataset_name,
diff --git a/autosklearn/automl_common b/autosklearn/automl_common
new file mode 160000
index 0000000000..c760e29b5f
--- /dev/null
+++ b/autosklearn/automl_common
@@ -0,0 +1 @@
+Subproject commit c760e29b5fbf90455d1a16aebbca6424468628c2
diff --git a/autosklearn/constants.py b/autosklearn/constants.py
index 60a025999e..2db372925c 100644
--- a/autosklearn/constants.py
+++ b/autosklearn/constants.py
@@ -7,21 +7,26 @@
MULTIOUTPUT_REGRESSION = 5
REGRESSION_TASKS = [REGRESSION, MULTIOUTPUT_REGRESSION]
-CLASSIFICATION_TASKS = [BINARY_CLASSIFICATION, MULTICLASS_CLASSIFICATION,
- MULTILABEL_CLASSIFICATION]
+CLASSIFICATION_TASKS = [
+ BINARY_CLASSIFICATION,
+ MULTICLASS_CLASSIFICATION,
+ MULTILABEL_CLASSIFICATION,
+]
TASK_TYPES = REGRESSION_TASKS + CLASSIFICATION_TASKS
-TASK_TYPES_TO_STRING = \
- {BINARY_CLASSIFICATION: 'binary.classification',
- MULTICLASS_CLASSIFICATION: 'multiclass.classification',
- MULTILABEL_CLASSIFICATION: 'multilabel.classification',
- REGRESSION: 'regression',
- MULTIOUTPUT_REGRESSION: 'multioutput.regression'}
+TASK_TYPES_TO_STRING = {
+ BINARY_CLASSIFICATION: "binary.classification",
+ MULTICLASS_CLASSIFICATION: "multiclass.classification",
+ MULTILABEL_CLASSIFICATION: "multilabel.classification",
+ REGRESSION: "regression",
+ MULTIOUTPUT_REGRESSION: "multioutput.regression",
+}
-STRING_TO_TASK_TYPES = \
- {'binary.classification': BINARY_CLASSIFICATION,
- 'multiclass.classification': MULTICLASS_CLASSIFICATION,
- 'multilabel.classification': MULTILABEL_CLASSIFICATION,
- 'regression': REGRESSION,
- 'multioutput.regression': MULTIOUTPUT_REGRESSION}
+STRING_TO_TASK_TYPES = {
+ "binary.classification": BINARY_CLASSIFICATION,
+ "multiclass.classification": MULTICLASS_CLASSIFICATION,
+ "multilabel.classification": MULTILABEL_CLASSIFICATION,
+ "regression": REGRESSION,
+ "multioutput.regression": MULTIOUTPUT_REGRESSION,
+}
diff --git a/autosklearn/data/abstract_data_manager.py b/autosklearn/data/abstract_data_manager.py
index 739e707334..0837d59ad0 100644
--- a/autosklearn/data/abstract_data_manager.py
+++ b/autosklearn/data/abstract_data_manager.py
@@ -2,14 +2,14 @@
from typing import Any, Dict, Union
import numpy as np
-
import scipy.sparse
-from autosklearn.pipeline.components.data_preprocessing.feature_type \
- import FeatTypeSplit
+from autosklearn.pipeline.components.data_preprocessing.feature_type import (
+ FeatTypeSplit,
+)
-class AbstractDataManager():
+class AbstractDataManager:
__metaclass__ = abc.ABCMeta
def __init__(self, name: str):
@@ -47,22 +47,26 @@ def encoder(self, value: FeatTypeSplit) -> FeatTypeSplit:
self._encoder = value
def __repr__(self) -> str:
- return 'DataManager : ' + self.name
+ return "DataManager : " + self.name
def __str__(self) -> str:
- val = 'DataManager : ' + self.name + '\ninfo:\n'
+ val = "DataManager : " + self.name + "\ninfo:\n"
for item in self.info:
- val = val + '\t' + item + ' = ' + str(self.info[item]) + '\n'
- val = val + 'data:\n'
+ val = val + "\t" + item + " = " + str(self.info[item]) + "\n"
+ val = val + "data:\n"
for subset in self.data:
- val = val + '\t%s = %s %s %s\n' % (subset, type(self.data[subset]),
- str(self.data[subset].shape),
- str(self.data[subset].dtype))
+ val = val + "\t%s = %s %s %s\n" % (
+ subset,
+ type(self.data[subset]),
+ str(self.data[subset].shape),
+ str(self.data[subset].dtype),
+ )
if isinstance(self.data[subset], scipy.sparse.spmatrix):
- val = val + '\tdensity: %f\n' % \
- (float(len(self.data[subset].data)) /
- self.data[subset].shape[0] /
- self.data[subset].shape[1])
- val = val + 'feat_type:\t' + str(self.feat_type) + '\n'
+ val = val + "\tdensity: %f\n" % (
+ float(len(self.data[subset].data))
+ / self.data[subset].shape[0]
+ / self.data[subset].shape[1]
+ )
+ val = val + "feat_type:\t" + str(self.feat_type) + "\n"
return val
diff --git a/autosklearn/data/feature_validator.py b/autosklearn/data/feature_validator.py
index cc4b578bdf..07c108390c 100644
--- a/autosklearn/data/feature_validator.py
+++ b/autosklearn/data/feature_validator.py
@@ -1,25 +1,18 @@
+from typing import Dict, List, Optional, Tuple, Union, cast
+
import logging
-import typing
+import warnings
import numpy as np
-
import pandas as pd
from pandas.api.types import is_numeric_dtype, is_sparse
-
-import scipy.sparse
-
+from scipy.sparse import csr_matrix, spmatrix
from sklearn.base import BaseEstimator
from sklearn.exceptions import NotFittedError
from autosklearn.util.logging_ import PickableLoggerAdapter
-
-SUPPORTED_FEAT_TYPES = typing.Union[
- typing.List,
- pd.DataFrame,
- np.ndarray,
- scipy.sparse.spmatrix,
-]
+SUPPORTED_FEAT_TYPES = Union[List, pd.DataFrame, np.ndarray, spmatrix]
class FeatureValidator(BaseEstimator):
@@ -31,34 +24,39 @@ class FeatureValidator(BaseEstimator):
Attributes
----------
- feat_type: typing.Optional[typing.List[str]]
- In case the dataset is not a pandas DataFrame:
- + If provided, this list indicates which columns should be treated as categorical
- it is internally transformed into a dictionary that indicates a mapping from
- column index to categorical/numerical
- + If not provided, by default all columns are treated as numerical
- If the input dataset is of type pandas dataframe, this argument
- must be none, as the column type will be inferred from the pandas dtypes.
-
- data_type:
- Class name of the data type provided during fit.
+ feat_type: Optional[List[str]] = None
+ In case the dataset is not a pandas DataFrame:
+ + If provided, this list indicates which columns should be treated as
+ categorical it is internally transformed into a dictionary that
+ indicates a mapping from column index to categorical/numerical.
+ + If not provided, by default all columns are treated as numerical
+
+ If the input dataset is of type pandas dataframe, this argument
+ must be none, as the column type will be inferred from the pandas dtypes.
+
+ data_type:
+ Class name of the data type provided during fit.
"""
- def __init__(self,
- feat_type: typing.Optional[typing.List[str]] = None,
- logger: typing.Optional[PickableLoggerAdapter] = None,
- ) -> None:
+
+ def __init__(
+ self,
+ feat_type: Optional[List[str]] = None,
+ logger: Optional[PickableLoggerAdapter] = None,
+ allow_string_features: bool = True,
+ ) -> None:
# If a dataframe was provided, we populate
- # this attribute with a mapping from column to {numerical | categorical}
- self.feat_type: typing.Optional[
- typing.Dict[typing.Union[str, int], str]
- ] = None
+ # this attribute with a mapping from column to
+ # {numerical | categorical | string}
+ self.feat_type: Optional[Dict[Union[str, int], str]] = None
if feat_type is not None:
if isinstance(feat_type, dict):
self.feat_type = feat_type
- elif not isinstance(feat_type, list):
- raise ValueError("Auto-Sklearn expects a list of categorical/"
- "numerical feature types, yet a"
- " {} was provided".format(type(feat_type)))
+ elif not isinstance(feat_type, List):
+ raise ValueError(
+ "Auto-Sklearn expects a list of categorical/"
+ "numerical/string feature types, yet a"
+ " {} was provided".format(type(feat_type))
+ )
else:
# Convert to a dictionary which will be passed to the ColumnTransformer
@@ -66,18 +64,19 @@ def __init__(self,
self.feat_type = {i: feat for i, feat in enumerate(feat_type)}
# Register types to detect unsupported data format changes
- self.data_type = None # type: typing.Optional[type]
- self.dtypes = {} # type: typing.Dict[str, str]
+ self.data_type = None # type: Optional[type]
+ self.dtypes = {} # type: Dict[str, str]
self.logger = logger if logger is not None else logging.getLogger(__name__)
self._is_fitted = False
+ self.allow_string_features = allow_string_features
def fit(
self,
X_train: SUPPORTED_FEAT_TYPES,
- X_test: typing.Optional[SUPPORTED_FEAT_TYPES] = None,
- ) -> BaseEstimator:
+ X_test: Optional[SUPPORTED_FEAT_TYPES] = None,
+ ) -> "FeatureValidator":
"""
Validates input data to Auto-Sklearn.
The supported data types are List, numpy arrays and pandas DataFrames.
@@ -88,12 +87,11 @@ def fit(
X_train: SUPPORTED_FEAT_TYPES
A set of features that are going to be validated (type and dimensionality
checks) and a encoder fitted in the case the data needs encoding
- X_test: typing.Optional[SUPPORTED_FEAT_TYPES]
+ X_test: Optional[SUPPORTED_FEAT_TYPES]
A hold out set of data used for checking
"""
-
# If a list was provided, it will be converted to pandas
- if isinstance(X_train, list):
+ if isinstance(X_train, List):
X_train, X_test = self.list_to_dataframe(X_train, X_test)
self._check_data(X_train)
@@ -101,47 +99,56 @@ def fit(
# Handle categorical feature identification for the pipeline
if hasattr(X_train, "iloc"):
if self.feat_type is not None:
- raise ValueError("When providing a DataFrame to Auto-Sklearn, we extract "
- "the feature types from the DataFrame.dtypes. That is, "
- "providing the option feat_type to the fit method is not "
- "supported when using a Dataframe. Please make sure that the "
- "type of each column in your DataFrame is properly set. "
- "More details about having the correct data type in your "
- "DataFrame can be seen in "
- "https://pandas.pydata.org/pandas-docs/stable/reference"
- "/api/pandas.DataFrame.astype.html")
+ raise ValueError(
+ "When providing a DataFrame to Auto-Sklearn, we extract "
+ "the feature types from the DataFrame.dtypes. That is, "
+ "providing the option feat_type to the fit method is not "
+ "supported when using a Dataframe. Please make sure that the "
+ "type of each column in your DataFrame is properly set. "
+ "More details about having the correct data type in your "
+ "DataFrame can be seen in "
+ "https://pandas.pydata.org/pandas-docs/stable/reference"
+ "/api/pandas.DataFrame.astype.html"
+ )
else:
self.feat_type = self.get_feat_type_from_columns(X_train)
else:
# Numpy array was provided
if self.feat_type is None:
# Assume numerical columns if a numpy array has no feature types
- self.feat_type = {i: 'numerical' for i in range(np.shape(X_train)[1])}
+ self.feat_type = {i: "numerical" for i in range(np.shape(X_train)[1])}
else:
# Check The feat type provided
if len(self.feat_type) != np.shape(X_train)[1]:
- raise ValueError('Array feat_type does not have same number of '
- 'variables as X has features. %d vs %d.' %
- (len(self.feat_type), np.shape(X_train)[1]))
+ raise ValueError(
+ "Array feat_type does not have same number of "
+ "variables as X has features. %d vs %d."
+ % (len(self.feat_type), np.shape(X_train)[1])
+ )
if not all([isinstance(f, str) for f in self.feat_type.values()]):
- raise ValueError("feat_type must only contain strings: {}".format(
- list(self.feat_type.values()),
- ))
+ raise ValueError(
+ "feat_type must only contain strings: {}".format(
+ list(self.feat_type.values()),
+ )
+ )
for ft in self.feat_type.values():
- if ft.lower() not in ['categorical', 'numerical']:
- raise ValueError('Only `Categorical` and `Numerical` are '
- 'valid feature types, you passed `%s`' % ft)
+ if ft.lower() not in ["categorical", "numerical", "string"]:
+ raise ValueError(
+ "Only `Categorical`, `Numerical` and `String` are "
+ "valid feature types"
+ )
if X_test is not None:
self._check_data(X_test)
if np.shape(X_train)[1] != np.shape(X_test)[1]:
- raise ValueError("The feature dimensionality of the train and test "
- "data does not match train({}) != test({})".format(
- np.shape(X_train)[1],
- np.shape(X_test)[1]
- ))
+ raise ValueError(
+ "The feature dimensionality of the train and test "
+ "data does not match train({}) != test({})".format(
+ np.shape(X_train)[1], np.shape(X_test)[1]
+ )
+ )
self._is_fitted = True
@@ -150,43 +157,50 @@ def fit(
def transform(
self,
X: SUPPORTED_FEAT_TYPES,
- ) -> np.ndarray:
+ ) -> Union[np.ndarray, spmatrix, pd.DataFrame]:
"""
Validates and fit a categorical encoder (if needed) to the features.
The supported data types are List, numpy arrays and pandas DataFrames.
Parameters
----------
- X_train: SUPPORTED_FEAT_TYPES
- A set of features, whose categorical features are going to be
- transformed
+ X_train: SUPPORTED_FEAT_TYPES
+ A set of features, whose categorical features are going to be
+ transformed
Return
------
- np.ndarray:
- The transformed array
+ np.ndarray | spmatrix | pd.DataFrame:
+ The transformed array
"""
if not self._is_fitted:
- raise NotFittedError("Cannot call transform on a validator that is not fitted")
+ raise NotFittedError(
+ "Cannot call transform on a validator that is not fitted"
+ )
# If a list was provided, it will be converted to pandas
- if isinstance(X, list):
- X, _ = self.list_to_dataframe(X)
+ if isinstance(X, List):
+ X_transformed, _ = self.list_to_dataframe(X)
+ else:
+ X_transformed = X
# Check the data here so we catch problems on new test data
- self._check_data(X)
+ self._check_data(X_transformed)
# Sparse related transformations
# Not all sparse format support index sorting
- if scipy.sparse.issparse(X):
- if not isinstance(X, scipy.sparse.csr_matrix):
- self.logger.warning(f"Sparse data provided is of type {type(X)} "
- "yet Auto-Sklearn only support csr_matrix. Auto-sklearn "
- "will convert the provided data to the csr_matrix format.")
- X = X.tocsr(copy=False)
- if hasattr(X, 'sort_indices'):
- X.sort_indices()
- return X
+ if isinstance(X_transformed, spmatrix):
+ if not isinstance(X_transformed, csr_matrix):
+ self.logger.warning(
+ f"Sparse data provided is of type {type(X_transformed)} "
+ "yet Auto-Sklearn only support csr_matrix. Auto-sklearn "
+ "will convert the provided data to the csr_matrix format."
+ )
+ X_transformed = X_transformed.tocsr(copy=False)
+
+ X_transformed.sort_indices()
+
+ return X_transformed
def _check_data(
self,
@@ -197,51 +211,48 @@ def _check_data(
Parameters
----------
- X: SUPPORTED_FEAT_TYPES
- A set of features that are going to be validated (type and dimensionality
- checks) and a encoder fitted in the case the data needs encoding
+ X: SUPPORTED_FEAT_TYPES
+ A set of features that are going to be validated (type and dimensionality)
+ and a encoder fitted in the case the data needs encoding
"""
-
# We consider columns that are all nan in a pandas frame as category
- if hasattr(X, 'columns'):
- for column in typing.cast(pd.DataFrame, X).columns:
+ if hasattr(X, "columns"):
+ for column in cast(pd.DataFrame, X).columns:
if X[column].isna().all():
- X[column] = X[column].astype('category')
+ X[column] = X[column].astype("category")
- if not isinstance(X, (np.ndarray, pd.DataFrame)) and not scipy.sparse.issparse(X):
- raise ValueError("Auto-sklearn only supports Numpy arrays, Pandas DataFrames,"
- " scipy sparse and Python Lists, yet, the provided input is"
- " of type {}".format(
- type(X)
- ))
+ if not isinstance(X, (np.ndarray, pd.DataFrame)) and not isinstance(
+ X, spmatrix
+ ):
+ raise ValueError(
+ "Auto-sklearn only supports Numpy arrays, Pandas DataFrames,"
+ " scipy sparse and Python Lists, yet, the provided input is"
+ " of type {}".format(type(X))
+ )
if self.data_type is None:
self.data_type = type(X)
if self.data_type != type(X):
- self.logger.warning("Auto-sklearn previously received features of type %s "
- "yet the current features have type %s. Changing the dtype "
- "of inputs to an estimator might cause problems" % (
- str(self.data_type),
- str(type(X)),
- ),
- )
+ self.logger.warning(
+ f"Auto-sklearn previously received features of type {self.data_type} "
+ f"yet the current features have type {type(X)}. Changing the dtype "
+ "of inputs to an estimator might cause problems"
+ )
# Do not support category/string numpy data. Only numbers
if hasattr(X, "dtype"):
if not np.issubdtype(X.dtype.type, np.number): # type: ignore[union-attr]
raise ValueError(
- "When providing a numpy array to Auto-sklearn, the only valid "
- "dtypes are numerical ones. The provided data type {} is not supported."
- "".format(
- X.dtype.type, # type: ignore[union-attr]
- )
+ "When providing a numpy array to Auto-sklearn, the only valid"
+ f" dtypes are numerical ones. The provided data type {X.dtype.type}"
+ " is not supported."
)
# Then for Pandas, we do not support Nan in categorical columns
if hasattr(X, "iloc"):
# If entered here, we have a pandas dataframe
- X = typing.cast(pd.DataFrame, X)
+ X = cast(pd.DataFrame, X)
dtypes = {col: X[col].dtype.name.lower() for col in X.columns}
if len(self.dtypes) > 0:
@@ -249,71 +260,84 @@ def _check_data(
# To support list, we need to support object inference.
# In extreme cases, the train column might be all integer,
# and the test column might be float.
- self.logger.warning("Changing the dtype of the features after fit() is "
- "not recommended. Fit() method was called with "
- "{} whereas the new features have {} as type".format(
- self.dtypes,
- dtypes,
- ))
+ self.logger.warning(
+ "Changing the dtype of the features after fit() is "
+ "not recommended. Fit() method was called with "
+ "{} whereas the new features have {} as type".format(
+ self.dtypes,
+ dtypes,
+ )
+ )
else:
self.dtypes = dtypes
def get_feat_type_from_columns(
self,
X: pd.DataFrame,
- ) -> typing.Dict[typing.Union[str, int], str]:
+ ) -> Dict[Union[str, int], str]:
"""
Returns a dictionary that maps pandas dataframe columns to a feature type.
- This feature type can be categorical or numerical
+ This feature type can be categorical, numerical or string
Parameters
----------
- X: pd.DataFrame
- A set of features that are going to be validated (type and dimensionality
- checks) and a encoder fitted in the case the data needs encoding
+ X: pd.DataFrame
+ A set of features that are going to be validated (type and dimensionality
+ checks) and a encoder fitted in the case the data needs encoding
+
Returns
-------
- feat_type:
- dictionary with column to feature type mapping
+ feat_type:
+ dictionary with column to feature type mapping
"""
-
# Also, register the feature types for the estimator
feat_type = {}
# Make sure each column is a valid type
for i, column in enumerate(X.columns):
if is_sparse(X[column]):
- raise ValueError("Auto-sklearn does not yet support sparse pandas Series."
- f" Please convert {column} to a dense format.")
- elif X[column].dtype.name in ['category', 'bool']:
-
- feat_type[column] = 'categorical'
+ raise ValueError(
+ "Auto-sklearn does not yet support sparse pandas Series."
+ f" Please convert {column} to a dense format."
+ )
+ elif X[column].dtype.name in ["category", "bool"]:
+ feat_type[column] = "categorical"
+ elif X[column].dtype.name == "string":
+ if self.allow_string_features:
+ feat_type[column] = "string"
+ else:
+ feat_type[column] = "categorical"
+ warnings.warn(
+ f"you disabled text encoding column {column} will be "
+ f"encoded as category"
+ )
# Move away from np.issubdtype as it causes
# TypeError: data type not understood in certain pandas types
elif not is_numeric_dtype(X[column]):
- if X[column].dtype.name == 'object':
- raise ValueError(
- "Input Column {} has invalid type object. "
- "Cast it to a valid dtype before using it in Auto-Sklearn. "
- "Valid types are numerical, categorical or boolean. "
- "You can cast it to a valid dtype using "
- "pandas.Series.astype ."
- "If working with string objects, the following "
- "tutorial illustrates how to work with text data: "
- "https://scikit-learn.org/stable/tutorial/text_analytics/working_with_text_data.html".format( # noqa: E501
- column,
- )
+ if X[column].dtype.name == "object":
+ warnings.warn(
+ f"Input Column {column} has generic type object. "
+ f"Autosklearn will treat this column as string. "
+ f"Please ensure that this setting is suitable for your task.",
+ UserWarning,
)
+ if self.allow_string_features:
+ feat_type[column] = "string"
+ else:
+ feat_type[column] = "categorical"
+ warnings.warn(
+ f"you disabled text encoding column {column} will be"
+ f"encoded as category"
+ )
elif pd.core.dtypes.common.is_datetime_or_timedelta_dtype(
X[column].dtype
):
raise ValueError(
- "Auto-sklearn does not support time and/or date datatype as given "
- "in column {}. Please convert the time information to a numerical value "
- "first. One example on how to do this can be found on "
- "https://stats.stackexchange.com/questions/311494/".format(
- column,
- )
+ "Auto-sklearn does not support time and/or date datatype as "
+ f"given in column {column}. Please convert the time "
+ " information to a numerical value first. One example on how to"
+ " do this can be found on "
+ " https://stats.stackexchange.com/questions/311494/"
)
else:
raise ValueError(
@@ -326,46 +350,41 @@ def get_feat_type_from_columns(
)
)
else:
- feat_type[column] = 'numerical'
+ feat_type[column] = "numerical"
return feat_type
def list_to_dataframe(
self,
X_train: SUPPORTED_FEAT_TYPES,
- X_test: typing.Optional[SUPPORTED_FEAT_TYPES] = None,
- ) -> typing.Tuple[pd.DataFrame, typing.Optional[pd.DataFrame]]:
+ X_test: Optional[SUPPORTED_FEAT_TYPES] = None,
+ ) -> Tuple[pd.DataFrame, Optional[pd.DataFrame]]:
"""
- Converts a list to a pandas DataFrame. In this process, column types are inferred.
+ Converts a list to a DataFrame. In this process, column types are inferred.
If test data is provided, we proactively match it to train data
Parameters
----------
- X_train: SUPPORTED_FEAT_TYPES
- A set of features that are going to be validated (type and dimensionality
- checks) and a encoder fitted in the case the data needs encoding
- X_test: typing.Optional[SUPPORTED_FEAT_TYPES]
- A hold out set of data used for checking
+ X_train: SUPPORTED_FEAT_TYPES
+ A set of features that are going to be validated (type and dimensionality
+ checks) and a encoder fitted in the case the data needs encoding
+
+ X_test: Optional[SUPPORTED_FEAT_TYPES]
+ A hold out set of data used for checking
+
Returns
-------
- pd.DataFrame:
- transformed train data from list to pandas DataFrame
- pd.DataFrame:
- transformed test data from list to pandas DataFrame
+ Union[pd.DataFrame, pd.DataFrame]:
+ transformed (train, test) data from list to pandas DataFrame
"""
-
# If a list was provided, it will be converted to pandas
X_train = pd.DataFrame(data=X_train).convert_dtypes()
# Store the dtypes and use in case of re-fit
if len(self.dtypes) == 0:
- # Categorical data is inferred as string. Convert to categorical.
- # Warn the user about dtypes or request him to use a dataframe
- for col in X_train.columns:
- if X_train[col].dtype.name == 'string':
- X_train[col] = X_train[col].astype('category')
-
- self.dtypes = {col: X_train[col].dtype.name.lower() for col in X_train.columns}
+ self.dtypes = {
+ col: X_train[col].dtype.name.lower() for col in X_train.columns
+ }
else:
for col in X_train.columns:
# Try to convert to the original dtype used to fit the validator
@@ -374,25 +393,31 @@ def list_to_dataframe(
try:
X_train[col] = X_train[col].astype(self.dtypes[col])
except Exception as e:
- self.logger.warning(f"Failed to format column {col} as {self.dtypes[col]}: {e}")
+ self.logger.warning(
+ f"Failed to format column {col} as {self.dtypes[col]}: {e}"
+ )
self.dtypes[col] = X_train[col].dtype.name.lower()
- self.logger.warning("The provided feature types to autosklearn are of type list."
- "Features have been interpreted as: {}".format(
- [(col, t) for col, t in zip(X_train.columns, X_train.dtypes)]
- ))
+ self.logger.warning(
+ "The provided feature types to autosklearn are of type list."
+ "Features have been interpreted as: {}".format(
+ [(col, t) for col, t in zip(X_train.columns, X_train.dtypes)]
+ )
+ )
if X_test is not None:
- if not isinstance(X_test, list):
- self.logger.warning("Train features are a list while the provided test data"
- "is {}. X_test will be casted as DataFrame.".format(
- type(X_test)
- ))
+ if not isinstance(X_test, List):
+ self.logger.warning(
+ "Train features are a list while the provided test data"
+ "is {}. X_test will be casted as DataFrame.".format(type(X_test))
+ )
X_test = pd.DataFrame(data=X_test)
for col in X_test.columns:
try:
X_test[col] = X_test[col].astype(self.dtypes[col])
except Exception as e:
- self.logger.warning(f"Failed to format column {col} as {self.dtypes[col]}: {e}")
+ self.logger.warning(
+ f"Failed to format column {col} as {self.dtypes[col]}: {e}"
+ )
self.dtypes[col] = X_test[col].dtype.name.lower()
return X_train, X_test
diff --git a/autosklearn/data/target_validator.py b/autosklearn/data/target_validator.py
index 55a3bfc5d0..740fa67cb2 100644
--- a/autosklearn/data/target_validator.py
+++ b/autosklearn/data/target_validator.py
@@ -1,71 +1,76 @@
+from typing import List, Optional, Type, Union, cast
+
import logging
-import typing
+import warnings
import numpy as np
-
import pandas as pd
-from pandas.api.types import is_numeric_dtype
-
-import scipy.sparse
-
import sklearn.utils
-from sklearn import preprocessing
+from pandas.api.types import is_numeric_dtype
+from pandas.core.dtypes.base import ExtensionDtype
+from scipy.sparse import spmatrix
from sklearn.base import BaseEstimator
from sklearn.exceptions import NotFittedError
+from sklearn.preprocessing import OrdinalEncoder
from sklearn.utils.multiclass import type_of_target
from autosklearn.util.logging_ import PickableLoggerAdapter
-
-SUPPORTED_TARGET_TYPES = typing.Union[
- typing.List,
- pd.Series,
- pd.DataFrame,
- np.ndarray,
- scipy.sparse.spmatrix,
-]
+SUPPORTED_TARGET_TYPES = Union[List, pd.Series, pd.DataFrame, np.ndarray, spmatrix]
+SUPPORTED_XDATA_TYPES = Union[pd.Series, pd.DataFrame, np.ndarray, spmatrix]
class TargetValidator(BaseEstimator):
- """
- A class to pre-process targets. It validates the data provided during fit (to make sure
- it matches Sklearn expectation) as well as encoding the targets in case of classification
+ """A class to pre-process targets.
+
+ It validates the data provided during fit to make sure it matches Sklearn
+ expectation as well as encoding the targets in case of classification
+
Attributes
----------
- is_classification: bool
- A bool that indicates if the validator should operate in classification mode.
- During classification, the targets are encoded.
- encoder: typing.Optional[BaseEstimator]
- Host a encoder object if the data requires transformation (for example,
- if provided a categorical column in a pandas DataFrame)
- enc_columns: typing.List[str]
- List of columns that where encoded
+ is_classification: bool
+ A bool that indicates if the validator should operate in classification mode.
+ During classification, the targets are encoded.
+
+ encoder: Optional[BaseEstimator]
+ Host a encoder object if the data requires transformation (for example, if
+ provided a categorical column in a pandas DataFrame).
+
+ enc_columns: List[str]
+ List of columns that where encoded
"""
- def __init__(self,
- is_classification: bool = False,
- logger: typing.Optional[PickableLoggerAdapter] = None,
- ) -> None:
+
+ def __init__(
+ self,
+ is_classification: bool = False,
+ logger: Optional[PickableLoggerAdapter] = None,
+ ) -> None:
self.is_classification = is_classification
- self.data_type = None # type: typing.Optional[type]
+ self.data_type = None # type: Optional[type]
- self.encoder = None # type: typing.Optional[BaseEstimator]
+ # NOTE: type update
+ # Encoders don't seems to have a unified base class for
+ # the methods 'transform'. Could make a `prototype` class that
+ # duck types for the 'transform', 'fit_transform' methods
+ # This is however fine while we only use ordinal encoder
+ self.encoder = None # type: Optional[OrdinalEncoder]
- self.out_dimensionality = None # type: typing.Optional[int]
- self.type_of_target = None # type: typing.Optional[str]
+ self.out_dimensionality = None # type: Optional[int]
+ self.type_of_target = None # type: Optional[str]
self.logger = logger if logger is not None else logging.getLogger(__name__)
# Store the dtype for remapping to correct type
- self.dtype = None # type: typing.Optional[type]
+ self.dtype = None # type: Optional[Type]
self._is_fitted = False
def fit(
self,
- y_train: SUPPORTED_TARGET_TYPES,
- y_test: typing.Optional[SUPPORTED_TARGET_TYPES] = None,
- ) -> BaseEstimator:
+ y_train: Union[List, np.ndarray, pd.Series, pd.DataFrame],
+ y_test: Optional[Union[List, np.ndarray, pd.Series, pd.DataFrame]] = None,
+ ) -> "TargetValidator":
"""
Validates and fit a categorical encoder (if needed) to the targets
The supported data types are List, numpy arrays and pandas DataFrames.
@@ -74,7 +79,7 @@ def fit(
----------
y_train: SUPPORTED_TARGET_TYPES
A set of targets set aside for training
- y_test: typing.Union[SUPPORTED_TARGET_TYPES]
+ y_test: Union[SUPPORTED_TARGET_TYPES]
A hold out set of data used of the targets. It is also used to fit the
categories of the encoder.
"""
@@ -86,34 +91,40 @@ def fit(
self._check_data(y_test)
if len(shape) != len(np.shape(y_test)) or (
- len(shape) > 1 and (shape[1] != np.shape(y_test)[1])):
- raise ValueError("The dimensionality of the train and test targets "
- "does not match train({}) != test({})".format(
- np.shape(y_train),
- np.shape(y_test)
- ))
+ len(shape) > 1 and (shape[1] != np.shape(y_test)[1])
+ ):
+ raise ValueError(
+ "The dimensionality of the train and test"
+ " targets do not match"
+ f" train {np.shape(y_train)}"
+ f" != test {np.shape(y_test)}"
+ )
+
if isinstance(y_train, pd.DataFrame):
- y_train = typing.cast(pd.DataFrame, y_train)
- y_test = typing.cast(pd.DataFrame, y_test)
+ if not isinstance(y_test, pd.DataFrame):
+ y_test = pd.DataFrame(y_test)
+
if y_train.columns.tolist() != y_test.columns.tolist():
raise ValueError(
- "Train and test targets must both have the same columns, yet "
- "y={} and y_test={} ".format(
- y_train.columns,
- y_test.columns
- )
+ "Train and test targets must both have the"
+ f" same columns, yet y={y_train.columns}"
+ f" and y_test={y_test.columns}"
)
if list(y_train.dtypes) != list(y_test.dtypes):
- raise ValueError("Train and test targets must both have the same dtypes")
+ raise ValueError(
+ "Train and test targets must both have the same dtypes"
+ )
if self.out_dimensionality is None:
self.out_dimensionality = 1 if len(shape) == 1 else shape[1]
else:
_n_outputs = 1 if len(shape) == 1 else shape[1]
if self.out_dimensionality != _n_outputs:
- raise ValueError('Number of outputs changed from %d to %d!' %
- (self.out_dimensionality, _n_outputs))
+ raise ValueError(
+ "Number of outputs changed from %d to %d!"
+ % (self.out_dimensionality, _n_outputs)
+ )
# Fit on the training data
self._fit(y_train, y_test)
@@ -124,9 +135,9 @@ def fit(
def _fit(
self,
- y_train: SUPPORTED_TARGET_TYPES,
- y_test: typing.Optional[SUPPORTED_TARGET_TYPES] = None,
- ) -> BaseEstimator:
+ y_train: Union[List, np.ndarray, pd.Series, pd.DataFrame],
+ y_test: Optional[Union[List, np.ndarray, pd.Series, pd.DataFrame]] = None,
+ ) -> "TargetValidator":
"""
If dealing with classification, this utility encodes the targets.
@@ -138,69 +149,84 @@ def _fit(
y_train: SUPPORTED_TARGET_TYPES
The labels of the current task. They are going to be encoded in case
of classification
- y_test: typing.Optional[SUPPORTED_TARGET_TYPES]
+ y_test: Optional[SUPPORTED_TARGET_TYPES]
A holdout set of labels
"""
- if not self.is_classification or self.type_of_target == 'multilabel-indicator':
+ if not self.is_classification or self.type_of_target == "multilabel-indicator":
# Only fit an encoder for classification tasks
# Also, encoding multilabel indicator data makes the data multiclass
# Let the user employ a MultiLabelBinarizer if needed
+ self.encoder = None
return self
- if y_test is not None:
- if hasattr(y_train, "iloc"):
- y_train = pd.concat([y_train, y_test], ignore_index=True, sort=False)
- elif isinstance(y_train, list):
- y_train = y_train + y_test
- elif isinstance(y_train, np.ndarray):
- y_train = np.concatenate((y_train, y_test))
-
- ndim = len(np.shape(y_train))
- if ndim == 1 or (ndim > 1 and np.shape(y_train)[1] == 1):
- # The label encoder makes sure data is, and remains
- # 1 dimensional
- self.encoder = preprocessing.OrdinalEncoder(handle_unknown='use_encoded_value',
- unknown_value=-1)
+ # Validate the shape of the input
+ shape = np.shape(y_train)
+ ndim = len(shape)
+ if ndim > 1 and shape[1] != 1:
+ # We should not reach this if statement, we check for type of targets before
+ raise ValueError(
+ "Multi-dimensional classification is not yet"
+ " supported. Encoding multidimensional data"
+ " converts multiple columns to a 1 dimensional encoding."
+ f" Data involved = {shape}/{self.type_of_target}"
+ )
+
+ # Create the encoder
+ self.encoder = OrdinalEncoder(handle_unknown="error")
+
+ # Clear typing to just numpy arrays and pandas
+ if isinstance(y_train, List):
+ y = np.asarray(y_train)
else:
- # We should not reach this if statement as we check for type of targets before
- raise ValueError("Multi-dimensional classification is not yet supported. "
- "Encoding multidimensional data converts multiple columns "
- "to a 1 dimensional encoding. Data involved = {}/{}".format(
- np.shape(y_train),
- self.type_of_target
- ))
-
- # Mypy redefinition
- assert self.encoder is not None
-
- # remove ravel warning from pandas Series
- if ndim > 1:
- self.encoder.fit(y_train)
+ y = y_train
+
+ # Attempt to store dtype if it's numeric, we use this in
+ # inverse_transform to try corretly restore it's dtype
+ if isinstance(y, pd.Series):
+ if isinstance(y.dtype, ExtensionDtype):
+ warnings.warn(
+ "Fitting transformer with a pandas series which"
+ f" has the dtype {y.dtype}. Inverse transform"
+ " may not be able preserve dtype when converting"
+ " to np.ndarray"
+ )
+ if is_numeric_dtype(y.dtype):
+ self.dtype = y.dtype
+ elif isinstance(y, pd.DataFrame):
+ if is_numeric_dtype(y.dtypes[0]):
+ self.dtype = y.dtypes[0]
else:
- if hasattr(y_train, 'iloc'):
- y_train = typing.cast(pd.DataFrame, y_train)
- self.encoder.fit(y_train.to_numpy().reshape(-1, 1))
+ if is_numeric_dtype(y.dtype):
+ self.dtype = y.dtype
+
+ # Merge y_test and y_train for encoding
+ if y_test is not None:
+ if isinstance(y, (pd.Series, pd.DataFrame)):
+ if isinstance(y, pd.Series):
+ y_test = pd.Series(y_test)
+ else:
+ y_test = pd.DataFrame(y_test)
+
+ y = pd.concat([y, y_test], ignore_index=True, sort=False)
else:
- self.encoder.fit(np.array(y_train).reshape(-1, 1))
-
- # we leave objects unchanged, so no need to store dtype in this case
- if hasattr(y_train, 'dtype'):
- # Series and numpy arrays are checked here
- # Cast is as numpy for mypy checks
- y_train = typing.cast(np.ndarray, y_train)
- if is_numeric_dtype(y_train.dtype):
- self.dtype = y_train.dtype
- elif hasattr(y_train, 'dtypes') and is_numeric_dtype(typing.cast(pd.DataFrame,
- y_train).dtypes[0]):
- # This case is for pandas array with a single column
- y_train = typing.cast(pd.DataFrame, y_train)
- self.dtype = y_train.dtypes[0]
+ y = np.concatenate([y, y_test])
+
+ # If one dimensional, we need to reshape it
+ # [1, 2, 3] -> [[1], [2], [3]]
+ if y.ndim == 1:
+ if isinstance(y, pd.DataFrame) or isinstance(y, pd.Series):
+ y = y.to_numpy()
+
+ y = y.reshape(-1, 1)
+
+ # Fit encoder
+ self.encoder.fit(y)
return self
def transform(
self,
- y: typing.Union[SUPPORTED_TARGET_TYPES],
+ y: Union[List, pd.Series, pd.DataFrame, np.ndarray],
) -> np.ndarray:
"""
Validates and fit a categorical encoder (if needed) to the features.
@@ -217,85 +243,96 @@ def transform(
The transformed array
"""
if not self._is_fitted:
- raise NotFittedError("Cannot call transform on a validator that is not fitted")
+ raise NotFittedError("TargetValidator must have fit() called first")
# Check the data here so we catch problems on new test data
self._check_data(y)
- if self.encoder is not None:
- # remove ravel warning from pandas Series
- shape = np.shape(y)
- if len(shape) > 1:
- y = self.encoder.transform(y)
- else:
- # The Ordinal encoder expects a 2 dimensional input.
- # The targets are 1 dimensional, so reshape to match the expected shape
- if hasattr(y, 'iloc'):
- y = typing.cast(pd.DataFrame, y)
- y = self.encoder.transform(y.to_numpy().reshape(-1, 1)).reshape(-1)
- else:
- y = self.encoder.transform(np.array(y).reshape(-1, 1)).reshape(-1)
+ # Clear the types List and DataFrame off of y
+ if isinstance(y, List):
+ y_transformed = np.asarray(y)
+ elif isinstance(y, pd.DataFrame) or isinstance(y, pd.Series):
+ y_transformed = y.to_numpy()
+ else:
+ y_transformed = y
+
+ if self.encoder is None:
+ return y_transformed
- # sklearn check array will make sure we have the
- # correct numerical features for the array
- # Also, a numpy array will be created
- y = sklearn.utils.check_array(
- y,
+ # The Ordinal encoder expects a 2 dimensional input.
+ # The targets are 1 dimensional, so reshape to match the expected shape
+ if y_transformed.ndim == 1:
+ y_transformed = y_transformed.reshape(-1, 1)
+ y_transformed = self.encoder.transform(y_transformed).reshape(-1)
+ else:
+ y_transformed = self.encoder.transform(y_transformed)
+
+ # check_array ensures correct numerical features for array
+ y_transformed = sklearn.utils.check_array(
+ y_transformed,
force_all_finite=True,
- accept_sparse='csr',
+ accept_sparse="csr",
ensure_2d=False,
)
- # When translating a dataframe to numpy, make sure we
- # honor the ravel requirement
- if y.ndim == 2 and y.shape[1] == 1:
- y = np.ravel(y)
+ # Ensure we flatten any arrays [[1], [2], [1], [1], ...] to [1,2,1,1,...]
+ if y_transformed.ndim == 2 and y_transformed.shape[1] == 1:
+ y_transformed = y_transformed.ravel()
- return y
+ return y_transformed
def inverse_transform(
self,
- y: SUPPORTED_TARGET_TYPES,
+ y: Union[List, pd.Series, pd.DataFrame, np.ndarray],
) -> np.ndarray:
- """
- Revert any encoding transformation done on a target array
+ """Revert any encoding transformation done on a target array.
Parameters
----------
- y: typing.Union[np.ndarray, pd.DataFrame, pd.Series]
- Target array to be transformed back to original form before encoding
+ y: Union[List, np.ndarray, pd.DataFrame, pd.Series]
+ Target array to be transformed back to original form before encoding
+
Return
------
- np.ndarray:
- The transformed array
+ np.ndarray:
+ The transformed array
"""
if not self._is_fitted:
- raise NotFittedError("Cannot call inverse_transform on a validator that is not fitted")
+ raise NotFittedError("TargetValidator must have fit() called first")
+ # If there's no encoder, just return it as the expected ndarray
if self.encoder is None:
- return y
+ return np.asarray(y)
+
+ # If it's one dimensional, we need to convert it
+ # [1,2,3] -> [[1], [2], [3]]
shape = np.shape(y)
- if len(shape) > 1:
- y = self.encoder.inverse_transform(y)
+ if len(shape) == 1:
+
+ if isinstance(y, pd.DataFrame) or isinstance(y, pd.Series):
+ y = y.to_numpy()
+ elif isinstance(y, List):
+ y = np.asarray(y)
+
+ # Perform the transformation and reshape it
+ y = y.reshape(-1, 1)
+ y_inv = self.encoder.inverse_transform(y)
+ y_inv = y_inv.reshape(-1)
+
else:
- # The targets should be a flattened array, hence reshape with -1
- if hasattr(y, 'iloc'):
- y = typing.cast(pd.DataFrame, y)
- y = self.encoder.inverse_transform(y.to_numpy().reshape(-1, 1)).reshape(-1)
- else:
- y = self.encoder.inverse_transform(np.array(y).reshape(-1, 1)).reshape(-1)
+ # Otherwise we can can just invert it as expected
+ y_inv = self.encoder.inverse_transform(y)
# Inverse transform returns a numpy array of type object
# This breaks certain metrics as accuracy, which makes type_of_target be unknown
# If while fit a dtype was observed, we try to honor that dtype
if self.dtype is not None:
- y = y.astype(self.dtype)
- return y
+ y_inv = y_inv.astype(self.dtype)
+
+ return y_inv
def is_single_column_target(self) -> bool:
- """
- Output is encoded with a single column encoding
- """
+ """Output is encoded with a single column encoding"""
return self.out_dimensionality == 1
def _check_data(
@@ -307,44 +344,44 @@ def _check_data(
Parameters
----------
- y: typing.Union[np.ndarray, pd.DataFrame, pd.Series]
- A set of features whose dimensionality and data type is going to be checked
+ y: Union[np.ndarray, pd.DataFrame, pd.Series]
+ A set of features whose dimensionality and data type is going to be checked
"""
-
if not isinstance(
- y, (np.ndarray, pd.DataFrame, list, pd.Series)) and not scipy.sparse.issparse(y):
- raise ValueError("Auto-sklearn only supports Numpy arrays, Pandas DataFrames,"
- " pd.Series, sparse data and Python Lists as targets, yet, "
- "the provided input is of type {}".format(
- type(y)
- ))
-
- # Sparse data muss be numerical
- # Type ignore on attribute because sparse targets have a dtype
- if scipy.sparse.issparse(y) and not np.issubdtype(y.dtype.type, # type: ignore[union-attr]
- np.number):
- raise ValueError("When providing a sparse matrix as targets, the only supported "
- "values are numerical. Please consider using a dense"
- " instead."
- )
+ y, (np.ndarray, pd.DataFrame, List, pd.Series)
+ ) and not isinstance(y, spmatrix):
+ raise ValueError(
+ "Auto-sklearn only supports Numpy arrays, Pandas DataFrames,"
+ " pd.Series, sparse data and Python Lists as targets, yet, "
+ "the provided input is of type {}".format(type(y))
+ )
+
+ if isinstance(y, spmatrix) and not np.issubdtype(y.dtype.type, np.number):
+ raise ValueError(
+ "When providing a sparse matrix as targets, the only supported "
+ "values are numerical. Please consider using a dense"
+ " instead."
+ )
if self.data_type is None:
self.data_type = type(y)
if self.data_type != type(y):
- self.logger.warning("Auto-sklearn previously received targets of type %s "
- "yet the current features have type %s. Changing the dtype "
- "of inputs to an estimator might cause problems" % (
- str(self.data_type),
- str(type(y)),
- ),
- )
+ self.logger.warning(
+ "Auto-sklearn previously received targets of type %s "
+ "yet the current features have type %s. Changing the dtype "
+ "of inputs to an estimator might cause problems"
+ % (
+ str(self.data_type),
+ str(type(y)),
+ ),
+ )
# No Nan is supported
has_nan_values = False
- if hasattr(y, 'iloc'):
- has_nan_values = typing.cast(pd.DataFrame, y).isnull().values.any()
- if scipy.sparse.issparse(y):
- y = typing.cast(scipy.sparse.spmatrix, y)
+ if hasattr(y, "iloc"):
+ has_nan_values = cast(pd.DataFrame, y).isnull().values.any()
+
+ if isinstance(y, spmatrix):
has_nan_values = not np.array_equal(y.data, y.data)
else:
# List and array like values are considered here
@@ -352,34 +389,39 @@ def _check_data(
# but NaN, are not equal to themselves:
has_nan_values = not np.array_equal(y, y)
if has_nan_values:
- raise ValueError("Target values cannot contain missing/NaN values. "
- "This is not supported by scikit-learn. "
- )
+ raise ValueError(
+ "Target values cannot contain missing/NaN values. "
+ "This is not supported by scikit-learn. "
+ )
# Pandas Series is not supported for multi-label indicator
# This format checks are done by type of target
try:
self.type_of_target = type_of_target(y)
except Exception as e:
- raise ValueError("The provided data could not be interpreted by Sklearn. "
- "While determining the type of the targets via type_of_target "
- "run into exception: {}.".format(e))
-
- supported_output_types = ('binary',
- 'continuous',
- 'continuous-multioutput',
- 'multiclass',
- 'multilabel-indicator',
- # Notice unknown/multiclass-multioutput are not supported
- # This can only happen during testing only as estimators
- # should filter out unsupported types.
- )
+ raise ValueError(
+ "The provided data could not be interpreted by Sklearn. "
+ "While determining the type of the targets via type_of_target "
+ "run into exception: {}.".format(e)
+ )
+
+ supported_output_types = (
+ "binary",
+ "continuous",
+ "continuous-multioutput",
+ "multiclass",
+ "multilabel-indicator",
+ # Notice unknown/multiclass-multioutput are not supported
+ # This can only happen during testing only as estimators
+ # should filter out unsupported types.
+ )
if self.type_of_target not in supported_output_types:
- raise ValueError("Provided targets are not supported by Auto-Sklearn. "
- "Provided type is {} whereas supported types are {}.".format(
- self.type_of_target,
- supported_output_types
- ))
+ raise ValueError(
+ "Provided targets are not supported by Auto-Sklearn. "
+ "Provided type is {} whereas supported types are {}.".format(
+ self.type_of_target, supported_output_types
+ )
+ )
@property
def classes_(self) -> np.ndarray:
@@ -388,10 +430,11 @@ def classes_(self) -> np.ndarray:
which consist of a ndarray of shape (n_classes,)
where n_classes are the number of classes seen while fitting
a encoder to the targets.
+
Returns
-------
- classes_: np.ndarray
- The unique classes seen during encoding of a classifier
+ classes_: np.ndarray
+ The unique classes seen during encoding of a classifier
"""
if self.encoder is None:
return np.array([])
diff --git a/autosklearn/data/validation.py b/autosklearn/data/validation.py
index 4d276dd9d1..324bab4895 100644
--- a/autosklearn/data/validation.py
+++ b/autosklearn/data/validation.py
@@ -1,22 +1,22 @@
# -*- encoding: utf-8 -*-
+from typing import List, Optional, Tuple, Union
+
import logging
-import typing
import numpy as np
-
-from scipy.sparse import issparse, spmatrix
-
+import pandas as pd
+from scipy.sparse import spmatrix
from sklearn.base import BaseEstimator
from sklearn.exceptions import NotFittedError
-from autosklearn.data.feature_validator import FeatureValidator, SUPPORTED_FEAT_TYPES
+from autosklearn.data.feature_validator import SUPPORTED_FEAT_TYPES, FeatureValidator
from autosklearn.data.target_validator import SUPPORTED_TARGET_TYPES, TargetValidator
from autosklearn.util.logging_ import get_named_client_logger
def convert_if_sparse(
- y: typing.Union[SUPPORTED_TARGET_TYPES, spmatrix]
-) -> SUPPORTED_TARGET_TYPES:
+ y: SUPPORTED_TARGET_TYPES,
+) -> Union[np.ndarray, List, pd.DataFrame, pd.Series]:
"""If the labels `y` are sparse, it will convert it to its dense representation
Parameters
@@ -28,16 +28,17 @@ def convert_if_sparse(
-------
np.ndarray of shape (n_samples, ) or (n_samples, n_outputs)
"""
- if issparse(y):
- y = typing.cast(spmatrix, y)
- y = y.toarray()
- y = typing.cast(np.ndarray, y)
+ if isinstance(y, spmatrix):
+ y_ = y.toarray()
- # For one dimensional data, toarray will return (1, nrows)
- if y.shape[0] == 1:
- y = y.flatten()
+ # For sparse one dimensional data, y.toarray will return [[1], [2], [3], ...]
+ # We need to flatten this before returning it
+ if y_.shape[0] == 1:
+ y_ = y_.flatten()
+ else:
+ y_ = y
- return y
+ return y_
class InputValidator(BaseEstimator):
@@ -48,100 +49,122 @@ class InputValidator(BaseEstimator):
This class also perform checks for data integrity and flags the user
via informative errors.
+
Attributes
----------
- feat_type: typing.Optional[typing.List[str]]
- In case the dataset is not a pandas DataFrame:
- + If provided, this list indicates which columns should be treated as categorical
- it is internally transformed into a dictionary that indicates a mapping from
- column index to categorical/numerical
- + If not provided, by default all columns are treated as numerical
- If the input dataset is of type pandas dataframe, this argument
- must be none, as the column type will be inferred from the pandas dtypes.
- is_classification: bool
- For classification task, this flag indicates that the target data
- should be encoded
- feature_validator: FeatureValidator
- A FeatureValidator instance used to validate and encode feature columns to match
- sklearn expectations on the data
- target_validator: TargetValidator
- A TargetValidator instance used to validate and encode (in case of classification)
- the target values
+ feat_type: Optional[List[str]] = None
+ In case the dataset is not a pandas DataFrame:
+ + If provided, this list indicates which columns should be treated as
+ categorical it is internally transformed into a dictionary that
+ indicates a mapping from column index to categorical/numerical.
+ + If not provided, by default all columns are treated as numerical
+
+ If the input dataset is of type pandas dataframe, this argument
+ must be none, as the column type will be inferred from the pandas dtypes.
+
+ is_classification: bool
+ For classification task, this flag indicates that the target data
+ should be encoded
+
+ feature_validator: FeatureValidator
+ A FeatureValidator instance used to validate and encode feature columns to match
+ sklearn expectations on the data
+
+ target_validator: TargetValidator
+ A TargetValidator instance used for classification to validate and encode the
+ target values
"""
+
def __init__(
self,
- feat_type: typing.Optional[typing.List[str]] = None,
+ feat_type: Optional[List[str]] = None,
is_classification: bool = False,
- logger_port: typing.Optional[int] = None,
+ logger_port: Optional[int] = None,
+ allow_string_features: bool = True,
) -> None:
self.feat_type = feat_type
self.is_classification = is_classification
self.logger_port = logger_port
if self.logger_port is not None:
self.logger = get_named_client_logger(
- name='Validation',
+ name="Validation",
port=self.logger_port,
)
else:
- self.logger = logging.getLogger('Validation')
-
- self.feature_validator = FeatureValidator(feat_type=self.feat_type,
- logger=self.logger)
- self.target_validator = TargetValidator(is_classification=self.is_classification,
- logger=self.logger)
+ self.logger = logging.getLogger("Validation")
+
+ self.allow_string_features = allow_string_features
+ self.feature_validator = FeatureValidator(
+ feat_type=self.feat_type,
+ logger=self.logger,
+ allow_string_features=self.allow_string_features,
+ )
+ self.target_validator = TargetValidator(
+ is_classification=self.is_classification, logger=self.logger
+ )
self._is_fitted = False
def fit(
self,
X_train: SUPPORTED_FEAT_TYPES,
y_train: SUPPORTED_TARGET_TYPES,
- X_test: typing.Optional[SUPPORTED_FEAT_TYPES] = None,
- y_test: typing.Optional[SUPPORTED_TARGET_TYPES] = None,
+ X_test: Optional[SUPPORTED_FEAT_TYPES] = None,
+ y_test: Optional[SUPPORTED_TARGET_TYPES] = None,
) -> BaseEstimator:
"""
Validates and fit a categorical encoder (if needed) to the features, and
a encoder for targets in the case of classification. Specifically:
For features:
- + Valid data types are enforced (List, np.ndarray, pd.DataFrame, pd.Series, scipy
- sparse) as well as dimensionality checks
- + If the provided data is a pandas DataFrame with categorical/boolean/int columns,
- such columns will be encoded using an Ordinal Encoder
+ Valid data types are enforced (List, np.ndarray, pd.DataFrame, pd.Series, scipy
+ sparse) as well as dimensionality checks
+
+ If the provided data is a pandas DataFrame with categorical/boolean/int columns,
+ such columns will be encoded using an Ordinal Encoder
+
For targets:
- + Checks for dimensionality as well as missing values are performed.
- + If performing a classification task, the data is going to be encoded
+ * Checks for dimensionality as well as missing values are performed.
+ * If performing a classification task, the data is going to be encoded
Parameters
----------
- X_train: SUPPORTED_FEAT_TYPES
- A set of features that are going to be validated (type and dimensionality
- checks). If this data contains categorical columns, an encoder is going to
- be instantiated and trained with this data.
- y_train: SUPPORTED_TARGET_TYPES
- A set of targets that are going to be encoded if the task is for classification
- X_test: typing.Optional[SUPPORTED_FEAT_TYPES]
- A hold out set of features used for checking
- y_test: SUPPORTED_TARGET_TYPES
- A hold out set of targets used for checking. Additionally, if the current task
- is a classification task, this y_test categories are also going to be used to
- fit a pre-processing encoding (to prevent errors on unseen classes).
+ X_train: SUPPORTED_FEAT_TYPES
+ A set of features that are going to be validated (type and dimensionality
+ checks). If this data contains categorical columns, an encoder is going to
+ be instantiated and trained with this data.
+
+ y_train: SUPPORTED_TARGET_TYPES
+ A set of targets to encoded if the task is for classification.
+
+ X_test: Optional[SUPPORTED_FEAT_TYPES]
+ A hold out set of features used for checking
+
+ y_test: SUPPORTED_TARGET_TYPES
+ A hold out set of targets used for checking. Additionally, if the current
+ task is a classification task, this y_test categories are also going to be
+ used to fit a pre-processing encoding (to prevent errors on unseen classes).
+
Returns
-------
- self
+ self
"""
# Check that the data is valid
if np.shape(X_train)[0] != np.shape(y_train)[0]:
- raise ValueError("Inconsistent number of train datapoints for features and targets,"
- " {} for features and {} for targets".format(
- np.shape(X_train)[0],
- np.shape(y_train)[0],
- ))
+ raise ValueError(
+ "Inconsistent number of train datapoints for features and targets,"
+ " {} for features and {} for targets".format(
+ np.shape(X_train)[0],
+ np.shape(y_train)[0],
+ )
+ )
if X_test is not None and np.shape(X_test)[0] != np.shape(y_test)[0]:
- raise ValueError("Inconsistent number of test datapoints for features and targets,"
- " {} for features and {} for targets".format(
- np.shape(X_test)[0],
- np.shape(y_test)[0],
- ))
+ raise ValueError(
+ "Inconsistent number of test datapoints for features and targets,"
+ " {} for features and {} for targets".format(
+ np.shape(X_test)[0],
+ np.shape(y_test)[0],
+ )
+ )
self.feature_validator.fit(X_train, X_test)
self.target_validator.fit(y_train, y_test)
@@ -152,8 +175,8 @@ def fit(
def transform(
self,
X: SUPPORTED_FEAT_TYPES,
- y: typing.Optional[SUPPORTED_TARGET_TYPES] = None,
- ) -> typing.Tuple[np.ndarray, typing.Optional[np.ndarray]]:
+ y: Optional[Union[List, pd.Series, pd.DataFrame, np.ndarray]] = None,
+ ) -> Tuple[Union[np.ndarray, pd.DataFrame, spmatrix], Optional[np.ndarray]]:
"""
Transform the given target or features to a numpy array
@@ -161,7 +184,7 @@ def transform(
----------
X: SUPPORTED_FEAT_TYPES
A set of features to transform
- y: typing.Optional[SUPPORTED_TARGET_TYPES]
+ y: Optional[SUPPORTED_TARGET_TYPES]
A set of targets to transform
Return
@@ -172,9 +195,13 @@ def transform(
The transformed targets array
"""
if not self._is_fitted:
- raise NotFittedError("Cannot call transform on a validator that is not fitted")
+ raise NotFittedError(
+ "Cannot call transform on a validator that is not fitted"
+ )
+
X_transformed = self.feature_validator.transform(X)
if y is not None:
- return X_transformed, self.target_validator.transform(y)
+ y_transformed = self.target_validator.transform(y)
+ return X_transformed, y_transformed
else:
- return X_transformed, y
+ return X_transformed, None
diff --git a/autosklearn/data/xy_data_manager.py b/autosklearn/data/xy_data_manager.py
index 4c539157ee..d8cd467214 100644
--- a/autosklearn/data/xy_data_manager.py
+++ b/autosklearn/data/xy_data_manager.py
@@ -2,9 +2,7 @@
from typing import Dict, Optional, Union, cast
import numpy as np
-
import pandas as pd
-
from scipy import sparse
from autosklearn.constants import (
@@ -15,14 +13,10 @@
REGRESSION,
)
from autosklearn.data.abstract_data_manager import AbstractDataManager
-from autosklearn.data.validation import (
- SUPPORTED_FEAT_TYPES,
- SUPPORTED_TARGET_TYPES,
-)
+from autosklearn.data.validation import SUPPORTED_FEAT_TYPES, SUPPORTED_TARGET_TYPES
class XYDataManager(AbstractDataManager):
-
def __init__(
self,
X: SUPPORTED_FEAT_TYPES,
@@ -31,52 +25,59 @@ def __init__(
y_test: Optional[SUPPORTED_TARGET_TYPES],
task: int,
feat_type: Dict[Union[str, int], str],
- dataset_name: str
+ dataset_name: str,
):
super(XYDataManager, self).__init__(dataset_name)
- self.info['task'] = task
+ self.info["task"] = task
if sparse.issparse(X):
- self.info['is_sparse'] = 1
- self.info['has_missing'] = np.all(np.isfinite(cast(sparse.csr_matrix, X).data))
+ self.info["is_sparse"] = 1
+ self.info["has_missing"] = np.all(
+ np.isfinite(cast(sparse.csr_matrix, X).data)
+ )
else:
- self.info['is_sparse'] = 0
- if hasattr(X, 'iloc'):
- self.info['has_missing'] = cast(pd.DataFrame, X).isnull().values.any()
+ self.info["is_sparse"] = 0
+ if hasattr(X, "iloc"):
+ self.info["has_missing"] = cast(pd.DataFrame, X).isnull().values.any()
else:
- self.info['has_missing'] = np.all(np.isfinite(X))
+ self.info["has_missing"] = np.all(np.isfinite(X))
label_num = {
REGRESSION: 1,
BINARY_CLASSIFICATION: 2,
MULTIOUTPUT_REGRESSION: np.shape(y)[-1],
MULTICLASS_CLASSIFICATION: len(np.unique(y)),
- MULTILABEL_CLASSIFICATION: np.shape(y)[-1]
+ MULTILABEL_CLASSIFICATION: np.shape(y)[-1],
}
- self.info['label_num'] = label_num[task]
+ self.info["label_num"] = label_num[task]
- self.data['X_train'] = X
- self.data['Y_train'] = y
+ self.data["X_train"] = X
+ self.data["Y_train"] = y
if X_test is not None:
- self.data['X_test'] = X_test
+ self.data["X_test"] = X_test
if y_test is not None:
- self.data['Y_test'] = y_test
+ self.data["Y_test"] = y_test
if isinstance(feat_type, dict):
self.feat_type = feat_type
else:
- raise ValueError("Unsupported feat_type provided. We expect the user to "
- "provide a Dict[str, str] mapping from column to categorical/ "
- "numerical.")
+ raise ValueError(
+ "Unsupported feat_type provided. We expect the user to "
+ "provide a Dict[str, str] mapping from column to categorical/ "
+ "numerical."
+ )
# TODO: try to guess task type!
if len(np.shape(y)) > 2:
- raise ValueError('y must not have more than two dimensions, '
- 'but has %d.' % len(np.shape(y)))
+ raise ValueError(
+ "y must not have more than two dimensions, "
+ "but has %d." % len(np.shape(y))
+ )
if np.shape(X)[0] != np.shape(y)[0]:
- raise ValueError('X and y must have the same number of '
- 'datapoints, but have %d and %d.' % (np.shape(X)[0],
- np.shape(y)[0]))
+ raise ValueError(
+ "X and y must have the same number of "
+ "datapoints, but have %d and %d." % (np.shape(X)[0], np.shape(y)[0])
+ )
diff --git a/autosklearn/ensemble_builder.py b/autosklearn/ensemble_builder.py
deleted file mode 100644
index 53539dd293..0000000000
--- a/autosklearn/ensemble_builder.py
+++ /dev/null
@@ -1,1501 +0,0 @@
-# -*- encoding: utf-8 -*-
-import glob
-import gzip
-import math
-import numbers
-import logging.handlers
-import multiprocessing
-import os
-import pickle
-import re
-import shutil
-import time
-import traceback
-from typing import List, Optional, Tuple, Union
-import zlib
-
-import dask.distributed
-
-import numpy as np
-import pandas as pd
-import pynisher
-from smac.callbacks import IncorporateRunResultCallback
-from smac.optimizer.smbo import SMBO
-from smac.runhistory.runhistory import RunInfo, RunValue
-from smac.tae.base import StatusType
-
-from autosklearn.util.backend import Backend
-from autosklearn.constants import BINARY_CLASSIFICATION
-from autosklearn.metrics import calculate_score, calculate_loss, Scorer
-from autosklearn.ensembles.ensemble_selection import EnsembleSelection
-from autosklearn.ensembles.abstract_ensemble import AbstractEnsemble
-from autosklearn.util.logging_ import get_named_client_logger
-from autosklearn.util.parallel import preload_modules
-
-Y_ENSEMBLE = 0
-Y_VALID = 1
-Y_TEST = 2
-
-MODEL_FN_RE = r'_([0-9]*)_([0-9]*)_([0-9]{1,3}\.[0-9]*)\.npy'
-
-
-class EnsembleBuilderManager(IncorporateRunResultCallback):
- def __init__(
- self,
- start_time: float,
- time_left_for_ensembles: float,
- backend: Backend,
- dataset_name: str,
- task: int,
- metric: Scorer,
- ensemble_size: int,
- ensemble_nbest: int,
- max_models_on_disc: Union[float, int],
- seed: int,
- precision: int,
- max_iterations: Optional[int],
- read_at_most: int,
- ensemble_memory_limit: Optional[int],
- random_state: Union[int, np.random.RandomState],
- logger_port: int = logging.handlers.DEFAULT_TCP_LOGGING_PORT,
- pynisher_context: str = 'fork',
- ):
- """ SMAC callback to handle ensemble building
-
- Parameters
- ----------
- start_time: int
- the time when this job was started, to account for any latency in job allocation
- time_left_for_ensemble: int
- How much time is left for the task. Job should finish within this allocated time
- backend: util.backend.Backend
- backend to write and read files
- dataset_name: str
- name of dataset
- task_type: int
- type of ML task
- metric: str
- name of metric to compute the loss of the given predictions
- ensemble_size: int
- maximal size of ensemble (passed to autosklearn.ensemble.ensemble_selection)
- ensemble_nbest: int/float
- if int: consider only the n best prediction
- if float: consider only this fraction of the best models
- Both wrt to validation predictions
- If performance_range_threshold > 0, might return less models
- max_models_on_disc: int
- Defines the maximum number of models that are kept in the disc.
- If int, it must be greater or equal than 1, and dictates the max number of
- models to keep.
- If float, it will be interpreted as the max megabytes allowed of disc space. That
- is, if the number of ensemble candidates require more disc space than this float
- value, the worst models will be deleted to keep within this budget.
- Models and predictions of the worst-performing models will be deleted then.
- If None, the feature is disabled.
- It defines an upper bound on the models that can be used in the ensemble.
- seed: int
- random seed
- max_iterations: int
- maximal number of iterations to run this script
- (default None --> deactivated)
- precision: [16,32,64,128]
- precision of floats to read the predictions
- memory_limit: Optional[int]
- memory limit in mb. If ``None``, no memory limit is enforced.
- read_at_most: int
- read at most n new prediction files in each iteration
- logger_port: int
- port that receives logging records
- pynisher_context: str
- The multiprocessing context for pynisher. One of spawn/fork/forkserver.
-
- Returns
- -------
- List[Tuple[int, float, float, float]]:
- A list with the performance history of this ensemble, of the form
- [[pandas_timestamp, train_performance, val_performance, test_performance], ...]
- """
- self.start_time = start_time
- self.time_left_for_ensembles = time_left_for_ensembles
- self.backend = backend
- self.dataset_name = dataset_name
- self.task = task
- self.metric = metric
- self.ensemble_size = ensemble_size
- self.ensemble_nbest = ensemble_nbest
- self.max_models_on_disc = max_models_on_disc
- self.seed = seed
- self.precision = precision
- self.max_iterations = max_iterations
- self.read_at_most = read_at_most
- self.ensemble_memory_limit = ensemble_memory_limit
- self.random_state = random_state
- self.logger_port = logger_port
- self.pynisher_context = pynisher_context
-
- # Store something similar to SMAC's runhistory
- self.history = []
-
- # We only submit new ensembles when there is not an active ensemble job
- self.futures = []
-
- # The last criteria is the number of iterations
- self.iteration = 0
-
- # Keep track of when we started to know when we need to finish!
- self.start_time = time.time()
-
- def __call__(
- self,
- smbo: 'SMBO',
- run_info: RunInfo,
- result: RunValue,
- time_left: float,
- ):
- if result.status in (StatusType.STOP, StatusType.ABORT) or smbo._stop:
- return
- self.build_ensemble(smbo.tae_runner.client)
-
- def build_ensemble(
- self,
- dask_client: dask.distributed.Client,
- unit_test: bool = False
- ) -> None:
-
- # The second criteria is elapsed time
- elapsed_time = time.time() - self.start_time
-
- logger = get_named_client_logger(
- name='EnsembleBuilder',
- port=self.logger_port,
- )
-
- # First test for termination conditions
- if self.time_left_for_ensembles < elapsed_time:
- logger.info(
- "Terminate ensemble building as not time is left (run for {}s)".format(
- elapsed_time
- ),
- )
- return
- if self.max_iterations is not None and self.max_iterations <= self.iteration:
- logger.info(
- "Terminate ensemble building because of max iterations: {} of {}".format(
- self.max_iterations,
- self.iteration
- )
- )
- return
-
- if len(self.futures) != 0:
- if self.futures[0].done():
- result = self.futures.pop().result()
- if result:
- ensemble_history, self.ensemble_nbest, _, _, _ = result
- logger.debug("iteration={} @ elapsed_time={} has history={}".format(
- self.iteration,
- elapsed_time,
- ensemble_history,
- ))
- self.history.extend(ensemble_history)
-
- # Only submit new jobs if the previous ensemble job finished
- if len(self.futures) == 0:
-
- # Add the result of the run
- # On the next while iteration, no references to
- # ensemble builder object, so it should be garbage collected to
- # save memory while waiting for resources
- # Also, notice how ensemble nbest is returned, so we don't waste
- # iterations testing if the deterministic predictions size can
- # be fitted in memory
- try:
- # Submit a Dask job from this job, to properly
- # see it in the dask diagnostic dashboard
- # Notice that the forked ensemble_builder_process will
- # wait for the below function to be done
- self.futures.append(dask_client.submit(
- fit_and_return_ensemble,
- backend=self.backend,
- dataset_name=self.dataset_name,
- task_type=self.task,
- metric=self.metric,
- ensemble_size=self.ensemble_size,
- ensemble_nbest=self.ensemble_nbest,
- max_models_on_disc=self.max_models_on_disc,
- seed=self.seed,
- precision=self.precision,
- memory_limit=self.ensemble_memory_limit,
- read_at_most=self.read_at_most,
- random_state=self.random_state,
- end_at=self.start_time + self.time_left_for_ensembles,
- iteration=self.iteration,
- return_predictions=False,
- priority=100,
- pynisher_context=self.pynisher_context,
- logger_port=self.logger_port,
- unit_test=unit_test,
- ))
-
- logger.info(
- "{}/{} Started Ensemble builder job at {} for iteration {}.".format(
- # Log the client to make sure we
- # remain connected to the scheduler
- self.futures[0],
- dask_client,
- time.strftime("%Y.%m.%d-%H.%M.%S"),
- self.iteration,
- ),
- )
- self.iteration += 1
- except Exception as e:
- exception_traceback = traceback.format_exc()
- error_message = repr(e)
- logger.critical(exception_traceback)
- logger.critical(error_message)
-
-
-def fit_and_return_ensemble(
- backend: Backend,
- dataset_name: str,
- task_type: str,
- metric: Scorer,
- ensemble_size: int,
- ensemble_nbest: int,
- max_models_on_disc: Union[float, int],
- seed: int,
- precision: int,
- read_at_most: int,
- end_at: float,
- iteration: int,
- return_predictions: bool,
- pynisher_context: str,
- logger_port: int = logging.handlers.DEFAULT_TCP_LOGGING_PORT,
- unit_test: bool = False,
- memory_limit: Optional[int] = None,
- random_state: Optional[Union[int, np.random.RandomState]] = None,
-) -> Tuple[
- List[Tuple[int, float, float, float]],
- int,
- Optional[np.ndarray],
- Optional[np.ndarray],
- Optional[np.ndarray],
-]:
- """
-
- A short function to fit and create an ensemble. It is just a wrapper to easily send
- a request to dask to create an ensemble and clean the memory when finished
-
- Parameters
- ----------
- backend: util.backend.Backend
- backend to write and read files
- dataset_name: str
- name of dataset
- metric: str
- name of metric to compute the loss of the given predictions
- task_type: int
- type of ML task
- ensemble_size: int
- maximal size of ensemble (passed to autosklearn.ensemble.ensemble_selection)
- ensemble_nbest: int/float
- if int: consider only the n best prediction
- if float: consider only this fraction of the best models
- Both wrt to validation predictions
- If performance_range_threshold > 0, might return less models
- max_models_on_disc: int
- Defines the maximum number of models that are kept in the disc.
- If int, it must be greater or equal than 1, and dictates the max number of
- models to keep.
- If float, it will be interpreted as the max megabytes allowed of disc space. That
- is, if the number of ensemble candidates require more disc space than this float
- value, the worst models will be deleted to keep within this budget.
- Models and predictions of the worst-performing models will be deleted then.
- If None, the feature is disabled.
- It defines an upper bound on the models that can be used in the ensemble.
- seed: int
- random seed
- precision: [16,32,64,128]
- precision of floats to read the predictions
- read_at_most: int
- read at most n new prediction files in each iteration
- end_at: float
- At what time the job must finish. Needs to be the endtime and not the time left
- because we do not know when dask schedules the job.
- iteration: int
- The current iteration
- pynisher_context: str
- Context to use for multiprocessing, can be either fork, spawn or forkserver.
- logger_port: int = DEFAULT_TCP_LOGGING_PORT
- The port where the logging server is listening to.
- unit_test: bool = False
- Turn on unit testing mode. This currently makes fit_ensemble raise a MemoryError.
- Having this is very bad coding style, but I did not find a way to make
- unittest.mock work through the pynisher with all spawn contexts. If you know a
- better solution, please let us know by opening an issue.
- memory_limit: Optional[int] = None
- memory limit in mb. If ``None``, no memory limit is enforced.
- random_state: Optional[int | RandomState] = None
- A random state used for the ensemble selection process.
-
- Returns
- -------
- List[Tuple[int, float, float, float]]
- A list with the performance history of this ensemble, of the form
- [[pandas_timestamp, train_performance, val_performance, test_performance], ...]
-
- """
- result = EnsembleBuilder(
- backend=backend,
- dataset_name=dataset_name,
- task_type=task_type,
- metric=metric,
- ensemble_size=ensemble_size,
- ensemble_nbest=ensemble_nbest,
- max_models_on_disc=max_models_on_disc,
- seed=seed,
- precision=precision,
- memory_limit=memory_limit,
- read_at_most=read_at_most,
- random_state=random_state,
- logger_port=logger_port,
- unit_test=unit_test,
- ).run(
- end_at=end_at,
- iteration=iteration,
- return_predictions=return_predictions,
- pynisher_context=pynisher_context,
- )
- return result
-
-
-class EnsembleBuilder(object):
- def __init__(
- self,
- backend: Backend,
- dataset_name: str,
- task_type: int,
- metric: Scorer,
- ensemble_size: int = 10,
- ensemble_nbest: Union[int, float] = 100,
- max_models_on_disc: int = 100,
- performance_range_threshold: float = 0,
- seed: int = 1,
- precision: int = 32,
- memory_limit: Optional[int] = 1024,
- read_at_most: int = 5,
- logger_port: int = logging.handlers.DEFAULT_TCP_LOGGING_PORT,
- random_state: Optional[Union[int, np.random.RandomState]] = None,
- unit_test: bool = False,
- ):
- """
- Constructor
-
- Parameters
- ----------
- backend: util.backend.Backend
- backend to write and read files
- dataset_name: str
- name of dataset
- task_type: int
- type of ML task
- metric: str
- name of metric to compute the loss of the given predictions
- ensemble_size: int = 10
- maximal size of ensemble (passed to autosklearn.ensemble.ensemble_selection)
- ensemble_nbest: int | float = 100
- if int: consider only the n best prediction
- if float: consider only this fraction of the best models
- Both with respect to the validation predictions
- If performance_range_threshold > 0, might return less models
- max_models_on_disc: int = 100
- Defines the maximum number of models that are kept in the disc.
- If int, it must be greater or equal than 1, and dictates the max number of
- models to keep.
- If float, it will be interpreted as the max megabytes allowed of disc space. That
- is, if the number of ensemble candidates require more disc space than this float
- value, the worst models will be deleted to keep within this budget.
- Models and predictions of the worst-performing models will be deleted then.
- If None, the feature is disabled.
- It defines an upper bound on the models that can be used in the ensemble.
- performance_range_threshold: float = 0
- Keep only models that are better than:
- dummy + (best - dummy)*performance_range_threshold
- E.g dummy=2, best=4, thresh=0.5 --> only consider models with loss > 3
- Will at most return the minimum between ensemble_nbest models,
- and max_models_on_disc. Might return less
- seed: int = 1
- random seed that is used as part of the filename
- precision: int in [16,32,64,128] = 32
- precision of floats to read the predictions
- memory_limit: Optional[int] = 1024
- memory limit in mb. If ``None``, no memory limit is enforced.
- read_at_most: int = 5
- read at most n new prediction files in each iteration
- logger_port: int = DEFAULT_TCP_LOGGING_PORT
- port that receives logging records
- random_state: Optional[int | RandomState] = None
- An int or RandomState object used for generating the ensemble.
- unit_test: bool = False
- Turn on unit testing mode. This currently makes fit_ensemble raise a MemoryError.
- Having this is very bad coding style, but I did not find a way to make
- unittest.mock work through the pynisher with all spawn contexts. If you know a
- better solution, please let us know by opening an issue.
- """
-
- super(EnsembleBuilder, self).__init__()
-
- self.backend = backend # communication with filesystem
- self.dataset_name = dataset_name
- self.task_type = task_type
- self.metric = metric
- self.ensemble_size = ensemble_size
- self.performance_range_threshold = performance_range_threshold
-
- if isinstance(ensemble_nbest, numbers.Integral) and ensemble_nbest < 1:
- raise ValueError("Integer ensemble_nbest has to be larger 1: %s" %
- ensemble_nbest)
- elif not isinstance(ensemble_nbest, numbers.Integral):
- if ensemble_nbest < 0 or ensemble_nbest > 1:
- raise ValueError(
- "Float ensemble_nbest best has to be >= 0 and <= 1: %s" %
- ensemble_nbest)
-
- self.ensemble_nbest = ensemble_nbest
-
- # max_models_on_disc can be a float, in such case we need to
- # remember the user specified Megabytes and translate this to
- # max number of ensemble models. max_resident_models keeps the
- # maximum number of models in disc
- if max_models_on_disc is not None and max_models_on_disc < 0:
- raise ValueError(
- "max_models_on_disc has to be a positive number or None"
- )
- self.max_models_on_disc = max_models_on_disc
- self.max_resident_models = None
-
- self.seed = seed
- self.precision = precision
- self.memory_limit = memory_limit
- self.read_at_most = read_at_most
- self.random_state = random_state
- self.unit_test = unit_test
-
- # Setup the logger
- self.logger_port = logger_port
- self.logger = get_named_client_logger(
- name='EnsembleBuilder',
- port=self.logger_port,
- )
-
- if ensemble_nbest == 1:
- self.logger.debug("Behaviour depends on int/float: %s, %s (ensemble_nbest, type)" %
- (ensemble_nbest, type(ensemble_nbest)))
-
- self.start_time = 0
- self.model_fn_re = re.compile(MODEL_FN_RE)
-
- self.last_hash = None # hash of ensemble training data
- self.y_true_ensemble = None
- self.SAVE2DISC = True
-
- # already read prediction files
- # {"file name": {
- # "ens_loss": float
- # "mtime_ens": str,
- # "mtime_valid": str,
- # "mtime_test": str,
- # "seed": int,
- # "num_run": int,
- # }}
- self.read_losses = {}
- # {"file_name": {
- # Y_ENSEMBLE: np.ndarray
- # Y_VALID: np.ndarray
- # Y_TEST: np.ndarray
- # }
- # }
- self.read_preds = {}
-
- # Depending on the dataset dimensions,
- # regenerating every iteration, the predictions
- # losses for self.read_preds
- # is too computationally expensive
- # As the ensemble builder is stateless
- # (every time the ensemble builder gets resources
- # from dask, it builds this object from scratch)
- # we save the state of this dictionary to memory
- # and read it if available
- self.ensemble_memory_file = os.path.join(
- self.backend.internals_directory,
- 'ensemble_read_preds.pkl'
- )
- if os.path.exists(self.ensemble_memory_file):
- try:
- with (open(self.ensemble_memory_file, "rb")) as memory:
- self.read_preds, self.last_hash = pickle.load(memory)
- except Exception as e:
- self.logger.warning(
- "Could not load the previous iterations of ensemble_builder predictions."
- "This might impact the quality of the run. Exception={} {}".format(
- e,
- traceback.format_exc(),
- )
- )
- self.ensemble_loss_file = os.path.join(
- self.backend.internals_directory,
- 'ensemble_read_losses.pkl'
- )
- if os.path.exists(self.ensemble_loss_file):
- try:
- with (open(self.ensemble_loss_file, "rb")) as memory:
- self.read_losses = pickle.load(memory)
- except Exception as e:
- self.logger.warning(
- "Could not load the previous iterations of ensemble_builder losses."
- "This might impact the quality of the run. Exception={} {}".format(
- e,
- traceback.format_exc(),
- )
- )
-
- # hidden feature which can be activated via an environment variable. This keeps all
- # models and predictions which have ever been a candidate. This is necessary to post-hoc
- # compute the whole ensemble building trajectory.
- self._has_been_candidate = set()
-
- self.validation_performance_ = np.inf
-
- # Track the ensemble performance
- datamanager = self.backend.load_datamanager()
- self.y_valid = datamanager.data.get('Y_valid')
- self.y_test = datamanager.data.get('Y_test')
- del datamanager
- self.ensemble_history = []
-
- def run(
- self,
- iteration: int,
- pynisher_context: str,
- time_left: Optional[float] = None,
- end_at: Optional[float] = None,
- time_buffer=5,
- return_predictions: bool = False,
- ):
-
- if time_left is None and end_at is None:
- raise ValueError('Must provide either time_left or end_at.')
- elif time_left is not None and end_at is not None:
- raise ValueError('Cannot provide both time_left and end_at.')
-
- self.logger = get_named_client_logger(
- name='EnsembleBuilder',
- port=self.logger_port,
- )
-
- process_start_time = time.time()
- while True:
-
- if time_left is not None:
- time_elapsed = time.time() - process_start_time
- time_left -= time_elapsed
- else:
- current_time = time.time()
- if current_time > end_at:
- break
- else:
- time_left = end_at - current_time
-
- wall_time_in_s = int(time_left - time_buffer)
- if wall_time_in_s < 1:
- break
- context = multiprocessing.get_context(pynisher_context)
- preload_modules(context)
-
- safe_ensemble_script = pynisher.enforce_limits(
- wall_time_in_s=wall_time_in_s,
- mem_in_mb=self.memory_limit,
- logger=self.logger,
- context=context,
- )(self.main)
- safe_ensemble_script(time_left, iteration, return_predictions)
- if safe_ensemble_script.exit_status is pynisher.MemorylimitException:
- # if ensemble script died because of memory error,
- # reduce nbest to reduce memory consumption and try it again
-
- # ATTENTION: main will start from scratch; # all data structures are empty again
- try:
- os.remove(self.ensemble_memory_file)
- except: # noqa E722
- pass
-
- if isinstance(self.ensemble_nbest, numbers.Integral) and self.ensemble_nbest <= 1:
- if self.read_at_most == 1:
- self.logger.error(
- "Memory Exception -- Unable to further reduce the number of ensemble "
- "members and can no further limit the number of ensemble members "
- "loaded per iteration -- please restart Auto-sklearn with a higher "
- "value for the argument `memory_limit` (current limit is %s MB). "
- "The ensemble builder will keep running to delete files from disk in "
- "case this was enabled.", self.memory_limit
- )
- self.ensemble_nbest = 0
- else:
- self.read_at_most = 1
- self.logger.warning(
- "Memory Exception -- Unable to further reduce the number of ensemble "
- "members -- Now reducing the number of predictions per call to read "
- "at most to 1."
- )
- else:
- if isinstance(self.ensemble_nbest, numbers.Integral):
- self.ensemble_nbest = max(1, int(self.ensemble_nbest / 2))
- else:
- self.ensemble_nbest = self.ensemble_nbest / 2
- self.logger.warning("Memory Exception -- restart with "
- "less ensemble_nbest: %d" % self.ensemble_nbest)
- return [], self.ensemble_nbest, None, None, None
- else:
- return safe_ensemble_script.result
-
- return [], self.ensemble_nbest, None, None, None
-
- def main(self, time_left, iteration, return_predictions):
-
- # Pynisher jobs inside dask 'forget'
- # the logger configuration. So we have to set it up
- # accordingly
- self.logger = get_named_client_logger(
- name='EnsembleBuilder',
- port=self.logger_port,
- )
-
- self.start_time = time.time()
- train_pred, valid_pred, test_pred = None, None, None
-
- used_time = time.time() - self.start_time
- self.logger.debug(
- 'Starting iteration %d, time left: %f',
- iteration,
- time_left - used_time,
- )
-
- # populates self.read_preds and self.read_losses
- if not self.compute_loss_per_model():
- if return_predictions:
- return self.ensemble_history, self.ensemble_nbest, train_pred, valid_pred, test_pred
- else:
- return self.ensemble_history, self.ensemble_nbest, None, None, None
-
- # Only the models with the n_best predictions are candidates
- # to be in the ensemble
- candidate_models = self.get_n_best_preds()
- if not candidate_models: # no candidates yet
- if return_predictions:
- return self.ensemble_history, self.ensemble_nbest, train_pred, valid_pred, test_pred
- else:
- return self.ensemble_history, self.ensemble_nbest, None, None, None
-
- # populates predictions in self.read_preds
- # reduces selected models if file reading failed
- n_sel_valid, n_sel_test = self. \
- get_valid_test_preds(selected_keys=candidate_models)
-
- # If valid/test predictions loaded, then reduce candidate models to this set
- if len(n_sel_test) != 0 and len(n_sel_valid) != 0 \
- and len(set(n_sel_valid).intersection(set(n_sel_test))) == 0:
- # Both n_sel_* have entries, but there is no overlap, this is critical
- self.logger.error("n_sel_valid and n_sel_test are not empty, but do not overlap")
- if return_predictions:
- return self.ensemble_history, self.ensemble_nbest, train_pred, valid_pred, test_pred
- else:
- return self.ensemble_history, self.ensemble_nbest, None, None, None
-
- # If any of n_sel_* is not empty and overlaps with candidate_models,
- # then ensure candidate_models AND n_sel_test are sorted the same
- candidate_models_set = set(candidate_models)
- if candidate_models_set.intersection(n_sel_valid).intersection(n_sel_test):
- candidate_models = sorted(list(candidate_models_set.intersection(
- n_sel_valid).intersection(n_sel_test)))
- n_sel_test = candidate_models
- n_sel_valid = candidate_models
- elif candidate_models_set.intersection(n_sel_valid):
- candidate_models = sorted(list(candidate_models_set.intersection(
- n_sel_valid)))
- n_sel_valid = candidate_models
- elif candidate_models_set.intersection(n_sel_test):
- candidate_models = sorted(list(candidate_models_set.intersection(
- n_sel_test)))
- n_sel_test = candidate_models
- else:
- # This has to be the case
- n_sel_test = []
- n_sel_valid = []
-
- if os.environ.get('ENSEMBLE_KEEP_ALL_CANDIDATES'):
- for candidate in candidate_models:
- self._has_been_candidate.add(candidate)
-
- # train ensemble
- ensemble = self.fit_ensemble(selected_keys=candidate_models)
-
- # Save the ensemble for later use in the main auto-sklearn module!
- if ensemble is not None and self.SAVE2DISC:
- self.backend.save_ensemble(ensemble, iteration, self.seed)
-
- # Delete files of non-candidate models - can only be done after fitting the ensemble and
- # saving it to disc so we do not accidentally delete models in the previous ensemble
- if self.max_resident_models is not None:
- self._delete_excess_models(selected_keys=candidate_models)
-
- # Save the read losses status for the next iteration
- with open(self.ensemble_loss_file, "wb") as memory:
- pickle.dump(self.read_losses, memory)
-
- if ensemble is not None:
- train_pred = self.predict(set_="train",
- ensemble=ensemble,
- selected_keys=candidate_models,
- n_preds=len(candidate_models),
- index_run=iteration)
- # We can't use candidate_models here, as n_sel_* might be empty
- valid_pred = self.predict(set_="valid",
- ensemble=ensemble,
- selected_keys=n_sel_valid,
- n_preds=len(candidate_models),
- index_run=iteration)
- # TODO if predictions fails, build the model again during the
- # next iteration!
- test_pred = self.predict(set_="test",
- ensemble=ensemble,
- selected_keys=n_sel_test,
- n_preds=len(candidate_models),
- index_run=iteration)
-
- # Add a score to run history to see ensemble progress
- self._add_ensemble_trajectory(
- train_pred,
- valid_pred,
- test_pred
- )
-
- # The loaded predictions and the hash can only be saved after the ensemble has been
- # built, because the hash is computed during the construction of the ensemble
- with open(self.ensemble_memory_file, "wb") as memory:
- pickle.dump((self.read_preds, self.last_hash), memory)
-
- if return_predictions:
- return self.ensemble_history, self.ensemble_nbest, train_pred, valid_pred, test_pred
- else:
- return self.ensemble_history, self.ensemble_nbest, None, None, None
-
- def get_disk_consumption(self, pred_path):
- """
- gets the cost of a model being on disc
- """
-
- match = self.model_fn_re.search(pred_path)
- if not match:
- raise ValueError("Invalid path format %s" % pred_path)
- _seed = int(match.group(1))
- _num_run = int(match.group(2))
- _budget = float(match.group(3))
-
- stored_files_for_run = os.listdir(
- self.backend.get_numrun_directory(_seed, _num_run, _budget))
- stored_files_for_run = [
- os.path.join(self.backend.get_numrun_directory(_seed, _num_run, _budget), file_name)
- for file_name in stored_files_for_run]
- this_model_cost = sum([os.path.getsize(path) for path in stored_files_for_run])
-
- # get the megabytes
- return round(this_model_cost / math.pow(1024, 2), 2)
-
- def compute_loss_per_model(self):
- """
- Compute the loss of the predictions on ensemble building data set;
- populates self.read_preds and self.read_losses
- """
-
- self.logger.debug("Read ensemble data set predictions")
-
- if self.y_true_ensemble is None:
- try:
- self.y_true_ensemble = self.backend.load_targets_ensemble()
- except FileNotFoundError:
- self.logger.debug(
- "Could not find true targets on ensemble data set: %s",
- traceback.format_exc(),
- )
- return False
-
- pred_path = os.path.join(
- glob.escape(self.backend.get_runs_directory()),
- '%d_*_*' % self.seed,
- 'predictions_ensemble_%s_*_*.npy*' % self.seed,
- )
- y_ens_files = glob.glob(pred_path)
- y_ens_files = [y_ens_file for y_ens_file in y_ens_files
- if y_ens_file.endswith('.npy') or y_ens_file.endswith('.npy.gz')]
- self.y_ens_files = y_ens_files
- # no validation predictions so far -- no files
- if len(self.y_ens_files) == 0:
- self.logger.debug("Found no prediction files on ensemble data set:"
- " %s" % pred_path)
- return False
-
- # First sort files chronologically
- to_read = []
- for y_ens_fn in self.y_ens_files:
- match = self.model_fn_re.search(y_ens_fn)
- _seed = int(match.group(1))
- _num_run = int(match.group(2))
- _budget = float(match.group(3))
- mtime = os.path.getmtime(y_ens_fn)
-
- to_read.append([y_ens_fn, match, _seed, _num_run, _budget, mtime])
-
- n_read_files = 0
- # Now read file wrt to num_run
- for y_ens_fn, match, _seed, _num_run, _budget, mtime in \
- sorted(to_read, key=lambda x: x[5]):
- if self.read_at_most and n_read_files >= self.read_at_most:
- # limit the number of files that will be read
- # to limit memory consumption
- break
-
- if not y_ens_fn.endswith(".npy") and not y_ens_fn.endswith(".npy.gz"):
- self.logger.info('Error loading file (not .npy or .npy.gz): %s', y_ens_fn)
- continue
-
- if not self.read_losses.get(y_ens_fn):
- self.read_losses[y_ens_fn] = {
- "ens_loss": np.inf,
- "mtime_ens": 0,
- "mtime_valid": 0,
- "mtime_test": 0,
- "seed": _seed,
- "num_run": _num_run,
- "budget": _budget,
- "disc_space_cost_mb": None,
- # Lazy keys so far:
- # 0 - not loaded
- # 1 - loaded and in memory
- # 2 - loaded but dropped again
- # 3 - deleted from disk due to space constraints
- "loaded": 0
- }
- if not self.read_preds.get(y_ens_fn):
- self.read_preds[y_ens_fn] = {
- Y_ENSEMBLE: None,
- Y_VALID: None,
- Y_TEST: None,
- }
-
- if self.read_losses[y_ens_fn]["mtime_ens"] == mtime:
- # same time stamp; nothing changed;
- continue
-
- # actually read the predictions and compute their respective loss
- try:
- y_ensemble = self._read_np_fn(y_ens_fn)
- loss = calculate_loss(solution=self.y_true_ensemble,
- prediction=y_ensemble,
- task_type=self.task_type,
- metric=self.metric,
- scoring_functions=None)
-
- if np.isfinite(self.read_losses[y_ens_fn]["ens_loss"]):
- self.logger.debug(
- 'Changing ensemble loss for file %s from %f to %f '
- 'because file modification time changed? %f - %f',
- y_ens_fn,
- self.read_losses[y_ens_fn]["ens_loss"],
- loss,
- self.read_losses[y_ens_fn]["mtime_ens"],
- os.path.getmtime(y_ens_fn),
- )
-
- self.read_losses[y_ens_fn]["ens_loss"] = loss
-
- # It is not needed to create the object here
- # To save memory, we just compute the loss.
- self.read_losses[y_ens_fn]["mtime_ens"] = os.path.getmtime(y_ens_fn)
- self.read_losses[y_ens_fn]["loaded"] = 2
- self.read_losses[y_ens_fn]["disc_space_cost_mb"] = self.get_disk_consumption(
- y_ens_fn
- )
-
- n_read_files += 1
-
- except Exception:
- self.logger.warning(
- 'Error loading %s: %s',
- y_ens_fn,
- traceback.format_exc(),
- )
- self.read_losses[y_ens_fn]["ens_loss"] = np.inf
-
- self.logger.debug(
- 'Done reading %d new prediction files. Loaded %d predictions in '
- 'total.',
- n_read_files,
- np.sum([pred["loaded"] > 0 for pred in self.read_losses.values()])
- )
- return True
-
- def get_n_best_preds(self):
- """
- get best n predictions (i.e., keys of self.read_losses)
- according to the loss on the "ensemble set"
- n: self.ensemble_nbest
-
- Side effects:
- ->Define the n-best models to use in ensemble
- ->Only the best models are loaded
- ->Any model that is not best is candidate to deletion
- if max models in disc is exceeded.
- """
-
- sorted_keys = self._get_list_of_sorted_preds()
-
- # number of models available
- num_keys = len(sorted_keys)
- # remove all that are at most as good as random
- # note: dummy model must have run_id=1 (there is no run_id=0)
- dummy_losses = list(filter(lambda x: x[2] == 1, sorted_keys))
- # number of dummy models
- num_dummy = len(dummy_losses)
- dummy_loss = dummy_losses[0]
- self.logger.debug("Use %f as dummy loss" % dummy_loss[1])
-
- # sorted_keys looks like: (k, v["ens_loss"], v["num_run"])
- # On position 1 we have the loss of a minimization problem.
- # keep only the predictions with a loss smaller than the dummy
- # prediction
- sorted_keys = filter(lambda x: x[1] < dummy_loss[1], sorted_keys)
-
- # remove Dummy Classifier
- sorted_keys = list(filter(lambda x: x[2] > 1, sorted_keys))
- if not sorted_keys:
- # no model left; try to use dummy loss (num_run==0)
- # log warning when there are other models but not better than dummy model
- if num_keys > num_dummy:
- self.logger.warning("No models better than random - using Dummy loss!"
- "Number of models besides current dummy model: %d. "
- "Number of dummy models: %d",
- num_keys - 1,
- num_dummy)
- sorted_keys = [
- (k, v["ens_loss"], v["num_run"]) for k, v in self.read_losses.items()
- if v["seed"] == self.seed and v["num_run"] == 1
- ]
- # reload predictions if losses changed over time and a model is
- # considered to be in the top models again!
- if not isinstance(self.ensemble_nbest, numbers.Integral):
- # Transform to number of models to keep. Keep at least one
- keep_nbest = max(1, min(len(sorted_keys),
- int(len(sorted_keys) * self.ensemble_nbest)))
- self.logger.debug(
- "Library pruning: using only top %f percent of the models for ensemble "
- "(%d out of %d)",
- self.ensemble_nbest * 100, keep_nbest, len(sorted_keys)
- )
- else:
- # Keep only at most ensemble_nbest
- keep_nbest = min(self.ensemble_nbest, len(sorted_keys))
- self.logger.debug("Library Pruning: using for ensemble only "
- " %d (out of %d) models" % (keep_nbest, len(sorted_keys)))
-
- # If max_models_on_disc is None, do nothing
- # One can only read at most max_models_on_disc models
- if self.max_models_on_disc is not None:
- if not isinstance(self.max_models_on_disc, numbers.Integral):
- consumption = [
- [
- v["ens_loss"],
- v["disc_space_cost_mb"],
- ] for v in self.read_losses.values() if v["disc_space_cost_mb"] is not None
- ]
- max_consumption = max(c[1] for c in consumption)
-
- # We are pessimistic with the consumption limit indicated by
- # max_models_on_disc by 1 model. Such model is assumed to spend
- # max_consumption megabytes
- if (sum(c[1] for c in consumption) + max_consumption) > self.max_models_on_disc:
-
- # just leave the best -- smaller is better!
- # This list is in descending order, to preserve the best models
- sorted_cum_consumption = np.cumsum([
- c[1] for c in list(sorted(consumption))
- ]) + max_consumption
- max_models = np.argmax(sorted_cum_consumption > self.max_models_on_disc)
-
- # Make sure that at least 1 model survives
- self.max_resident_models = max(1, max_models)
- self.logger.warning(
- "Limiting num of models via float max_models_on_disc={}"
- " as accumulated={} worst={} num_models={}".format(
- self.max_models_on_disc,
- (sum(c[1] for c in consumption) + max_consumption),
- max_consumption,
- self.max_resident_models
- )
- )
- else:
- self.max_resident_models = None
- else:
- self.max_resident_models = self.max_models_on_disc
-
- if self.max_resident_models is not None and keep_nbest > self.max_resident_models:
- self.logger.debug(
- "Restricting the number of models to %d instead of %d due to argument "
- "max_models_on_disc",
- self.max_resident_models, keep_nbest,
- )
- keep_nbest = self.max_resident_models
-
- # consider performance_range_threshold
- if self.performance_range_threshold > 0:
- best_loss = sorted_keys[0][1]
- worst_loss = dummy_loss[1]
- worst_loss -= (worst_loss - best_loss) * self.performance_range_threshold
- if sorted_keys[keep_nbest - 1][1] > worst_loss:
- # We can further reduce number of models
- # since worst model is worse than thresh
- for i in range(0, keep_nbest):
- # Look at most at keep_nbest models,
- # but always keep at least one model
- current_loss = sorted_keys[i][1]
- if current_loss >= worst_loss:
- self.logger.debug("Dynamic Performance range: "
- "Further reduce from %d to %d models",
- keep_nbest, max(1, i))
- keep_nbest = max(1, i)
- break
- ensemble_n_best = keep_nbest
-
- # reduce to keys
- sorted_keys = list(map(lambda x: x[0], sorted_keys))
-
- # remove loaded predictions for non-winning models
- for k in sorted_keys[ensemble_n_best:]:
- if k in self.read_preds:
- self.read_preds[k][Y_ENSEMBLE] = None
- self.read_preds[k][Y_VALID] = None
- self.read_preds[k][Y_TEST] = None
- if self.read_losses[k]['loaded'] == 1:
- self.logger.debug(
- 'Dropping model %s (%d,%d) with loss %f.',
- k,
- self.read_losses[k]['seed'],
- self.read_losses[k]['num_run'],
- self.read_losses[k]['ens_loss'],
- )
- self.read_losses[k]['loaded'] = 2
-
- # Load the predictions for the winning
- for k in sorted_keys[:ensemble_n_best]:
- if (
- (
- k not in self.read_preds or
- self.read_preds[k][Y_ENSEMBLE] is None
- )
- and self.read_losses[k]['loaded'] != 3
- ):
- self.read_preds[k][Y_ENSEMBLE] = self._read_np_fn(k)
- # No need to load valid and test here because they are loaded
- # only if the model ends up in the ensemble
- self.read_losses[k]['loaded'] = 1
-
- # return keys of self.read_losses with lowest losses
- return sorted_keys[:ensemble_n_best]
-
- def get_valid_test_preds(self, selected_keys: List[str]) -> Tuple[List[str], List[str]]:
- """
- get valid and test predictions from disc
- and store them in self.read_preds
-
- Parameters
- ---------
- selected_keys: list
- list of selected keys of self.read_preds
-
- Return
- ------
- success_keys:
- all keys in selected keys for which we could read the valid and
- test predictions
- """
- success_keys_valid = []
- success_keys_test = []
-
- for k in selected_keys:
- valid_fn = glob.glob(
- os.path.join(
- glob.escape(self.backend.get_runs_directory()),
- '%d_%d_%s' % (
- self.read_losses[k]["seed"],
- self.read_losses[k]["num_run"],
- self.read_losses[k]["budget"],
- ),
- 'predictions_valid_%d_%d_%s.npy*' % (
- self.read_losses[k]["seed"],
- self.read_losses[k]["num_run"],
- self.read_losses[k]["budget"],
- )
- )
- )
- valid_fn = [vfn for vfn in valid_fn if vfn.endswith('.npy') or vfn.endswith('.npy.gz')]
- test_fn = glob.glob(
- os.path.join(
- glob.escape(self.backend.get_runs_directory()),
- '%d_%d_%s' % (
- self.read_losses[k]["seed"],
- self.read_losses[k]["num_run"],
- self.read_losses[k]["budget"],
- ),
- 'predictions_test_%d_%d_%s.npy*' % (
- self.read_losses[k]["seed"],
- self.read_losses[k]["num_run"],
- self.read_losses[k]["budget"]
- )
- )
- )
- test_fn = [tfn for tfn in test_fn if tfn.endswith('.npy') or tfn.endswith('.npy.gz')]
-
- if len(valid_fn) == 0:
- # self.logger.debug("Not found validation prediction file "
- # "(although ensemble predictions available): "
- # "%s" % valid_fn)
- pass
- else:
- valid_fn = valid_fn[0]
- if (
- self.read_losses[k]["mtime_valid"] == os.path.getmtime(valid_fn)
- and k in self.read_preds
- and self.read_preds[k][Y_VALID] is not None
- ):
- success_keys_valid.append(k)
- continue
- try:
- y_valid = self._read_np_fn(valid_fn)
- self.read_preds[k][Y_VALID] = y_valid
- success_keys_valid.append(k)
- self.read_losses[k]["mtime_valid"] = os.path.getmtime(valid_fn)
- except Exception:
- self.logger.warning('Error loading %s: %s',
- valid_fn, traceback.format_exc())
-
- if len(test_fn) == 0:
- # self.logger.debug("Not found test prediction file (although "
- # "ensemble predictions available):%s" %
- # test_fn)
- pass
- else:
- test_fn = test_fn[0]
- if (
- self.read_losses[k]["mtime_test"] == os.path.getmtime(test_fn)
- and k in self.read_preds
- and self.read_preds[k][Y_TEST] is not None
- ):
- success_keys_test.append(k)
- continue
- try:
- y_test = self._read_np_fn(test_fn)
- self.read_preds[k][Y_TEST] = y_test
- success_keys_test.append(k)
- self.read_losses[k]["mtime_test"] = os.path.getmtime(test_fn)
- except Exception:
- self.logger.warning('Error loading %s: %s',
- test_fn, traceback.format_exc())
-
- return success_keys_valid, success_keys_test
-
- def fit_ensemble(self, selected_keys: list):
- """
- fit ensemble
-
- Parameters
- ---------
- selected_keys: list
- list of selected keys of self.read_losses
-
- Returns
- -------
- ensemble: EnsembleSelection
- trained Ensemble
- """
-
- if self.unit_test:
- raise MemoryError()
-
- predictions_train = [self.read_preds[k][Y_ENSEMBLE] for k in selected_keys]
- include_num_runs = [
- (
- self.read_losses[k]["seed"],
- self.read_losses[k]["num_run"],
- self.read_losses[k]["budget"],
- )
- for k in selected_keys]
-
- # check hash if ensemble training data changed
- current_hash = "".join([
- str(zlib.adler32(predictions_train[i].data.tobytes()))
- for i in range(len(predictions_train))
- ])
- if self.last_hash == current_hash:
- self.logger.debug(
- "No new model predictions selected -- skip ensemble building "
- "-- current performance: %f",
- self.validation_performance_,
- )
-
- return None
- self.last_hash = current_hash
-
- ensemble = EnsembleSelection(
- ensemble_size=self.ensemble_size,
- task_type=self.task_type,
- metric=self.metric,
- random_state=self.random_state,
- )
-
- try:
- self.logger.debug(
- "Fitting the ensemble on %d models.",
- len(predictions_train),
- )
- start_time = time.time()
- ensemble.fit(predictions_train, self.y_true_ensemble,
- include_num_runs)
- end_time = time.time()
- self.logger.debug(
- "Fitting the ensemble took %.2f seconds.",
- end_time - start_time,
- )
- self.logger.info(ensemble)
- self.validation_performance_ = min(
- self.validation_performance_,
- ensemble.get_validation_performance(),
- )
-
- except ValueError:
- self.logger.error('Caught ValueError: %s', traceback.format_exc())
- return None
- except IndexError:
- self.logger.error('Caught IndexError: %s' + traceback.format_exc())
- return None
- finally:
- # Explicitly free memory
- del predictions_train
-
- return ensemble
-
- def predict(self, set_: str,
- ensemble: AbstractEnsemble,
- selected_keys: list,
- n_preds: int,
- index_run: int):
- """
- save preditions on ensemble, validation and test data on disc
-
- Parameters
- ----------
- set_: ["valid","test"]
- data split name
- ensemble: EnsembleSelection
- trained Ensemble
- selected_keys: list
- list of selected keys of self.read_losses
- n_preds: int
- number of prediction models used for ensemble building
- same number of predictions on valid and test are necessary
- index_run: int
- n-th time that ensemble predictions are written to disc
-
- Return
- ------
- y: np.ndarray
- """
- self.logger.debug("Predicting the %s set with the ensemble!", set_)
-
- if set_ == 'valid':
- pred_set = Y_VALID
- elif set_ == 'test':
- pred_set = Y_TEST
- else:
- pred_set = Y_ENSEMBLE
- predictions = [self.read_preds[k][pred_set] for k in selected_keys]
-
- if n_preds == len(predictions):
- y = ensemble.predict(predictions)
- if self.task_type == BINARY_CLASSIFICATION:
- y = y[:, 1]
- return y
- else:
- self.logger.info(
- "Found inconsistent number of predictions and models (%d vs "
- "%d) for subset %s",
- len(predictions),
- n_preds,
- set_,
- )
- return None
-
- def _add_ensemble_trajectory(self, train_pred, valid_pred, test_pred):
- """
- Records a snapshot of how the performance look at a given training
- time.
-
- Parameters
- ----------
- ensemble: EnsembleSelection
- The ensemble selection object to record
- valid_pred: np.ndarray
- The predictions on the validation set using ensemble
- test_pred: np.ndarray
- The predictions on the test set using ensemble
-
- """
- if self.task_type == BINARY_CLASSIFICATION:
- if len(train_pred.shape) == 1 or train_pred.shape[1] == 1:
- train_pred = np.vstack(
- ((1 - train_pred).reshape((1, -1)), train_pred.reshape((1, -1)))
- ).transpose()
- if valid_pred is not None and (len(valid_pred.shape) == 1 or valid_pred.shape[1] == 1):
- valid_pred = np.vstack(
- ((1 - valid_pred).reshape((1, -1)), valid_pred.reshape((1, -1)))
- ).transpose()
- if test_pred is not None and (len(test_pred.shape) == 1 or test_pred.shape[1] == 1):
- test_pred = np.vstack(
- ((1 - test_pred).reshape((1, -1)), test_pred.reshape((1, -1)))
- ).transpose()
-
- performance_stamp = {
- 'Timestamp': pd.Timestamp.now(),
- 'ensemble_optimization_score': calculate_score(
- solution=self.y_true_ensemble,
- prediction=train_pred,
- task_type=self.task_type,
- metric=self.metric,
- scoring_functions=None
- )
- }
- if valid_pred is not None:
- # TODO: valid_pred are a legacy from competition manager
- # and this if never happens. Re-evaluate Y_valid support
- performance_stamp['ensemble_val_score'] = calculate_score(
- solution=self.y_valid,
- prediction=valid_pred,
- task_type=self.task_type,
- metric=self.metric,
- scoring_functions=None
- )
-
- # In case test_pred was provided
- if test_pred is not None:
- performance_stamp['ensemble_test_score'] = calculate_score(
- solution=self.y_test,
- prediction=test_pred,
- task_type=self.task_type,
- metric=self.metric,
- scoring_functions=None
- )
-
- self.ensemble_history.append(performance_stamp)
-
- def _get_list_of_sorted_preds(self):
- """
- Returns a list of sorted predictions in descending order
- Losses are taken from self.read_losses.
-
- Parameters
- ----------
- None
-
- Return
- ------
- sorted_keys: list
- """
- # Sort by loss - smaller is better!
- sorted_keys = list(sorted(
- [
- (k, v["ens_loss"], v["num_run"])
- for k, v in self.read_losses.items()
- ],
- # Sort by loss as priority 1 and then by num_run on a ascending order
- # We want small num_run first
- key=lambda x: (x[1], x[2]),
- ))
- return sorted_keys
-
- def _delete_excess_models(self, selected_keys: List[str]):
- """
- Deletes models excess models on disc. self.max_models_on_disc
- defines the upper limit on how many models to keep.
- Any additional model with a worst loss than the top
- self.max_models_on_disc is deleted.
-
- """
-
- # Loop through the files currently in the directory
- for pred_path in self.y_ens_files:
-
- # Do not delete candidates
- if pred_path in selected_keys:
- continue
-
- if pred_path in self._has_been_candidate:
- continue
-
- match = self.model_fn_re.search(pred_path)
- _seed = int(match.group(1))
- _num_run = int(match.group(2))
- _budget = float(match.group(3))
-
- # Do not delete the dummy prediction
- if _num_run == 1:
- continue
-
- numrun_dir = self.backend.get_numrun_directory(_seed, _num_run, _budget)
- try:
- os.rename(numrun_dir, numrun_dir + '.old')
- shutil.rmtree(numrun_dir + '.old')
- self.logger.info("Deleted files of non-candidate model %s", pred_path)
- self.read_losses[pred_path]["disc_space_cost_mb"] = None
- self.read_losses[pred_path]["loaded"] = 3
- self.read_losses[pred_path]["ens_loss"] = np.inf
- except Exception as e:
- self.logger.error(
- "Failed to delete files of non-candidate model %s due"
- " to error %s", pred_path, e
- )
-
- def _read_np_fn(self, path):
-
- # Support for string precision
- if isinstance(self.precision, str):
- precision = int(self.precision)
- self.logger.warning("Interpreted str-precision as {}".format(
- precision
- ))
- else:
- precision = self.precision
-
- if path.endswith("gz"):
- open_method = gzip.open
- elif path.endswith("npy"):
- open_method = open
- else:
- raise ValueError("Unknown filetype %s" % path)
- with open_method(path, 'rb') as fp:
- if precision == 16:
- predictions = np.load(fp, allow_pickle=True).astype(dtype=np.float16)
- elif precision == 32:
- predictions = np.load(fp, allow_pickle=True).astype(dtype=np.float32)
- elif precision == 64:
- predictions = np.load(fp, allow_pickle=True).astype(dtype=np.float64)
- else:
- predictions = np.load(fp, allow_pickle=True)
- return predictions
diff --git a/autosklearn/ensemble_building/__init__.py b/autosklearn/ensemble_building/__init__.py
new file mode 100644
index 0000000000..4c63165e1b
--- /dev/null
+++ b/autosklearn/ensemble_building/__init__.py
@@ -0,0 +1,5 @@
+from autosklearn.ensemble_building.builder import EnsembleBuilder
+from autosklearn.ensemble_building.manager import EnsembleBuilderManager
+from autosklearn.ensemble_building.run import Run
+
+__all__ = ["EnsembleBuilder", "EnsembleBuilderManager", "Run"]
diff --git a/autosklearn/ensemble_building/builder.py b/autosklearn/ensemble_building/builder.py
new file mode 100644
index 0000000000..50f69eb35a
--- /dev/null
+++ b/autosklearn/ensemble_building/builder.py
@@ -0,0 +1,1107 @@
+from __future__ import annotations
+
+from typing import Any, Iterable, Mapping, Sequence, Type, cast
+
+import logging.handlers
+import multiprocessing
+import numbers
+import os
+import pickle
+import time
+import traceback
+from itertools import accumulate
+from pathlib import Path
+
+import numpy as np
+import pandas as pd
+import pynisher
+from sklearn.utils.validation import check_random_state
+
+from autosklearn.automl_common.common.utils.backend import Backend
+from autosklearn.data.validation import SUPPORTED_FEAT_TYPES
+from autosklearn.data.xy_data_manager import XYDataManager
+from autosklearn.ensemble_building.run import Run, RunID
+from autosklearn.ensembles.abstract_ensemble import AbstractEnsemble
+from autosklearn.ensembles.ensemble_selection import EnsembleSelection
+from autosklearn.metrics import Scorer, calculate_loss, calculate_scores
+from autosklearn.util.disk import rmtree
+from autosklearn.util.functional import cut, findwhere, roundrobin, split
+from autosklearn.util.logging_ import get_named_client_logger
+from autosklearn.util.parallel import preload_modules
+
+CANDIDATES_FILENAME = "previous_ensemble_building_candidates.pkl"
+
+
+class EnsembleBuilder:
+ """Builds ensembles out of runs that exist in the Backend
+
+ This is used by EnsembleBuilderManager and created in a dask-client
+ every time a run finishes and there is currently no EnsembleBuilder active.
+ """
+
+ def __init__(
+ self,
+ backend: Backend,
+ dataset_name: str,
+ task_type: int,
+ metrics: Sequence[Scorer],
+ ensemble_class: Type[AbstractEnsemble] = EnsembleSelection,
+ ensemble_kwargs: Mapping[str, Any] | None = None,
+ ensemble_nbest: int | float = 50,
+ max_models_on_disc: int | float | None = 100,
+ seed: int = 1,
+ precision: int = 32,
+ memory_limit: int | None = 1024,
+ read_at_most: int | None = None,
+ logger_port: int = logging.handlers.DEFAULT_TCP_LOGGING_PORT,
+ random_state: int | np.random.RandomState | None = None,
+ ):
+ """
+ Parameters
+ ----------
+ backend: Backend
+ backend to write and read files
+
+ dataset_name: str
+ name of dataset
+
+ task_type: int
+ type of ML task
+
+ metrics: Sequence[Scorer]
+ Metrics to optimize the ensemble for. These must be non-duplicated.
+
+ ensemble_class: Type[AbstractEnsemble]
+ Implementation of the ensemble algorithm.
+
+ ensemble_kwargs: Mapping[str, Any] | None
+ Arguments passed to the constructor of the ensemble algorithm.
+
+ ensemble_nbest: int | float = 50
+
+ * int: consider only the n best prediction (> 0)
+
+ * float: consider only this fraction of the best, between (0, 1)
+
+ Both with respect to the validation predictions.
+
+ max_models_on_disc: int | float | None = 100
+ Defines the maximum number of models that are kept in the disc.
+ It defines an upper bound on the models that can be used in the ensemble.
+
+ * int: and dictates the max number of models to keep. (>= 1)
+
+ * float: it will be interpreted as the max megabytes allowed of disc space.
+ If the number of ensemble candidates require more disc space than
+ this float value, the worst models are deleted to keep within this budget.
+ Models and predictions of the worst-performing models will be deleted then.
+
+ * None: the feature is disabled.
+
+ seed: int = 1
+ random seed that is used as part of the filename
+
+ precision: int [16 | 32 | 64 | 128] = 32
+ precision of floats to read the predictions
+
+ memory_limit: int | None = 1024
+ memory limit in mb. If ``None``, no memory limit is enforced.
+
+ read_at_most: int | None = None
+ read at most n new prediction files in each iteration. If `None`, will read
+ the predictions and calculate losses for all runs that require it.
+
+
+ logger_port: int = DEFAULT_TCP_LOGGING_PORT
+ port that receives logging records
+
+ random_state: int | RandomState | None = None
+ An int or RandomState object used for generating the ensemble.
+ """
+ if isinstance(ensemble_nbest, int) and ensemble_nbest < 1:
+ raise ValueError(f"int ensemble_nbest ({ensemble_nbest}) must be (>1)")
+
+ if isinstance(ensemble_nbest, float) and not (0 <= ensemble_nbest <= 1):
+ raise ValueError(f"float ensemble_nbest ({ensemble_nbest}) not in (0,1)")
+
+ if max_models_on_disc is not None and max_models_on_disc < 0:
+ raise ValueError("max_models_on_disc must be positive or None")
+
+ if read_at_most is not None and (read_at_most < 1 or read_at_most == np.inf):
+ raise ValueError("Read at most must be int greater than 1 or None")
+
+ # Setup the logger
+ self.logger = get_named_client_logger(name="EnsembleBuilder", port=logger_port)
+ self.logger_port = logger_port
+
+ # Log the behaviour
+ if ensemble_nbest == 1:
+ t = type(ensemble_nbest)
+ self.logger.debug(f"Using behaviour when {t} for {ensemble_nbest}:{t}")
+
+ self.seed = seed
+ self.metrics = metrics
+ self.backend = backend
+ self.precision = precision
+ self.task_type = task_type
+ self.memory_limit = memory_limit
+ self.read_at_most = read_at_most
+ self.random_state = check_random_state(random_state)
+ self.dataset_name = dataset_name
+ self.ensemble_class = ensemble_class
+ self.ensemble_kwargs = ensemble_kwargs
+ self.ensemble_nbest = ensemble_nbest
+
+ # Decide if self.max_models_on_disc is a memory limit or model limit
+ self.max_models_on_disc: int | None = None
+ self.model_memory_limit: float | None = None
+
+ if isinstance(max_models_on_disc, int):
+ self.max_models_on_disc = max_models_on_disc
+ elif isinstance(max_models_on_disc, float):
+ self.model_memory_limit = max_models_on_disc
+
+ # The starting time of the procedure
+ self.start_time: float = 0.0
+
+ # Track the ensemble performance
+ self.ensemble_history: list[dict[str, Any]] = []
+
+ # Keep running knowledge of its validation performance
+ self.validation_performance_ = np.inf
+
+ # Data we may need
+ # TODO: The test data is needlessly loaded but automl_common has no concept of
+ # these and is perhaps too rigid
+ datamanager: XYDataManager = self.backend.load_datamanager()
+ self._X_test: SUPPORTED_FEAT_TYPES | None = datamanager.data.get("X_test", None)
+ self._y_test: np.ndarray | None = datamanager.data.get("Y_test", None)
+ self._X_ensemble: SUPPORTED_FEAT_TYPES | None = None
+ self._y_ensemble: np.ndarray | None = None
+
+ @property
+ def previous_candidates_path(self) -> Path:
+ """Path to the cached losses we store between runs"""
+ return Path(self.backend.internals_directory) / CANDIDATES_FILENAME
+
+ def previous_candidates(self) -> dict[RunID, Run]:
+ """Load any previous candidates that were saved from previous runs
+
+ Returns
+ -------
+ dict[RunID, Run]
+ A dictionary from RunId's to the previous candidates
+ """
+ if self.previous_candidates_path.exists():
+ with self.previous_candidates_path.open("rb") as f:
+ return pickle.load(f)
+ else:
+ return {}
+
+ def available_runs(self) -> dict[RunID, Run]:
+ """Get a dictionary of all available runs on the filesystem
+
+ Returns
+ -------
+ dict[RunID, Run]
+ A dictionary from RunId's to the available runs
+ """
+ runs_dir = Path(self.backend.get_runs_directory())
+ runs = iter(Run(path=dir) for dir in runs_dir.iterdir() if Run.valid(dir))
+ return {run.id: run for run in runs}
+
+ def targets(self, kind: str = "ensemble") -> np.ndarray | None:
+ """The ensemble targets used for training the ensemble
+
+ It will attempt to load and cache them in memory but
+ return None if it can't.
+
+ Returns
+ -------
+ np.ndarray | None
+ The ensemble targets, if they can be loaded
+ """
+ if kind == "ensemble":
+ if self._y_ensemble is None:
+ if os.path.exists(self.backend._get_targets_ensemble_filename()):
+ self._y_ensemble = self.backend.load_targets_ensemble()
+ return self._y_ensemble
+
+ elif kind == "test":
+ return self._y_test
+
+ else:
+ raise NotImplementedError(kind)
+
+ def X_data(self, kind: str = "ensemble") -> SUPPORTED_FEAT_TYPES:
+ """The ensemble targets used for training the ensemble
+
+ It will attempt to load and cache them in memory but
+ return None if it can't.
+
+ Returns
+ -------
+ np.ndarray | None
+ The ensemble targets, if they can be loaded
+ """
+ if kind == "ensemble":
+ if self._X_ensemble is None:
+ if os.path.exists(self.backend._get_input_ensemble_filename()):
+ self._X_ensemble = self.backend.load_input_ensemble()
+ return self._X_ensemble
+
+ elif kind == "test":
+ return self._X_test
+
+ else:
+ raise NotImplementedError(kind)
+
+ def run(
+ self,
+ iteration: int,
+ pynisher_context: str | None = None,
+ time_left: float | None = None,
+ end_at: float | None = None,
+ time_buffer: int = 5,
+ ) -> tuple[list[dict[str, Any]], int | float]:
+ """Run the ensemble building process
+
+ Parameters
+ ----------
+ iteration : int
+ What iteration to associate with this run
+
+ pynisher_context : str | None = None
+ The pynisher context to run in. If None, defaults to
+ multiprocessing.get_context(None)
+
+ time_left : float | None = None
+ How much time should be left for this run. Either this or `end_at` must
+ be provided.
+
+ end_at : float | None = None
+ When this run should end. Either this or `time_left` must be provided.
+
+ time_buffer : int = 5
+ How much extra time to add as a buffer to this run. This means there is
+ always some amount of time to do something useful.
+
+ Returns
+ -------
+ (ensemble_history, nbest)
+ """
+ if time_left is None and end_at is None:
+ raise ValueError("Must provide either time_left or end_at.")
+
+ elif time_left is not None and end_at is not None:
+ raise ValueError("Cannot provide both time_left and end_at.")
+
+ if not self.logger:
+ self.logger = get_named_client_logger(
+ name="EnsembleBuilder",
+ port=self.logger_port,
+ )
+
+ process_start_time = time.time()
+ while True:
+
+ if time_left is not None:
+ time_elapsed = time.time() - process_start_time
+ time_left -= time_elapsed
+ else:
+ assert end_at is not None
+ current_time = time.time()
+ if current_time > end_at:
+ break
+ else:
+ time_left = end_at - current_time
+
+ wall_time_in_s = int(time_left - time_buffer)
+ if wall_time_in_s < 1:
+ break
+
+ context = multiprocessing.get_context(pynisher_context)
+ preload_modules(context)
+
+ safe_ensemble_script = pynisher.enforce_limits(
+ wall_time_in_s=wall_time_in_s,
+ mem_in_mb=self.memory_limit,
+ logger=self.logger,
+ context=context,
+ )(self.main)
+
+ safe_ensemble_script(time_left, iteration)
+
+ status = safe_ensemble_script.exit_status
+ if isinstance(status, pynisher.MemorylimitException):
+ # if ensemble script died because of memory error,
+ # reduce nbest to reduce memory consumption and try it again
+
+ # ATTENTION: main will start from scratch;
+ # all data structures are empty again
+ try:
+ self.previous_candidates_path.unlink()
+ except: # noqa E722
+ pass
+
+ if (
+ isinstance(self.ensemble_nbest, numbers.Integral)
+ and self.ensemble_nbest <= 1
+ ):
+ if self.read_at_most == 1:
+ self.logger.error(
+ "Memory Exception -- Unable to further reduce the number"
+ " of ensemble members and can no further limit the number"
+ " of ensemble members loaded per iteration, please restart"
+ " Auto-sklearn with a higher value for the argument"
+ f" `memory_limit` (current limit is {self.memory_limit}MB)."
+ " The ensemble builder will keep running to delete files"
+ " from disk in case this was enabled.",
+ )
+ self.ensemble_nbest = 0
+ else:
+ self.read_at_most = 1
+ self.logger.warning(
+ "Memory Exception -- Unable to further reduce the number of"
+ " ensemble members. Now reducing the number of predictions"
+ " per call to read at most to 1."
+ )
+ else:
+ if isinstance(self.ensemble_nbest, numbers.Integral):
+ self.ensemble_nbest = max(1, int(self.ensemble_nbest / 2))
+ else:
+ self.ensemble_nbest = self.ensemble_nbest / 2
+ self.logger.warning(
+ "Memory Exception -- restart with "
+ "less ensemble_nbest: %d" % self.ensemble_nbest
+ )
+ return [], self.ensemble_nbest
+ elif isinstance(status, pynisher.AnythingException):
+ return ([], self.ensemble_nbest)
+ else:
+ return safe_ensemble_script.result
+
+ return [], self.ensemble_nbest
+
+ def main(
+ self,
+ time_left: float | None = None,
+ iteration: int = 0,
+ ) -> tuple[list[dict[str, Any]], int | float]:
+ """Run the main loop of ensemble building
+
+ The process is:
+ * Load all available runs + previous candidates (if any)
+ * Update the loss of those that require
+ * From these runs, get a list of candidates
+ * Save candidates
+ * Delete models that are not candidates
+ * Build an ensemble from the candidates if there are new candidates
+
+ Parameters
+ ----------
+ time_left : float | None = None
+ How much time is left for this run
+
+ iteration : int = 0
+ The iteration of this run
+
+ Returns
+ -------
+ (ensemble_history: list[dict[str, Any]], nbest: int | float)
+ """
+ # Pynisher jobs inside dask 'forget' the logger configuration.
+ # So we have to set it up accordingly
+ self.logger = get_named_client_logger(
+ name="EnsembleBuilder",
+ port=self.logger_port,
+ )
+
+ if time_left is not None:
+ self.start_time = time.time()
+ used_time = time.time() - self.start_time
+ left_for_iter = time_left - used_time
+ itr = iteration if str(iteration) is not None else ""
+ self.logger.debug(f"Starting iteration {itr}, time left: {left_for_iter}")
+
+ # Can't load data, exit early
+ if not os.path.exists(self.backend._get_targets_ensemble_filename()):
+ self.logger.debug(f"No targets for ensemble: {traceback.format_exc()}")
+ raise RuntimeError("No targets for ensemble")
+
+ # We will delete runs once we are complete
+ deletable_runs: set[Run] = set()
+
+ # Load in information from previous candidates and also runs
+ available_runs = self.available_runs()
+
+ # Update runs with information of available previous candidates
+ previous_candidates = self.previous_candidates()
+ available_runs.update(previous_candidates)
+
+ # We just need the values now, not the key value pairs {run.id: Run}
+ runs = list(available_runs.values())
+
+ if len(runs) == 0:
+ self.logger.debug("Found no runs")
+ raise RuntimeError("Found no runs")
+
+ # We load in `X_data` if we need it
+ if any(m._needs_X for m in self.metrics):
+ ensemble_X_data = self.X_data("ensemble")
+
+ if ensemble_X_data is None:
+ msg = "No `X_data` for 'ensemble' which was required by metrics"
+ self.logger.debug(msg)
+ raise RuntimeError(msg)
+ else:
+ ensemble_X_data = None
+
+ # Calculate the loss for those that require it
+ requires_update = self.requires_loss_update(runs)
+ if self.read_at_most is not None:
+ requires_update = requires_update[: self.read_at_most]
+
+ for run in requires_update:
+ run.record_modified_times() # So we don't count as modified next time
+ run.losses = {
+ metric.name: self.loss(run, metric=metric, X_data=ensemble_X_data)
+ for metric in self.metrics
+ }
+
+ # Get the dummy and real runs
+ dummies, candidates = split(runs, by=lambda r: r.is_dummy())
+
+ # We see if we need to delete any of the real runs before we waste compute
+ # on evaluating their candidacy for ensemble building
+ if any(candidates):
+ candidates, to_delete = self.requires_deletion(
+ candidates,
+ max_models=self.max_models_on_disc,
+ memory_limit=self.model_memory_limit,
+ )
+
+ # If there are no candidates left, we just keep the best one, ordered by
+ # objectives and finally num_run
+ if not any(candidates):
+ best = min(
+ to_delete,
+ key=lambda r: (
+ *[r.losses.get(m.name, np.inf) for m in self.metrics],
+ r.num_run,
+ ),
+ )
+ candidates = [best]
+ to_delete.remove(best)
+
+ if any(to_delete):
+ self.logger.info(
+ f"Deleting runs {to_delete} due to"
+ f" max_models={self.max_models_on_disc} and/or"
+ f" memory_limit={self.model_memory_limit}"
+ )
+ deletable_runs.update(to_delete)
+
+ # If there are any candidates, perform candidates selection
+ if any(candidates):
+ candidates, to_delete = self.candidate_selection(
+ runs=candidates,
+ dummies=dummies,
+ better_than_dummy=True,
+ nbest=self.ensemble_nbest,
+ )
+ if any(to_delete):
+ self.logger.info(
+ f"Deleting runs {to_delete} due to"
+ f" nbest={self.ensemble_nbest} or worse than dummies"
+ )
+ deletable_runs.update(to_delete)
+ else:
+ candidates = dummies
+ self.logger.warning("No runs were available to build an ensemble from")
+
+ # In case we record test predictions and not every model has test predictions,
+ # only use the subset of models that has predictions for both the test set and
+ # the ensemble optimization set.
+ candidates_set = set(candidates)
+ test_subset = {r for r in candidates if r.pred_path("test").exists()}
+
+ if len(test_subset) > 0:
+ candidates = sorted(test_subset, key=lambda r: r.id)
+ test_models = candidates
+
+ to_delete = candidates_set - test_subset
+ if any(to_delete):
+ self.logger.info(
+ f"Deleting runs {to_delete} due to runs not"
+ ' having "test_predictions" while others do not:'
+ f"\nHave test_predictions = {test_subset}"
+ f"\nNo test_predictions = {to_delete}"
+ )
+ deletable_runs.update(to_delete)
+
+ else:
+ candidates = sorted(candidates_set, key=lambda r: r.id)
+ test_models = []
+
+ # Save the candidates for the next round
+ with self.previous_candidates_path.open("wb") as f:
+ pickle.dump({run.id: run for run in candidates}, f)
+
+ # If there was any change from the previous run, either in terms of
+ # runs or one of those runs had its loss updated, then we need to
+ # fit the ensemble builder
+ previous_candidate_ids = set(previous_candidates)
+ current_candidate_ids = set(run.id for run in candidates)
+ difference = previous_candidate_ids ^ current_candidate_ids
+
+ was_updated_candidates = list(run in candidates for run in requires_update)
+
+ if not any(difference) and not any(was_updated_candidates):
+ self.logger.info("All ensemble candidates the same, no update required")
+ return self.ensemble_history, self.ensemble_nbest
+
+ targets = cast(np.ndarray, self.targets("ensemble")) # Sure they exist
+
+ ensemble = self.fit_ensemble(
+ candidates=candidates,
+ targets=targets,
+ runs=runs,
+ ensemble_class=self.ensemble_class,
+ ensemble_kwargs=self.ensemble_kwargs,
+ X_data=ensemble_X_data,
+ task=self.task_type,
+ metrics=self.metrics,
+ precision=self.precision,
+ random_state=self.random_state,
+ )
+
+ self.logger.info(str(ensemble))
+ ens_perf = ensemble.get_validation_performance()
+ self.validation_performance_ = min(self.validation_performance_, ens_perf)
+ self.backend.save_ensemble(
+ ensemble=ensemble, idx=iteration, seed=self.seed # type: ignore
+ )
+
+ performance_stamp = {"Timestamp": pd.Timestamp.now()}
+
+ for kind, score_name, models in [
+ ("ensemble", "optimization", candidates),
+ ("test", "test", test_models),
+ ]:
+ if len(models) == 0:
+ continue
+
+ pred_targets = self.targets(kind)
+ if pred_targets is None:
+ self.logger.warning(f"No ensemble targets for {kind}")
+ continue
+
+ run_preds = [r.predictions(kind, precision=self.precision) for r in models]
+ pred = ensemble.predict(run_preds)
+
+ if any(m._needs_X for m in self.metrics):
+ X_data = self.X_data(kind)
+ if X_data is None:
+ msg = f"No `X` data for '{kind}' which was required by metrics"
+ self.logger.debug(msg)
+ raise RuntimeError(msg)
+ else:
+ X_data = None
+
+ scores = calculate_scores(
+ solution=pred_targets,
+ prediction=pred,
+ task_type=self.task_type,
+ metrics=self.metrics,
+ X_data=X_data,
+ scoring_functions=None,
+ )
+
+ # TODO only one metric in history
+ #
+ # We should probably return for all metrics but this makes
+ # automl::performance_history a lot more complicated, will
+ # tackle in a future PR
+ first_metric = self.metrics[0]
+ performance_stamp[f"ensemble_{score_name}_score"] = scores[
+ first_metric.name
+ ]
+
+ # Add the performance stamp to the history
+ self.ensemble_history.append(performance_stamp)
+
+ # Lastly, delete any runs that need to be deleted. We save this as the last step
+ # so that we have an ensemble saved that is up to date. If we do not do so,
+ # there could be runs deleted that are in th previous ensemble and we do not
+ # manage to update the ensemble due to a crash or the process being killed
+ # before it could be updated
+ self.delete_runs(deletable_runs)
+
+ return self.ensemble_history, self.ensemble_nbest
+
+ def requires_loss_update(
+ self,
+ runs: Sequence[Run],
+ metrics: Sequence[Scorer] | None = None,
+ ) -> list[Run]:
+ """
+
+ Parameters
+ ----------
+ runs : Sequence[Run]
+ The runs to process
+
+ metrics: Sequence[Scorer] | None = None
+ The metrics to check for, defaults to builders
+
+ Returns
+ -------
+ list[Run]
+ The runs that require a loss to be calculated
+ """
+ metrics = metrics if metrics is not None else self.metrics
+
+ queue = []
+ for run in sorted(runs, key=lambda run: run.recorded_mtimes["ensemble"]):
+ if any(metric.name not in run.losses for metric in metrics):
+ queue.append(run)
+
+ elif run.was_modified():
+ self.logger.debug(f"{run.id} had its predictions modified?")
+ queue.append(run)
+
+ return queue
+
+ def candidate_selection(
+ self,
+ runs: Sequence[Run],
+ dummies: Run | list[Run],
+ *,
+ better_than_dummy: bool = False,
+ nbest: int | float | None = None,
+ metrics: Sequence[Scorer] | None = None,
+ ) -> tuple[list[Run], set[Run]]:
+ """Get a list of candidates from `runs`, garuanteeing at least one
+
+ Applies a set of reductions in order of parameters to reach a set of final
+ candidates.
+
+ Expects at least one `dummies` run.
+
+ Parameters
+ ----------
+ runs : Sequence[Run]
+ The runs to evaluate candidates from.
+
+ dummies: Run | Sequence[Run]
+ The dummy run to base from
+
+ better_than_dummy: bool = False
+ Whether the run must be better than the best dummy run to be a candidate.
+ In the case where there are no candidates left, the dummies will then be
+ used.
+
+ nbest : int | float | None
+ The nbest models to select. If `int`, acts as an absolute limit.
+ If `float`, acts as a percentage of available candidates.
+
+ metrics : Sequence[Scorer] | None = None
+ The metrics to consider, defaults to the builders
+
+ Returns
+ -------
+ (candidates: list[Run], discarded: set[Run])
+ A tuple of runs that are candidates and also those that didn't make it
+ """
+ metrics = metrics if metrics else self.metrics
+
+ if isinstance(dummies, Run):
+ dummies = [dummies]
+
+ assert len(dummies) > 0 and len(runs) > 0, "At least 1 real run and dummy run"
+
+ all_discarded: set[Run] = set()
+
+ # We filter out all runs that don't have any predictions for the ensemble
+ candidates, discarded = split(
+ runs,
+ by=lambda run: run.pred_path("ensemble").exists(),
+ )
+ all_discarded.update(discarded)
+
+ if len(candidates) == 0:
+ self.logger.debug("No runs with predictions on ensemble set, using dummies")
+ return dummies, all_discarded
+
+ for run in discarded:
+ self.logger.warning(f"Have no ensemble predictions for {run}")
+
+ # Get all the ones that have all metrics and tangible losses
+ has_metrics = lambda r: all(m.name in r.losses for m in metrics)
+ tangible_losses = lambda r: all(loss < np.inf for loss in r.losses.values())
+
+ candidates, discarded = split(
+ candidates,
+ by=lambda r: has_metrics(r) and tangible_losses(r),
+ )
+ all_discarded.update(discarded)
+
+ if len(candidates) == 0:
+ self.logger.debug("No runs with a usable loss, using dummies")
+ return dummies, all_discarded
+
+ if better_than_dummy:
+
+ # In the single objective case, they must explicitly be better than dummy
+ if len(metrics) == 1:
+ metric = metrics[0]
+
+ best_dummy = min(dummies, key=lambda d: d.losses[metric.name])
+ self.logger.debug(f"Using dummy {best_dummy} to filter candidates")
+
+ candidates, discarded = split(
+ candidates,
+ by=lambda r: r.losses[metric.name] < best_dummy.losses[metric.name],
+ )
+
+ # In the multiobjective case, they must be better than at least one of the
+ # best dummies, where there is a best dummy for each metric
+ else:
+ best_dummies = {
+ metric.name: min(dummies, key=lambda d: d.losses[metric.name])
+ for metric in metrics
+ }
+ self.logger.debug(f"Using dummies {best_dummies} to filter candidates")
+
+ candidates, discarded = split(
+ candidates,
+ by=lambda r: any(
+ r.losses[metric_name] < best_dummy.losses[metric_name]
+ for metric_name, best_dummy in best_dummies.items()
+ ),
+ )
+
+ all_discarded.update(discarded)
+
+ # If there are no real candidates left, use the dummies
+ if len(candidates) == 0:
+ self.logger.warning(
+ "No models better than random - using Dummy losses!"
+ f"\n\tModels besides current dummy model: {len(candidates)}"
+ f"\n\tDummy models: {len(dummies)}",
+ )
+ return dummies, all_discarded
+
+ if nbest is not None:
+ # We order the runs according to a roundrobin of the metrics
+ # i.e. the 1st run will be best in metric[0],
+ # the 2nd run will be best in metric[1],
+ # the 3rd run will be second best in metric[0],
+ # the 4th run will be second best in metric[1],
+ # ... whoops, what would be the 5th run, third best in metric[0] was
+ # also the same as the 4th run. This was skipped
+ # the 5th run will be third best in metric[1] **as we skipped above**
+ #
+ # With removing duplicates, this means if a run was already contained, it
+ # will skip to the next member in the roundrobin fashion.
+ sorted_candidates = [
+ sorted(
+ candidates,
+ key=lambda r: (r.losses.get(m.name, np.inf), r.num_run),
+ )
+ for m in metrics
+ ]
+ candidates = list(roundrobin(*sorted_candidates, duplicates=False))
+
+ # Determine how many to keep, always keeping one
+ if isinstance(nbest, float):
+ nkeep = int(len(candidates) * nbest)
+ else:
+ nkeep = nbest
+
+ candidates, discarded = cut(candidates, nkeep)
+ self.logger.info(f"Discarding {len(discarded)}/{len(candidates)} runs")
+
+ # Always preserve at least one, the best
+ if len(candidates) == 0:
+ candidates, discared = cut(discarded, 1)
+ self.logger.warning("nbest too aggresive, using single best")
+
+ all_discarded.update(discarded)
+
+ return candidates, all_discarded
+
+ def fit_ensemble(
+ self,
+ candidates: Sequence[Run],
+ runs: Sequence[Run],
+ *,
+ targets: np.ndarray | None = None,
+ ensemble_class: Type[AbstractEnsemble] = EnsembleSelection,
+ ensemble_kwargs: Mapping[str, Any] | None = None,
+ X_data: SUPPORTED_FEAT_TYPES | None = None,
+ task: int | None = None,
+ metrics: Sequence[Scorer] | None = None,
+ precision: int | None = None,
+ random_state: int | np.random.RandomState | None = None,
+ ) -> AbstractEnsemble:
+ """Fit an ensemble from the provided runs.
+
+ Note
+ ----
+ Expects all runs to have the "ensemble" predictions present
+
+ Parameters
+ ----------
+ candidates: Sequence[Run]
+ List of runs to build an ensemble from
+
+ runs: Sequence[Run]
+ List of all runs (also pruned ones and dummy runs)
+
+ targets: np.ndarray | None = None
+ The targets to build the ensemble with
+
+ ensemble_class: Type[AbstractEnsemble]
+ Implementation of the ensemble algorithm.
+
+ ensemble_kwargs: Mapping[str, Any] | None
+ Arguments passed to the constructor of the ensemble algorithm.
+
+ X_data: SUPPORTED_FEAT_TYPES | None = None
+ The base level data.
+
+ task: int | None = None
+ The kind of task performed
+
+ metrics: Sequence[Scorer] | None = None
+ Metrics to optimize the ensemble for.
+
+ precision: int | None = None
+ The precision with which to load run predictions
+
+ random_state: int | RandomState | None = None
+ The random state to use
+
+ Returns
+ -------
+ AbstractEnsemble
+ """
+ # Validate we have targets if None specified
+ if targets is None:
+ targets = self.targets("ensemble")
+ if targets is None:
+ path = self.backend._get_targets_ensemble_filename()
+ raise ValueError(f"`fit_ensemble` could not find any targets at {path}")
+
+ ensemble_class = (
+ ensemble_class if ensemble_class is not None else self.ensemble_class
+ )
+
+ # Create the ensemble_kwargs, favouring in order:
+ # 1) function kwargs, 2) function params 3) init_kwargs 4) init_params
+
+ # Collect func params in dict if they're not None
+ params = {
+ k: v
+ for k, v in [
+ ("task_type", task),
+ ("metrics", metrics),
+ ("random_state", random_state),
+ ]
+ if v is not None
+ }
+
+ kwargs = {
+ "backend": self.backend,
+ "task_type": self.task_type,
+ "metrics": self.metrics,
+ "random_state": self.random_state,
+ **(self.ensemble_kwargs or {}),
+ **params,
+ **(ensemble_kwargs or {}),
+ }
+
+ ensemble = ensemble_class(**kwargs) # type: AbstractEnsemble
+
+ self.logger.debug(f"Fitting ensemble on {len(candidates)} models")
+ start_time = time.time()
+
+ precision = precision if precision is not None else self.precision
+ predictions_train = [
+ run.predictions("ensemble", precision=precision) for run in candidates
+ ]
+
+ ensemble.fit(
+ base_models_predictions=predictions_train,
+ X_data=X_data,
+ true_targets=targets,
+ model_identifiers=[run.id for run in candidates],
+ runs=runs,
+ )
+
+ duration = time.time() - start_time
+ self.logger.debug(f"Fitting the ensemble took {duration} seconds.")
+ return ensemble
+
+ def requires_deletion(
+ self,
+ runs: Sequence[Run],
+ *,
+ max_models: int | None = None,
+ memory_limit: float | None = None,
+ metrics: Sequence[Scorer] | None = None,
+ ) -> tuple[list[Run], set[Run]]:
+ """Cut a list of runs into those to keep and those to delete
+
+ If neither params are specified, this method should do nothing.
+
+ In the case of multiple metrics, we interleave these according to the order of
+ the provided metrics.
+
+ Parameters
+ ----------
+ runs : Sequence[Run]
+ The runs to check
+
+ max_models : int | None = None
+ The maximum amount of models to have on disk. Leave `None` for no effect
+
+ memory_limit : float | None = None
+ The memory limit in MB, leave `None` for no effect
+
+ metrics: Sequence[Scorer] | None = None
+ The metrics to consider for the runs, defaults to `self.metrics` passed
+ during construction if `None`.
+
+ Returns
+ -------
+ (keep: list[Run], delete: set[Run])
+ The list of runs to keep and those to delete
+ """
+ if memory_limit is None and max_models is None:
+ return list(runs), set()
+
+ metrics = metrics if metrics is not None else self.metrics
+
+ # We order the runs according to a roundrobin of the metrics
+ # i.e. the 1st run will be best in metric[0],
+ # the 2nd run will be best in metric[1],
+ # the 3rd run will be second best in metric[0],
+ # the 4th run will be second best in metric[1],
+ # ... whoops, what would be the 5th run, third best in metric[0] was also
+ # ... the same as the 4th run. This was skipped
+ # the 5th run will be third best in metric[1] **as we skipped above**
+ #
+ # With removing duplicates, this means if a run was already contained, it will
+ # skip to the next member in the roundrobin fashion.
+ sorted_runs = [
+ sorted(runs, key=lambda r: (r.losses.get(metric.name, np.inf), r.num_run))
+ for metric in metrics
+ ]
+ keep = list(roundrobin(*sorted_runs, duplicates=False))
+
+ delete: set[Run] = set()
+
+ if max_models is not None and max_models < len(runs):
+ keep, to_delete = cut(keep, max_models)
+
+ if any(to_delete):
+ self.logger.info(
+ "Limiting num of models via `max_models_on_disc`"
+ f" max_models={max_models}"
+ f" remaining={len(keep)}"
+ f" discarded={len(to_delete)}"
+ )
+ delete.update(to_delete)
+
+ if memory_limit is not None:
+ largest = max(runs, key=lambda r: r.mem_usage)
+ cutoff = memory_limit - largest.mem_usage
+
+ accumulated_mem_usage = accumulate(r.mem_usage for r in runs)
+
+ cutpoint = findwhere(accumulated_mem_usage, lambda mem: mem > cutoff)
+ keep, to_delete = cut(keep, cutpoint)
+
+ if any(to_delete):
+ self.logger.info(
+ "Limiting num of models via `memory_limit`"
+ f" memory_limit={memory_limit}"
+ f" cutoff={cutoff}"
+ f" largest={largest.mem_usage}"
+ f" remaining={len(keep)}"
+ f" discarded={len(to_delete)}"
+ )
+ delete.update(to_delete)
+
+ return keep, delete
+
+ def loss(
+ self,
+ run: Run,
+ metric: Scorer,
+ *,
+ X_data: SUPPORTED_FEAT_TYPES | None = None,
+ kind: str = "ensemble",
+ ) -> float:
+ """Calculate the loss for a run
+
+ Parameters
+ ----------
+ run: Run
+ The run to calculate the loss for
+
+ metric: Scorer
+ The metric to calculate the loss of
+
+ X_data: SUPPORTED_FEAT_TYPES | None = None
+ Any X_data required to be passed to the metric
+
+ kind: str = "ensemble"
+ The kind of targets to use for the run
+
+ Returns
+ -------
+ float
+ The loss for the run
+ """
+ targets = self.targets(kind)
+ if targets is None:
+ self.logger.error(f"No targets of {kind}")
+ return np.inf
+
+ try:
+ predictions = run.predictions(kind, precision=self.precision)
+ loss: float = calculate_loss(
+ solution=targets,
+ prediction=predictions,
+ task_type=self.task_type,
+ metric=metric,
+ X_data=X_data,
+ )
+ except Exception as e:
+ tb = traceback.format_exc()
+ self.logger.error(f"Error getting loss `{metric.name}` for {run}:{e}{tb}")
+ loss = np.inf
+ finally:
+ return loss
+
+ def delete_runs(self, runs: Iterable[Run]) -> None:
+ """Delete runs
+
+ Will not delete dummy runs
+
+ Parameters
+ ----------
+ runs : Sequence[Run]
+ The runs to delete
+ """
+ items = iter(run for run in runs if not run.is_dummy() and run.dir.exists())
+ for run in items:
+ try:
+ rmtree(run.dir, atomic=True)
+ self.logger.info(f"Deleted files for {run}")
+ except Exception as e:
+ self.logger.error(f"Failed to delete files for {run}: \n{e}")
diff --git a/autosklearn/ensemble_building/manager.py b/autosklearn/ensemble_building/manager.py
new file mode 100644
index 0000000000..31a42b2dbe
--- /dev/null
+++ b/autosklearn/ensemble_building/manager.py
@@ -0,0 +1,393 @@
+from __future__ import annotations
+
+from typing import Any, Dict, Sequence, Type
+
+import logging.handlers
+import time
+import traceback
+
+import dask.distributed
+import numpy as np
+from sklearn.utils.validation import check_random_state
+from smac.callbacks import IncorporateRunResultCallback
+from smac.optimizer.smbo import SMBO
+from smac.runhistory.runhistory import RunInfo, RunValue
+from smac.tae.base import StatusType
+
+from autosklearn.automl_common.common.utils.backend import Backend
+from autosklearn.ensemble_building.builder import EnsembleBuilder
+from autosklearn.ensembles.abstract_ensemble import AbstractEnsemble
+from autosklearn.ensembles.ensemble_selection import EnsembleSelection
+from autosklearn.metrics import Scorer
+from autosklearn.util.logging_ import get_named_client_logger
+
+
+class EnsembleBuilderManager(IncorporateRunResultCallback):
+ def __init__(
+ self,
+ backend: Backend,
+ dataset_name: str,
+ task: int,
+ metrics: Sequence[Scorer],
+ time_left_for_ensembles: float = np.inf,
+ max_iterations: int | None = None,
+ pynisher_context: str = "fork",
+ ensemble_class: Type[AbstractEnsemble] = EnsembleSelection,
+ ensemble_kwargs: Dict[str, Any] | None = None,
+ ensemble_nbest: int | float = 50,
+ max_models_on_disc: int | float | None = None,
+ seed: int = 1,
+ precision: int = 32,
+ memory_limit: int | None = None,
+ read_at_most: int | None = None,
+ logger_port: int = logging.handlers.DEFAULT_TCP_LOGGING_PORT,
+ random_state: int | np.random.RandomState | None = None,
+ start_time: float | None = None,
+ ):
+ """SMAC callback to handle ensemble building
+
+ Parameters
+ ----------
+ backend: Backend
+ backend to write and read files
+
+ dataset_name: str
+ name of dataset
+
+ task: int
+ Type of ML task
+
+ metrics: Sequence[Scorer]
+ Metrics to optimize the ensemble for
+
+ time_left_for_ensemble: float = np.inf
+ How much time is left for the task in seconds.
+ Job should finish within this allocated time
+
+ max_iterations: int | None = None
+ maximal number of iterations to run this script. None indicates no limit
+ on iterations.
+
+ pynisher_context: "spawn" | "fork" | "forkserver" = "fork"
+ The multiprocessing context for pynisher.
+
+ ensemble_class : Type[AbstractEnsemble] (default=EnsembleSelection)
+ Class implementing the post-hoc ensemble algorithm. Set to
+ ``None`` to disable ensemble building or use ``SingleBest``
+ to obtain only use the single best model instead of an
+ ensemble.
+
+ ensemble_kwargs : Dict, optional
+ Keyword arguments that are passed to the ensemble class upon
+ initialization.
+
+ ensemble_nbest: int | float = 50
+ If int: consider only the n best prediction
+ If float: consider only this fraction of the best models
+
+ max_models_on_disc: int | float | None = None
+ Defines the maximum number of models that are kept in the disc.
+
+ If int, it must be greater or equal than 1, and dictates the max
+ number of models to keep.
+
+ If float, it will be interpreted as the max megabytes allowed of
+ disc space. That is, if the number of ensemble candidates require more
+ disc space than this float value, the worst models will be deleted to
+ keep within this budget. Models and predictions of the worst-performing
+ models will be deleted then.
+
+ If None, the feature is disabled. It defines an upper bound on the
+ models that can be used in the ensemble.
+
+ seed: int = 1
+ Seed used for the inidividual runs
+
+ precision: 16 | 32 | 64 | 128 = 32
+ Precision of floats to read the predictions
+
+ memory_limit: int | None = None
+ Memory limit in mb. If ``None``, no memory limit is enforced.
+
+ read_at_most: int | None = None
+ read at most n new prediction files in each iteration. If `None`, will read
+ the predictions and calculate losses for all runs that require it.
+
+ logger_port: int = DEFAULT_TCP_LOGGING_PORT
+ Port that receives logging records
+
+ start_time: float | None = None
+ DISABLED: Just using time.time() to set it
+ The time when this job was started, to account for any latency in job
+ allocation.
+ """
+ self.time_left_for_ensembles = time_left_for_ensembles
+ self.backend = backend
+ self.dataset_name = dataset_name
+ self.task = task
+ self.metrics = metrics
+ self.ensemble_class = ensemble_class
+ self.ensemble_kwargs = ensemble_kwargs
+ self.ensemble_nbest = ensemble_nbest
+ self.max_models_on_disc = max_models_on_disc
+ self.seed = seed
+ self.precision = precision
+ self.max_iterations = max_iterations
+ self.read_at_most = read_at_most
+ self.memory_limit = memory_limit
+ self.random_state = check_random_state(random_state)
+ self.logger_port = logger_port
+ self.pynisher_context = pynisher_context
+
+ # Store something similar to SMAC's runhistory
+ self.history: list[dict[str, Any]] = []
+
+ # We only submit new ensembles when there is not an active ensemble job
+ self.futures: list[dask.distributed.Future] = []
+
+ # The last criteria is the number of iterations
+ self.iteration = 0
+
+ # Keep track of when we started to know when we need to finish!
+ self.start_time = time.time()
+
+ def __call__(
+ self,
+ smbo: "SMBO",
+ run_info: RunInfo,
+ result: RunValue,
+ time_left: float,
+ ) -> None:
+ """
+ Returns
+ -------
+ List[Tuple[int, float, float, float]]:
+ A list with the performance history of this ensemble, of the form
+ [(pandas_timestamp, train_performance, val_performance, test_performance)]
+ """
+ if result.status in (StatusType.STOP, StatusType.ABORT) or smbo._stop:
+ return
+ self.build_ensemble(smbo.tae_runner.client)
+
+ def build_ensemble(
+ self,
+ dask_client: dask.distributed.Client,
+ ) -> None:
+ """Build the ensemble
+
+ Parameters
+ ----------
+ dask_client: dask.distributed.Client
+ The dask client to use
+ """
+ # The second criteria is elapsed time
+ elapsed_time = time.time() - self.start_time
+
+ logger = get_named_client_logger(
+ name="EnsembleBuilder",
+ port=self.logger_port,
+ )
+
+ # First test for termination conditions
+ if self.time_left_for_ensembles < elapsed_time:
+ logger.info(
+ "Terminate ensemble building as not time is left (run for {}s)".format(
+ elapsed_time
+ ),
+ )
+ return
+ if self.max_iterations is not None and self.max_iterations <= self.iteration:
+ logger.info(
+ "Terminate ensemble building because of max iterations:"
+ f" {self.max_iterations} of {self.iteration}"
+ )
+ return
+
+ if len(self.futures) != 0:
+ if self.futures[0].done():
+ result = self.futures.pop().result()
+ if result:
+ ensemble_history, self.ensemble_nbest = result
+ logger.debug(
+ f"iteration={self.iteration} @ elapsed_time={elapsed_time}"
+ f" has history={ensemble_history}"
+ )
+ self.history.extend(ensemble_history)
+
+ # Only submit new jobs if the previous ensemble job finished
+ if len(self.futures) == 0:
+
+ # Add the result of the run
+ # On the next while iteration, no references to
+ # ensemble builder object, so it should be garbage collected to
+ # save memory while waiting for resources
+ # Also, notice how ensemble nbest is returned, so we don't waste
+ # iterations testing if the deterministic predictions size can
+ # be fitted in memory
+ try:
+ # Submit a Dask job from this job, to properly
+ # see it in the dask diagnostic dashboard
+ # Notice that the forked ensemble_builder_process will
+ # wait for the below function to be done
+ self.futures.append(
+ dask_client.submit(
+ EnsembleBuilderManager.fit_and_return_ensemble,
+ backend=self.backend,
+ dataset_name=self.dataset_name,
+ task_type=self.task,
+ metrics=self.metrics,
+ ensemble_class=self.ensemble_class,
+ ensemble_kwargs=self.ensemble_kwargs,
+ ensemble_nbest=self.ensemble_nbest,
+ max_models_on_disc=self.max_models_on_disc,
+ seed=self.seed,
+ precision=self.precision,
+ memory_limit=self.memory_limit,
+ read_at_most=self.read_at_most,
+ random_state=self.random_state,
+ end_at=self.start_time + self.time_left_for_ensembles,
+ iteration=self.iteration,
+ pynisher_context=self.pynisher_context,
+ logger_port=self.logger_port,
+ )
+ )
+
+ logger.info(
+ "{}/{} Started Ensemble builder job at {} for iteration {}.".format(
+ # Log the client to make sure we
+ # remain connected to the scheduler
+ self.futures[0],
+ dask_client,
+ time.strftime("%Y.%m.%d-%H.%M.%S"),
+ self.iteration,
+ ),
+ )
+ self.iteration += 1
+ except Exception as e:
+ exception_traceback = traceback.format_exc()
+ error_message = repr(e)
+ logger.critical(exception_traceback)
+ logger.critical(error_message)
+
+ @staticmethod
+ def fit_and_return_ensemble(
+ iteration: int,
+ end_at: float,
+ backend: Backend,
+ dataset_name: str,
+ task_type: int,
+ metrics: Sequence[Scorer],
+ pynisher_context: str,
+ ensemble_class: Type[AbstractEnsemble] = EnsembleSelection,
+ ensemble_kwargs: Dict[str, Any] | None = None,
+ ensemble_nbest: int | float = 50,
+ max_models_on_disc: int | float | None = None,
+ seed: int = 1,
+ precision: int = 32,
+ memory_limit: int | None = None,
+ read_at_most: int | None = None,
+ logger_port: int = logging.handlers.DEFAULT_TCP_LOGGING_PORT,
+ random_state: int | np.random.RandomState | None = None,
+ ) -> tuple[list[dict[str, Any]], int | float]:
+ """
+ A short function to fit and create an ensemble. It is just a wrapper to easily
+ send a request to dask to create an ensemble and clean the memory when finished
+
+ Parameters
+ ----------
+ iteration: int
+ The current iteration
+
+ end_at: float
+ At what time the job must finish. Needs to be the endtime and not the
+ time left because we do not know when dask schedules the job.
+
+ backend: Backend
+ Backend to write and read files
+
+ dataset_name: str
+ name of dataset
+
+ task_type: int
+ type of ML task
+
+ metrics: Sequence[Scorer]
+ Metrics to optimize the ensemble for.
+
+ pynisher_context: "fork" | "spawn" | "forkserver" = "fork"
+ Context to use for multiprocessing, can be either fork, spawn or forkserver.
+
+ ensemble_class : Type[AbstractEnsemble] (default=EnsembleSelection)
+ Class implementing the post-hoc ensemble algorithm. Set to
+ ``None`` to disable ensemble building or use ``SingleBest``
+ to obtain only use the single best model instead of an
+ ensemble.
+
+ ensemble_kwargs : Dict, optional
+ Keyword arguments that are passed to the ensemble class upon
+ initialization.
+
+ ensemble_nbest: int | float = 50
+ If int: consider only the n best prediction
+ If float: consider only this fraction of the best models
+
+ max_models_on_disc: int | float | None = 100
+ Defines the maximum number of models that are kept in the disc.
+
+ If int, it must be greater or equal than 1, and dictates the max number of
+ models to keep.
+
+ If float, it will be interpreted as the max megabytes allowed of disc space.
+ That is, if the number of ensemble candidates require more disc space than
+ this float value, the worst models will be deleted to keep within this
+ budget. Models and predictions of the worst-performing models will be
+ deleted then.
+
+ If None, the feature is disabled.
+
+ seed: int = 1
+ Seed used for training the models in the backend
+
+ precision: 16 | 32 | 64 | 128 = 32
+ Precision of floats to read the predictions
+
+ memory_limit: int | None = None
+ Memory limit in mb. If ``None``, no memory limit is enforced.
+
+ read_at_most: int | None = None
+ read at most n new prediction files in each iteration. If `None`, will read
+ the predictions and calculate losses for all runs that require it.
+
+ logger_port: int = DEFAULT_TCP_LOGGING_PORT
+ The port where the logging server is listening to.
+
+ random_state: int | RandomState | None = None
+ A random state used for the ensemble selection process.
+
+ Returns
+ -------
+ (ensemble_history: list[dict[str, Any]], nbest: int | float)
+ The ensemble history and the nbest chosen members
+ """
+ random_state = check_random_state(random_state)
+ result = EnsembleBuilder(
+ backend=backend,
+ dataset_name=dataset_name,
+ task_type=task_type,
+ metrics=metrics,
+ ensemble_class=ensemble_class,
+ ensemble_kwargs=ensemble_kwargs,
+ ensemble_nbest=ensemble_nbest,
+ max_models_on_disc=max_models_on_disc,
+ seed=seed,
+ precision=precision,
+ memory_limit=memory_limit,
+ read_at_most=read_at_most,
+ random_state=random_state.randint(10000000),
+ logger_port=logger_port,
+ ).run(
+ end_at=end_at,
+ iteration=iteration,
+ pynisher_context=pynisher_context,
+ )
+ return result
diff --git a/autosklearn/ensemble_building/run.py b/autosklearn/ensemble_building/run.py
new file mode 100644
index 0000000000..9687169944
--- /dev/null
+++ b/autosklearn/ensemble_building/run.py
@@ -0,0 +1,183 @@
+from __future__ import annotations
+
+from typing import Tuple
+
+import re
+from pathlib import Path
+
+import numpy as np
+
+from autosklearn.util.disk import sizeof
+
+RunID = Tuple[int, int, float]
+
+
+class Run:
+ """Class for storing information about a run used during ensemble building.
+
+ Note
+ ----
+ This is for internal use by the EnsembleBuilder and not for general usage.
+ """
+
+ # For matching prediction files
+ RE_MODEL_PREDICTION_FILE = (
+ r"^predictions_ensemble_([0-9]*)_([0-9]*)_([0-9]{1,3}\.[0-9]*).npy$"
+ )
+
+ # For matching run directories
+ RE_MODEL_DIR = r"^([0-9]*)_([0-9]*)_([0-9]{1,3}\.[0-9]*)$"
+
+ def __init__(self, path: Path) -> None:
+ """Creates a Run from a path pointing to the directory of a run
+
+ Parameters
+ ----------
+ path: Path
+ Expects something like /path/to/{seed}_{numrun}_{budget}
+ """
+ name = path.name
+ seed, num_run, budget = name.split("_")
+
+ self.dir = path
+ self.seed = int(seed)
+ self.num_run = int(num_run)
+ self.budget = float(budget)
+
+ # These are ordered based on preference
+ self.losses: dict[str, float] = {}
+
+ self._mem_usage: float | None = None
+
+ # Items that will be delete when the run is saved back to file
+ self._cache: dict[str, np.ndarray] = {}
+
+ # The recorded time of ensemble/test predictions modified
+ self.recorded_mtimes: dict[str, float] = {}
+ self.record_modified_times()
+
+ @property
+ def mem_usage(self) -> float:
+ """The memory usage of this run based on it's directory"""
+ if self._mem_usage is None:
+ self._mem_usage = round(sizeof(self.dir, unit="MB"), 2)
+
+ return self._mem_usage
+
+ def is_dummy(self) -> bool:
+ """Whether this run is a dummy run or not"""
+ return self.num_run == 1
+
+ def was_modified(self) -> bool:
+ """Query for when the ens file was last modified"""
+ recorded = self.recorded_mtimes.get("ensemble")
+ last = self.pred_path().stat().st_mtime
+ return recorded != last
+
+ def pred_path(self, kind: str = "ensemble") -> Path:
+ """Get the path to certain predictions"""
+ fname = f"predictions_{kind}_{self.seed}_{self.num_run}_{self.budget}.npy"
+ return self.dir / fname
+
+ def record_modified_times(self) -> None:
+ """Records the last time each prediction file type was modified, if it exists"""
+ self.recorded_mtimes = {}
+ for kind in ["ensemble", "test"]:
+ path = self.pred_path(kind) # type: ignore
+ if path.exists():
+ self.recorded_mtimes[kind] = path.stat().st_mtime
+
+ def has_predictions(self, kind: str = "ensemble") -> bool:
+ """
+ Parameters
+ ----------
+ kind: "ensemble" | "test" = "ensemble"
+ The kind of predictions to query for
+
+ Returns
+ -------
+ bool
+ Whether this run has the kind of predictions queried for
+ """
+ return self.pred_path(kind).exists()
+
+ def predictions(
+ self,
+ kind: str = "ensemble",
+ precision: int | None = None,
+ ) -> np.ndarray:
+ """Load the predictions for this run
+
+ Parameters
+ ----------
+ kind : "ensemble" | "test"
+ The kind of predictions to load
+
+ precisions : type | None = None
+ What kind of precision reduction to apply
+
+ Returns
+ -------
+ np.ndarray
+ The loaded predictions
+ """
+ key = f"predictions_{kind}"
+ if key in self._cache:
+ return self._cache[key]
+
+ path = self.pred_path(kind)
+
+ with path.open("rb") as f:
+ # TODO: We should probably remove this requirement. I'm not sure why model
+ # predictions are being saved as pickled
+ predictions = np.load(f, allow_pickle=True)
+
+ if precision:
+ dtypes: dict[int, type] = {16: np.float16, 32: np.float32, 64: np.float64}
+ dtype = dtypes.get(precision, None)
+
+ if dtype is not None:
+ predictions = predictions.astype(dtype=dtype, copy=False)
+
+ self._cache[key] = predictions
+ return predictions
+
+ def __getstate__(self) -> dict:
+ """Remove the cache when pickling."""
+ state = self.__dict__.copy()
+ del state["_cache"]
+ return state
+
+ def __setstate__(self, state: dict) -> None:
+ """Reset state and instansiate blank cache."""
+ self.__dict__.update(state)
+ self._cache = {}
+
+ @property
+ def id(self) -> RunID:
+ """Get the three components of it's id"""
+ return self.seed, self.num_run, self.budget
+
+ def __hash__(self) -> int:
+ return hash(self.id)
+
+ def __repr__(self) -> str:
+ return f"Run(id={self.id}, losses={self.losses})"
+
+ def __eq__(self, other: object) -> bool:
+ return isinstance(other, Run) and other.id == self.id
+
+ @staticmethod
+ def valid(path: Path) -> bool:
+ """
+ Parameters
+ ----------
+ path: Path
+ The path to check
+
+ Returns
+ -------
+ bool
+ Whether the path is a valid run dir
+ """
+ return re.match(Run.RE_MODEL_DIR, path.name) is not None
diff --git a/autosklearn/ensembles/__init__.py b/autosklearn/ensembles/__init__.py
index e69de29bb2..4e1af926cf 100644
--- a/autosklearn/ensembles/__init__.py
+++ b/autosklearn/ensembles/__init__.py
@@ -0,0 +1,10 @@
+from .abstract_ensemble import AbstractEnsemble, AbstractMultiObjectiveEnsemble
+from .ensemble_selection import EnsembleSelection
+from .singlebest_ensemble import SingleBest
+
+__all__ = [
+ "AbstractEnsemble",
+ "AbstractMultiObjectiveEnsemble",
+ "EnsembleSelection",
+ "SingleBest",
+]
diff --git a/autosklearn/ensembles/abstract_ensemble.py b/autosklearn/ensembles/abstract_ensemble.py
index 752131407f..ea14a45e90 100644
--- a/autosklearn/ensembles/abstract_ensemble.py
+++ b/autosklearn/ensembles/abstract_ensemble.py
@@ -1,21 +1,37 @@
-from abc import ABCMeta, abstractmethod
-from typing import Dict, List, Tuple, Union
+from __future__ import annotations
+
+from abc import ABC, abstractmethod
+from typing import Dict, List, Sequence, Tuple, Union
import numpy as np
+from autosklearn.automl_common.common.utils.backend import Backend
+from autosklearn.data.validation import SUPPORTED_FEAT_TYPES
+from autosklearn.ensemble_building.run import Run
+from autosklearn.metrics import Scorer
from autosklearn.pipeline.base import BasePipeline
-class AbstractEnsemble(object):
- __metaclass__ = ABCMeta
+class AbstractEnsemble(ABC):
+ @abstractmethod
+ def __init__(
+ self,
+ task_type: int,
+ metrics: Sequence[Scorer] | Scorer,
+ random_state: int | np.random.RandomState | None,
+ backend: Backend,
+ ):
+ pass
@abstractmethod
def fit(
self,
- base_models_predictions: np.ndarray,
+ base_models_predictions: np.ndarray | List[np.ndarray],
+ X_data: SUPPORTED_FEAT_TYPES,
true_targets: np.ndarray,
model_identifiers: List[Tuple[int, int, float]],
- ) -> 'AbstractEnsemble':
+ runs: Sequence[Run],
+ ) -> "AbstractEnsemble":
"""Fit an ensemble given predictions of base models and targets.
Ensemble building maximizes performance (in contrast to
@@ -23,15 +39,25 @@ def fit(
Parameters
----------
- base_models_predictions : array of shape = [n_base_models, n_data_points, n_targets]
+ base_models_predictions: np.ndarray
+ shape = (n_base_models, n_data_points, n_targets)
n_targets is the number of classes in case of classification,
n_targets is 0 or 1 in case of regression
+ Can be a list of 2d numpy arrays as well to prevent copying all
+ predictions into a single, large numpy array.
+
+ X_data : list-like or sparse data
+
true_targets : array of shape [n_targets]
model_identifiers : identifier for each base model.
Can be used for practical text output of the ensemble.
+ runs: Sequence[Run]
+ Additional information for each run executed by SMAC that was
+ considered by the ensemble builder.
+
Returns
-------
self
@@ -40,12 +66,15 @@ def fit(
pass
@abstractmethod
- def predict(self, base_models_predictions: Union[np.ndarray, List[np.ndarray]]) -> np.ndarray:
+ def predict(
+ self, base_models_predictions: Union[np.ndarray, List[np.ndarray]]
+ ) -> np.ndarray:
"""Create ensemble predictions from the base model predictions.
Parameters
----------
- base_models_predictions : array of shape = [n_base_models, n_data_points, n_targets]
+ base_models_predictions : np.ndarray
+ shape = (n_base_models, n_data_points, n_targets)
Same as in the fit method.
Returns
@@ -55,8 +84,10 @@ def predict(self, base_models_predictions: Union[np.ndarray, List[np.ndarray]])
pass
@abstractmethod
- def get_models_with_weights(self, models: Dict) -> List[Tuple[float, BasePipeline]]:
- """Return a list of (weight, model) pairs
+ def get_models_with_weights(
+ self, models: Dict[Tuple[int, int, float], BasePipeline]
+ ) -> List[Tuple[float, BasePipeline]]:
+ """Return a list of (weight, model) pairs for all models included in the ensemble.
Parameters
----------
@@ -69,6 +100,24 @@ def get_models_with_weights(self, models: Dict) -> List[Tuple[float, BasePipelin
array : [(weight_1, model_1), ..., (weight_n, model_n)]
"""
+ @abstractmethod
+ def get_identifiers_with_weights(
+ self,
+ ) -> List[Tuple[Tuple[int, int, float], float]]:
+ """Return a (identifier, weight)-pairs for all models that were passed to the
+ ensemble builder.
+
+ Parameters
+ ----------
+ models : dict {identifier : model object}
+ The identifiers are the same as the one presented to the fit()
+ method. Models can be used for nice printing.
+
+ Returns
+ -------
+ array : [(identifier_1, weight_1), ..., (identifier_n, weight_n)]
+ """
+
@abstractmethod
def get_selected_model_identifiers(self) -> List[Tuple[int, int, float]]:
"""Return identifiers of models in the ensemble.
@@ -88,3 +137,8 @@ def get_validation_performance(self) -> float:
------
float
"""
+
+
+class AbstractMultiObjectiveEnsemble(AbstractEnsemble):
+ def get_pareto_set(self) -> Sequence[AbstractEnsemble]:
+ pass
diff --git a/autosklearn/ensembles/ensemble_selection.py b/autosklearn/ensembles/ensemble_selection.py
index 1546c763c2..310f2eb6a4 100644
--- a/autosklearn/ensembles/ensemble_selection.py
+++ b/autosklearn/ensembles/ensemble_selection.py
@@ -1,28 +1,35 @@
+from __future__ import annotations
+
+from typing import Any, Dict, List, Optional, Sequence, Tuple, Union
+
import random
+import warnings
from collections import Counter
-from typing import Any, Dict, List, Optional, Tuple, Union, cast
import numpy as np
-
from sklearn.utils import check_random_state
+from autosklearn.automl_common.common.utils.backend import Backend
from autosklearn.constants import TASK_TYPES
+from autosklearn.data.validation import SUPPORTED_FEAT_TYPES
+from autosklearn.ensemble_building.run import Run
from autosklearn.ensembles.abstract_ensemble import AbstractEnsemble
-from autosklearn.metrics import Scorer, calculate_loss
+from autosklearn.metrics import Scorer, calculate_losses
from autosklearn.pipeline.base import BasePipeline
class EnsembleSelection(AbstractEnsemble):
def __init__(
self,
- ensemble_size: int,
task_type: int,
- metric: Scorer,
+ metrics: Sequence[Scorer] | Scorer,
+ random_state: Optional[Union[int, np.random.RandomState]],
+ backend: Backend,
+ ensemble_size: int = 50,
bagging: bool = False,
- mode: str = 'fast',
- random_state: Optional[Union[int, np.random.RandomState]] = None,
+ mode: str = "fast",
) -> None:
- """ An ensemble of selected algorithms
+ """An ensemble of selected algorithms
Fitting an EnsembleSelection generates an ensemble from the the models
generated during the search process. Can be further used for prediction.
@@ -31,25 +38,50 @@ def __init__(
----------
task_type: int
An identifier indicating which task is being performed.
- metric: Scorer
- The metric used to evaluate the models
+
+ metrics: Sequence[Scorer] | Scorer
+ The metric used to evaluate the models. If multiple metrics are passed,
+ ensemble selection only optimizes for the first
+
+ random_state: Optional[int | RandomState] = None
+ The random_state used for ensemble selection.
+
+ * None - Uses numpy's default RandomState object
+ * int - Successive calls to fit will produce the same results
+ * RandomState - Truly random, each call to fit will produce
+ different results, even with the same object.
+
+ backend : Backend
+ Gives access to the backend of Auto-sklearn. Not used by Ensemble Selection.
+
bagging: bool = False
Whether to use bagging in ensemble selection
+
mode: str in ['fast', 'slow'] = 'fast'
Which kind of ensemble generation to use
- * 'slow' - The original method used in Rich Caruana's ensemble selection.
- * 'fast' - A faster version of Rich Caruanas' ensemble selection.
+ * 'slow' - The original method used in Rich Caruana's ensemble selection.
+ * 'fast' - A faster version of Rich Caruanas' ensemble selection.
- random_state: Optional[int | RandomState] = None
- The random_state used for ensemble selection.
- * None - Uses numpy's default RandomState object
- * int - Successive calls to fit will produce the same results
- * RandomState - Truely random, each call to fit will produce
- different results, even with the same object.
- """
+ References
+ ----------
+ | Ensemble selection from libraries of models
+ | Rich Caruana, Alexandru Niculescu-Mizil, Geoff Crew and Alex Ksikes
+ | ICML 2004
+ | https://dl.acm.org/doi/10.1145/1015330.1015432
+ | https://www.cs.cornell.edu/~caruana/ctp/ct.papers/caruana.icml04.icdm06long.pdf
+ """ # noqa: E501
self.ensemble_size = ensemble_size
self.task_type = task_type
- self.metric = metric
+ if isinstance(metrics, Sequence):
+ if len(metrics) > 1:
+ warnings.warn(
+ "Ensemble selection can only optimize one metric, "
+ "but multiple metrics were passed, dropping all "
+ "except for the first metric."
+ )
+ self.metric = metrics[0]
+ else:
+ self.metric = metrics
self.bagging = bagging
self.mode = mode
@@ -68,51 +100,60 @@ def __getstate__(self) -> Dict[str, Any]:
# one in __main__. we don't use the metric
# in the EnsembleSelection so this should
# be fine
- self.metric = None # type: ignore
- return self.__dict__
+ return {key: value for key, value in self.__dict__.items() if key != "metrics"}
def fit(
self,
- predictions: List[np.ndarray],
- labels: np.ndarray,
- identifiers: List[Tuple[int, int, float]],
- ) -> AbstractEnsemble:
+ base_models_predictions: List[np.ndarray],
+ X_data: SUPPORTED_FEAT_TYPES,
+ true_targets: np.ndarray,
+ model_identifiers: List[Tuple[int, int, float]],
+ runs: Sequence[Run],
+ ) -> EnsembleSelection:
self.ensemble_size = int(self.ensemble_size)
if self.ensemble_size < 1:
- raise ValueError('Ensemble size cannot be less than one!')
+ raise ValueError("Ensemble size cannot be less than one!")
if self.task_type not in TASK_TYPES:
- raise ValueError('Unknown task type %s.' % self.task_type)
+ raise ValueError("Unknown task type %s." % self.task_type)
if not isinstance(self.metric, Scorer):
- raise ValueError("The provided metric must be an instance of Scorer, "
- "nevertheless it is {}({})".format(
- self.metric,
- type(self.metric),
- ))
- if self.mode not in ('fast', 'slow'):
- raise ValueError('Unknown mode %s' % self.mode)
+ raise ValueError(
+ "The provided metric must be an instance of Scorer, "
+ "nevertheless it is {}({})".format(
+ self.metric,
+ type(self.metric),
+ )
+ )
+ if self.mode not in ("fast", "slow"):
+ raise ValueError("Unknown mode %s" % self.mode)
if self.bagging:
- self._bagging(predictions, labels)
+ self._bagging(base_models_predictions, true_targets)
else:
- self._fit(predictions, labels)
+ self._fit(
+ predictions=base_models_predictions,
+ X_data=X_data,
+ labels=true_targets,
+ )
self._calculate_weights()
- self.identifiers_ = identifiers
+ self.identifiers_ = model_identifiers
return self
def _fit(
self,
predictions: List[np.ndarray],
+ X_data: SUPPORTED_FEAT_TYPES,
labels: np.ndarray,
- ) -> AbstractEnsemble:
- if self.mode == 'fast':
- self._fast(predictions, labels)
+ ) -> EnsembleSelection:
+ if self.mode == "fast":
+ self._fast(predictions, X_data, labels)
else:
- self._slow(predictions, labels)
+ self._slow(predictions, X_data, labels)
return self
def _fast(
self,
predictions: List[np.ndarray],
+ X_data: SUPPORTED_FEAT_TYPES,
labels: np.ndarray,
) -> None:
"""Fast version of Rich Caruana's ensemble selection method."""
@@ -149,32 +190,27 @@ def _fast(
# Memory-efficient averaging!
for j, pred in enumerate(predictions):
# fant_ensemble_prediction is the prediction of the current ensemble
- # and should be ([predictions[selected_prev_iterations] + predictions[j])/(s+1)
- # We overwrite the contents of fant_ensemble_prediction
- # directly with weighted_ensemble_prediction + new_prediction and then scale for avg
- np.add(
- weighted_ensemble_prediction,
- pred,
- out=fant_ensemble_prediction
- )
+ # and should be
+ #
+ # ([predictions[selected_prev_iterations] + predictions[j])/(s+1)
+ #
+ # We overwrite the contents of fant_ensemble_prediction directly with
+ # weighted_ensemble_prediction + new_prediction and then scale for avg
+ np.add(weighted_ensemble_prediction, pred, out=fant_ensemble_prediction)
np.multiply(
fant_ensemble_prediction,
- (1. / float(s + 1)),
- out=fant_ensemble_prediction
+ (1.0 / float(s + 1)),
+ out=fant_ensemble_prediction,
)
- # calculate_loss is versatile and can return a dict of losses
- # when scoring_functions=None, we know it will be a float
- losses[j] = cast(
- float,
- calculate_loss(
- solution=labels,
- prediction=fant_ensemble_prediction,
- task_type=self.task_type,
- metric=self.metric,
- scoring_functions=None
- )
- )
+ losses[j] = calculate_losses(
+ solution=labels,
+ prediction=fant_ensemble_prediction,
+ task_type=self.task_type,
+ metrics=[self.metric],
+ X_data=X_data,
+ scoring_functions=None,
+ )[self.metric.name]
all_best = np.argwhere(losses == np.nanmin(losses)).flatten()
@@ -195,7 +231,8 @@ def _fast(
def _slow(
self,
predictions: List[np.ndarray],
- labels: np.ndarray
+ X_data: SUPPORTED_FEAT_TYPES,
+ labels: np.ndarray,
) -> None:
"""Rich Caruana's ensemble selection method."""
self.num_input_models_ = len(predictions)
@@ -214,18 +251,14 @@ def _slow(
for j, pred in enumerate(predictions):
ensemble.append(pred)
ensemble_prediction = np.mean(np.array(ensemble), axis=0)
- # calculate_loss is versatile and can return a dict of losses
- # when scoring_functions=None, we know it will be a float
- losses[j] = cast(
- float,
- calculate_loss(
- solution=labels,
- prediction=ensemble_prediction,
- task_type=self.task_type,
- metric=self.metric,
- scoring_functions=None
- )
- )
+ losses[j] = calculate_losses(
+ solution=labels,
+ prediction=ensemble_prediction,
+ task_type=self.task_type,
+ metrics=[self.metric],
+ X_data=X_data,
+ scoring_functions=None,
+ )[self.metric.name]
ensemble.pop()
best = np.nanargmin(losses)
ensemble.append(predictions[best])
@@ -269,7 +302,7 @@ def _bagging(
n_bags: int = 20,
) -> np.ndarray:
"""Rich Caruana's ensemble selection method with bagging."""
- raise ValueError('Bagging might not work with class-based interface!')
+ raise ValueError("Bagging might not work with class-based interface!")
n_models = predictions.shape[0]
bag_size = int(n_models * fraction)
@@ -286,52 +319,58 @@ def _bagging(
dtype=np.int64,
)
- def predict(self, predictions: Union[np.ndarray, List[np.ndarray]]) -> np.ndarray:
+ def predict(
+ self, base_models_predictions: Union[np.ndarray, List[np.ndarray]]
+ ) -> np.ndarray:
- average = np.zeros_like(predictions[0], dtype=np.float64)
- tmp_predictions = np.empty_like(predictions[0], dtype=np.float64)
+ average = np.zeros_like(base_models_predictions[0], dtype=np.float64)
+ tmp_predictions = np.empty_like(base_models_predictions[0], dtype=np.float64)
# if predictions.shape[0] == len(self.weights_),
# predictions include those of zero-weight models.
- if len(predictions) == len(self.weights_):
- for pred, weight in zip(predictions, self.weights_):
+ if len(base_models_predictions) == len(self.weights_):
+ for pred, weight in zip(base_models_predictions, self.weights_):
np.multiply(pred, weight, out=tmp_predictions)
np.add(average, tmp_predictions, out=average)
# if prediction model.shape[0] == len(non_null_weights),
# predictions do not include those of zero-weight models.
- elif len(predictions) == np.count_nonzero(self.weights_):
+ elif len(base_models_predictions) == np.count_nonzero(self.weights_):
non_null_weights = [w for w in self.weights_ if w > 0]
- for pred, weight in zip(predictions, non_null_weights):
+ for pred, weight in zip(base_models_predictions, non_null_weights):
np.multiply(pred, weight, out=tmp_predictions)
np.add(average, tmp_predictions, out=average)
# If none of the above applies, then something must have gone wrong.
else:
- raise ValueError("The dimensions of ensemble predictions"
- " and ensemble weights do not match!")
+ raise ValueError(
+ "The dimensions of ensemble predictions"
+ " and ensemble weights do not match!"
+ )
del tmp_predictions
return average
def __str__(self) -> str:
- trajectory_str = ' '.join([
- f'{id}: {perf:.5f}'
- for id, perf in enumerate(self.trajectory_)
- ])
- identifiers_str = ' '.join([
- f'{identifier}'
- for idx, identifier in enumerate(self.identifiers_)
- if self.weights_[idx] > 0
- ])
- return ("Ensemble Selection:\n"
- f"\tTrajectory: {trajectory_str}\n"
- f"\tMembers: {self.indices_}\n"
- f"\tWeights: {self.weights_}\n"
- f"\tIdentifiers: {identifiers_str}\n")
+ trajectory_str = " ".join(
+ [f"{id}: {perf:.5f}" for id, perf in enumerate(self.trajectory_)]
+ )
+ identifiers_str = " ".join(
+ [
+ f"{identifier}"
+ for idx, identifier in enumerate(self.identifiers_)
+ if self.weights_[idx] > 0
+ ]
+ )
+ return (
+ "Ensemble Selection:\n"
+ f"\tTrajectory: {trajectory_str}\n"
+ f"\tMembers: {self.indices_}\n"
+ f"\tWeights: {self.weights_}\n"
+ f"\tIdentifiers: {identifiers_str}\n"
+ )
def get_models_with_weights(
- self,
- models: BasePipeline
+ self, models: Dict[Tuple[int, int, float], BasePipeline]
) -> List[Tuple[float, BasePipeline]]:
output = []
for i, weight in enumerate(self.weights_):
@@ -344,6 +383,11 @@ def get_models_with_weights(
return output
+ def get_identifiers_with_weights(
+ self,
+ ) -> List[Tuple[Tuple[int, int, float], float]]:
+ return list(zip(self.identifiers_, self.weights_))
+
def get_selected_model_identifiers(self) -> List[Tuple[int, int, float]]:
output = []
diff --git a/autosklearn/ensembles/singlebest_ensemble.py b/autosklearn/ensembles/singlebest_ensemble.py
index 31a69ae904..497ee4d495 100644
--- a/autosklearn/ensembles/singlebest_ensemble.py
+++ b/autosklearn/ensembles/singlebest_ensemble.py
@@ -1,14 +1,18 @@
+from __future__ import annotations
+
+from typing import Dict, List, Sequence, Tuple, Union
+
import os
-from typing import List, Tuple, Union
import numpy as np
-
from smac.runhistory.runhistory import RunHistory
+from autosklearn.automl_common.common.utils.backend import Backend
+from autosklearn.data.validation import SUPPORTED_FEAT_TYPES
+from autosklearn.ensemble_building.run import Run
from autosklearn.ensembles.abstract_ensemble import AbstractEnsemble
from autosklearn.metrics import Scorer
from autosklearn.pipeline.base import BasePipeline
-from autosklearn.util.backend import Backend
class SingleBest(AbstractEnsemble):
@@ -20,14 +24,23 @@ class SingleBest(AbstractEnsemble):
object, to comply with the expected interface of an
AbstractEnsemble.
"""
+
def __init__(
self,
- metric: Scorer,
+ task_type: int,
+ metrics: Sequence[Scorer] | Scorer,
+ random_state: int | np.random.RandomState | None,
+ backend: Backend,
run_history: RunHistory,
seed: int,
- backend: Backend,
):
- self.metric = metric
+ self.task_type = task_type
+ if isinstance(metrics, Sequence):
+ self.metrics = metrics
+ elif isinstance(metrics, Scorer):
+ self.metrics = [metrics]
+ else:
+ raise TypeError(type(metrics))
self.seed = seed
self.backend = backend
@@ -37,67 +50,95 @@ def __init__(
self.run_history = run_history
self.identifiers_ = self.get_identifiers_from_run_history()
+ def fit(
+ self,
+ base_models_predictions: np.ndarray | List[np.ndarray],
+ X_data: SUPPORTED_FEAT_TYPES,
+ true_targets: np.ndarray,
+ model_identifiers: List[Tuple[int, int, float]],
+ runs: Sequence[Run],
+ ) -> "AbstractEnsemble":
+ return self
+
def get_identifiers_from_run_history(self) -> List[Tuple[int, int, float]]:
- """
- This method parses the run history, to identify
- the best performing model
+ """Parses the run history, to identify the best performing model
- It populates the identifiers attribute, which is used
- by the backend to access the actual model
+ Populates the identifiers attribute, which is used by the backend to access
+ the actual model.
"""
best_model_identifier = []
- best_model_score = self.metric._worst_possible_result
+ best_model_score = self.metrics[0]._worst_possible_result
for run_key in self.run_history.data.keys():
run_value = self.run_history.data[run_key]
- score = self.metric._optimum - (self.metric._sign * run_value.cost)
+ print(run_key, run_value)
+ if len(self.metrics) == 1:
+ cost = run_value.cost
+ else:
+ cost = run_value.cost[0]
+ score = self.metrics[0]._optimum - (self.metrics[0]._sign * cost)
- if (score > best_model_score and self.metric._sign > 0) \
- or (score < best_model_score and self.metric._sign < 0):
+ if (score > best_model_score and self.metrics[0]._sign > 0) or (
+ score < best_model_score and self.metrics[0]._sign < 0
+ ):
# Make sure that the individual best model actually exists
model_dir = self.backend.get_numrun_directory(
self.seed,
- run_value.additional_info['num_run'],
+ run_value.additional_info["num_run"],
run_key.budget,
)
model_file_name = self.backend.get_model_filename(
self.seed,
- run_value.additional_info['num_run'],
+ run_value.additional_info["num_run"],
run_key.budget,
)
file_path = os.path.join(model_dir, model_file_name)
if not os.path.exists(file_path):
continue
- best_model_identifier = [(
- self.seed,
- run_value.additional_info['num_run'],
- run_key.budget,
- )]
+ best_model_identifier = [
+ (
+ self.seed,
+ run_value.additional_info["num_run"],
+ run_key.budget,
+ )
+ ]
best_model_score = score
if not best_model_identifier:
raise ValueError(
- "No valid model found in run history. This means smac was not able to fit"
- " a valid model. Please check the log file for errors."
+ "No valid model found in run history. This means smac was not able to"
+ " fit a valid model. Please check the log file for errors."
)
+ self.best_model_score_ = best_model_score
+
return best_model_identifier
def predict(self, predictions: Union[np.ndarray, List[np.ndarray]]) -> np.ndarray:
return predictions[0]
def __str__(self) -> str:
- return 'Single Model Selection:\n\tMembers: %s' \
- '\n\tWeights: %s\n\tIdentifiers: %s' % \
- (self.indices_, self.weights_,
- ' '.join([str(identifier) for idx, identifier in
- enumerate(self.identifiers_)
- if self.weights_[idx] > 0]))
-
- def get_models_with_weights(self, models: BasePipeline
- ) -> List[Tuple[float, BasePipeline]]:
+ return (
+ "Single Model Selection:\n\tMembers: %s"
+ "\n\tWeights: %s\n\tIdentifiers: %s"
+ % (
+ self.indices_,
+ self.weights_,
+ " ".join(
+ [
+ str(identifier)
+ for idx, identifier in enumerate(self.identifiers_)
+ if self.weights_[idx] > 0
+ ]
+ ),
+ )
+ )
+
+ def get_models_with_weights(
+ self, models: Dict[Tuple[int, int, float], BasePipeline]
+ ) -> List[Tuple[float, BasePipeline]]:
output = []
for i, weight in enumerate(self.weights_):
if weight > 0.0:
@@ -109,6 +150,11 @@ def get_models_with_weights(self, models: BasePipeline
return output
+ def get_identifiers_with_weights(
+ self,
+ ) -> List[Tuple[Tuple[int, int, float], float]]:
+ return list(zip(self.identifiers_, self.weights_))
+
def get_selected_model_identifiers(self) -> List[Tuple[int, int, float]]:
output = []
@@ -118,3 +164,6 @@ def get_selected_model_identifiers(self) -> List[Tuple[int, int, float]]:
output.append(identifier)
return output
+
+ def get_validation_performance(self) -> float:
+ return self.best_model_score_
diff --git a/autosklearn/estimators.py b/autosklearn/estimators.py
index 87aa2be317..5afd8c597c 100644
--- a/autosklearn/estimators.py
+++ b/autosklearn/estimators.py
@@ -1,42 +1,62 @@
# -*- encoding: utf-8 -*-
-from typing import Optional, Dict, List, Tuple, Union, Iterable
-from typing_extensions import Literal
+from __future__ import annotations
+
+from typing import (
+ Any,
+ Dict,
+ Iterable,
+ List,
+ Mapping,
+ Optional,
+ Sequence,
+ Tuple,
+ Type,
+ Union,
+)
+
+import warnings
-from ConfigSpace.configuration_space import Configuration, ConfigurationSpace
import dask.distributed
import joblib
import numpy as np
import pandas as pd
+from ConfigSpace.configuration_space import Configuration, ConfigurationSpace
from scipy.sparse import spmatrix
from sklearn.base import BaseEstimator, ClassifierMixin, RegressorMixin
+from sklearn.ensemble import VotingClassifier, VotingRegressor
from sklearn.utils.multiclass import type_of_target
from smac.runhistory.runhistory import RunInfo, RunValue
+from typing_extensions import Literal
+from autosklearn.automl import AutoML, AutoMLClassifier, AutoMLRegressor
from autosklearn.data.validation import (
- convert_if_sparse,
SUPPORTED_FEAT_TYPES,
SUPPORTED_TARGET_TYPES,
+ convert_if_sparse,
)
-from autosklearn.pipeline.base import BasePipeline
-from autosklearn.automl import AutoMLClassifier, AutoMLRegressor, AutoML
+from autosklearn.ensembles.abstract_ensemble import AbstractEnsemble
+from autosklearn.ensembles.ensemble_selection import EnsembleSelection
from autosklearn.metrics import Scorer
+from autosklearn.pipeline.base import BasePipeline
+from autosklearn.util.smac_wrap import SMACCallback
class AutoSklearnEstimator(BaseEstimator):
-
def __init__(
self,
time_left_for_this_task=3600,
per_run_time_limit=None,
initial_configurations_via_metalearning=25,
- ensemble_size: int = 50,
+ ensemble_size: int | None = None,
+ ensemble_class: Type[AbstractEnsemble] | None = EnsembleSelection,
+ ensemble_kwargs: Dict[str, Any] | None = None,
ensemble_nbest=50,
max_models_on_disc=50,
seed=1,
memory_limit=3072,
- include=None,
- exclude=None,
- resampling_strategy='holdout',
+ include: Optional[Dict[str, List[str]]] = None,
+ exclude: Optional[Dict[str, List[str]]] = None,
+ resampling_strategy="holdout",
resampling_strategy_arguments=None,
tmp_folder=None,
delete_tmp_folder_after_terminate=True,
@@ -47,10 +67,12 @@ def __init__(
smac_scenario_args=None,
logging_config=None,
metadata_directory=None,
- metric=None,
+ metric: Scorer | Sequence[Scorer] | None = None,
scoring_functions: Optional[List[Scorer]] = None,
load_models: bool = True,
- get_trials_callback=None
+ get_trials_callback: SMACCallback | None = None,
+ dataset_compression: Union[bool, Mapping[str, Any]] = True,
+ allow_string_features: bool = True,
):
"""
Parameters
@@ -73,14 +95,31 @@ def __init__(
datasets. Disable if the hyperparameter optimization algorithm
should start from scratch.
- ensemble_size : int, optional (default=50)
+ ensemble_size : int, optional
Number of models added to the ensemble built by *Ensemble
selection from libraries of models*. Models are drawn with
- replacement.
+ replacement. If set to ``0`` no ensemble is fit.
+
+ Deprecated - will be removed in Auto-sklearn 0.16. Please pass
+ this argument via ``ensemble_kwargs={"ensemble_size": int}``
+ if you want to change the ensemble size for ensemble selection.
+
+ ensemble_class : Type[AbstractEnsemble], optional (default=EnsembleSelection)
+ Class implementing the post-hoc ensemble algorithm. Set to
+ ``None`` to disable ensemble building or use ``SingleBest``
+ to obtain only use the single best model instead of an
+ ensemble.
+
+ ensemble_kwargs : Dict, optional
+ Keyword arguments that are passed to the ensemble class upon
+ initialization.
ensemble_nbest : int, optional (default=50)
Only consider the ``ensemble_nbest`` models when building an
- ensemble.
+ ensemble. This is inspired by a concept called library pruning
+ introduced in `Getting Most out of Ensemble Selection`. This
+ is independent of the ``ensemble_class`` argument and this
+ pruning step is done prior to constructing an ensemble.
max_models_on_disc: int, optional (default=50),
Defines the maximum number of models that are kept in the disc.
@@ -96,59 +135,100 @@ def __init__(
memory_limit : int, optional (3072)
Memory limit in MB for the machine learning algorithm.
`auto-sklearn` will stop fitting the machine learning algorithm if
- it tries to allocate more than `memory_limit` MB.
- If None is provided, no memory limit is set.
- In case of multi-processing, `memory_limit` will be per job.
- This memory limit also applies to the ensemble creation process.
-
- include : dict, optional (None)
- If None, all possible algorithms are used. Otherwise specifies
- set of algorithms for each added component is used. Include and
- exclude are incompatible if used together on the same component
-
- exclude : dict, optional (None)
- If None, all possible algorithms are used. Otherwise specifies
- set of algorithms for each added component is not used.
- Incompatible with include. Include and exclude are incompatible
- if used together on the same component
-
- resampling_strategy : string or object, optional ('holdout')
- how to to handle overfitting, might need 'resampling_strategy_arguments'
-
- * 'holdout': 67:33 (train:test) split
- * 'holdout-iterative-fit': 67:33 (train:test) split, calls iterative
- fit where possible
- * 'cv': crossvalidation, requires 'folds'
- * 'cv-iterative-fit': crossvalidation, calls iterative fit where possible
- * 'partial-cv': crossvalidation with intensification, requires
- 'folds'
- * BaseCrossValidator object: any BaseCrossValidator class found
- in scikit-learn model_selection module
- * _RepeatedSplits object: any _RepeatedSplits class found
- in scikit-learn model_selection module
- * BaseShuffleSplit object: any BaseShuffleSplit class found
- in scikit-learn model_selection module
-
- resampling_strategy_arguments : dict, optional if 'holdout' (train_size default=0.67)
- Additional arguments for resampling_strategy:
-
- * ``train_size`` should be between 0.0 and 1.0 and represent the
- proportion of the dataset to include in the train split.
- * ``shuffle`` determines whether the data is shuffled prior to
- splitting it into train and validation.
-
- Available arguments:
-
- * 'holdout': {'train_size': float}
- * 'holdout-iterative-fit': {'train_size': float}
- * 'cv': {'folds': int}
- * 'cv-iterative-fit': {'folds': int}
- * 'partial-cv': {'folds': int, 'shuffle': bool}
- * BaseCrossValidator or _RepeatedSplits or BaseShuffleSplit object: all arguments
- required by chosen class as specified in scikit-learn documentation.
- If arguments are not provided, scikit-learn defaults are used.
- If no defaults are available, an exception is raised.
- Refer to the 'n_splits' argument as 'folds'.
+ it tries to allocate more than ``memory_limit`` MB.
+
+ **Important notes:**
+
+ * If ``None`` is provided, no memory limit is set.
+ * In case of multi-processing, ``memory_limit`` will be *per job*, so the total usage is
+ ``n_jobs x memory_limit``.
+ * The memory limit also applies to the ensemble creation process.
+
+ include : Optional[Dict[str, List[str]]] = None
+ If None, all possible algorithms are used.
+
+ Otherwise, specifies a step and the components that are included in search.
+ See ``/pipeline/components//*`` for available components.
+
+ Incompatible with parameter ``exclude``.
+
+ **Possible Steps**:
+
+ * ``"data_preprocessor"``
+ * ``"balancing"``
+ * ``"feature_preprocessor"``
+ * ``"classifier"`` - Only for when when using ``AutoSklearnClasssifier``
+ * ``"regressor"`` - Only for when when using ``AutoSklearnRegressor``
+
+ **Example**:
+
+ .. code-block:: python
+
+ include = {
+ 'classifier': ["random_forest"],
+ 'feature_preprocessor': ["no_preprocessing"]
+ }
+
+ exclude : Optional[Dict[str, List[str]]] = None
+ If None, all possible algorithms are used.
+
+ Otherwise, specifies a step and the components that are excluded from search.
+ See ``/pipeline/components//*`` for available components.
+
+ Incompatible with parameter ``include``.
+
+ **Possible Steps**:
+
+ * ``"data_preprocessor"``
+ * ``"balancing"``
+ * ``"feature_preprocessor"``
+ * ``"classifier"`` - Only for when when using ``AutoSklearnClasssifier``
+ * ``"regressor"`` - Only for when when using ``AutoSklearnRegressor``
+
+ **Example**:
+
+ .. code-block:: python
+
+ exclude = {
+ 'classifier': ["random_forest"],
+ 'feature_preprocessor': ["no_preprocessing"]
+ }
+
+ resampling_strategy : str | BaseCrossValidator | _RepeatedSplits | BaseShuffleSplit = "holdout"
+ How to to handle overfitting, might need to use ``resampling_strategy_arguments``
+ if using ``"cv"`` based method or a Splitter object.
+
+ * **Options**
+ * ``"holdout"`` - Use a 67:33 (train:test) split
+ * ``"cv"``: perform cross validation, requires "folds" in ``resampling_strategy_arguments``
+ * ``"holdout-iterative-fit"`` - Same as "holdout" but iterative fit where possible
+ * ``"cv-iterative-fit"``: Same as "cv" but iterative fit where possible
+ * ``"partial-cv"``: Same as "cv" but uses intensification.
+ * ``BaseCrossValidator`` - any BaseCrossValidator subclass (found in scikit-learn model_selection module)
+ * ``_RepeatedSplits`` - any _RepeatedSplits subclass (found in scikit-learn model_selection module)
+ * ``BaseShuffleSplit`` - any BaseShuffleSplit subclass (found in scikit-learn model_selection module)
+
+ If using a Splitter object that relies on the dataset retaining it's current
+ size and order, you will need to look at the ``dataset_compression`` argument
+ and ensure that ``"subsample"`` is not included in the applied compression
+ ``"methods"`` or disable it entirely with ``False``.
+
+ resampling_strategy_arguments : Optional[Dict] = None
+ Additional arguments for ``resampling_strategy``, this is required if
+ using a ``cv`` based strategy. The default arguments if left as ``None``
+ are:
+
+ .. code-block:: python
+
+ {
+ "train_size": 0.67, # The size of the training set
+ "shuffle": True, # Whether to shuffle before splitting data
+ "folds": 5 # Used in 'cv' based resampling strategies
+ }
+
+ If using a custom splitter class, which takes ``n_splits`` such as
+ `PredefinedSplit `_,
+ the value of ``"folds"`` will be used.
tmp_folder : string, optional (None)
folder to store configuration output and log files, if ``None``
@@ -160,13 +240,15 @@ def __init__(
n_jobs : int, optional, experimental
The number of jobs to run in parallel for ``fit()``. ``-1`` means
- using all processors. By default, Auto-sklearn uses a single core
- for fitting the machine learning model and a single core for fitting
- an ensemble. Ensemble building is not affected by ``n_jobs`` but
- can be controlled by the number of models in the ensemble. In
- contrast to most scikit-learn models, ``n_jobs`` given in the
- constructor is not applied to the ``predict()`` method. If
- ``dask_client`` is None, a new dask client is created.
+ using all processors.
+
+ **Important notes**:
+
+ * By default, Auto-sklearn uses one core.
+ * Ensemble building is not affected by ``n_jobs`` but can be controlled by the number
+ of models in the ensemble.
+ * ``predict()`` is not affected by ``n_jobs`` (in contrast to most scikit-learn models)
+ * If ``dask_client`` is ``None``, a new dask client is created.
dask_client : dask.distributed.Client, optional
User-created dask client, can be used to start a dask cluster and then
@@ -180,9 +262,9 @@ def __init__(
list are:
* ``'y_optimization'`` : do not save the predictions for the
- optimization/validation set, which would later on be used to build
- an ensemble.
- * ``'model'`` : do not save any model files
+ optimization set, which would later on be used to build an ensemble.
+
+ * ``model`` : do not save any model files
smac_scenario_args : dict, optional (None)
Additional arguments inserted into the scenario of SMAC. See the
@@ -218,23 +300,89 @@ def __init__(
load_models : bool, optional (True)
Whether to load the models after fitting Auto-sklearn.
-
+
get_trials_callback: callable
- Callback function to create an object of subclass defined in module
- `smac.callbacks `_.
- This is an advanced feature. Use only if you are familiar with
- `SMAC `_.
+ A callable with the following definition.
+
+ * (smac.SMBO, smac.RunInfo, smac.RunValue, time_left: float) -> bool | None
+
+ This will be called after SMAC, the underlying optimizer for autosklearn,
+ finishes training each run.
+
+ You can use this to record your own information about the optimization
+ process. You can also use this to enable a early stopping based on some
+ critera.
+
+ See the example:
+ :ref:`Early Stopping And Callbacks `.
+
+ dataset_compression: Union[bool, Mapping[str, Any]] = True
+ We compress datasets so that they fit into some predefined amount of memory.
+ Currently this does not apply to dataframes or sparse arrays, only to raw
+ numpy arrays.
+
+ **NOTE** - If using a custom ``resampling_strategy`` that relies on specific
+ size or ordering of data, this must be disabled to preserve these properties.
+
+ You can disable this entirely by passing ``False`` or leave as the default
+ ``True`` for configuration below.
+
+ .. code-block:: python
+
+ {
+ "memory_allocation": 0.1,
+ "methods": ["precision", "subsample"]
+ }
+
+ You can also pass your own configuration with the same keys and choosing
+ from the available ``"methods"``.
+
+ The available options are described here:
+
+ * **memory_allocation**
+ By default, we attempt to fit the dataset into ``0.1 * memory_limit``.
+ This float value can be set with ``"memory_allocation": 0.1``.
+ We also allow for specifying absolute memory in MB, e.g. 10MB is
+ ``"memory_allocation": 10``.
+
+ The memory used by the dataset is checked after each reduction method is
+ performed. If the dataset fits into the allocated memory, any further
+ methods listed in ``"methods"`` will not be performed.
+
+ For example, if ``methods: ["precision", "subsample"]`` and the
+ ``"precision"`` reduction step was enough to make the dataset fit into
+ memory, then the ``"subsample"`` reduction step will not be performed.
+
+ * **methods**
+ We provide the following methods for reducing the dataset size.
+ These can be provided in a list and are performed in the order as given.
+
+ * ``"precision"`` - We reduce floating point precision as follows:
+ * ``np.float128 -> np.float64``
+ * ``np.float96 -> np.float64``
+ * ``np.float64 -> np.float32``
+
+ * ``subsample`` - We subsample data such that it **fits directly into
+ the memory allocation** ``memory_allocation * memory_limit``.
+ Therefore, this should likely be the last method listed in
+ ``"methods"``.
+ Subsampling takes into account classification labels and stratifies
+ accordingly. We guarantee that at least one occurrence of each
+ label is included in the sampled set.
+
+ allow_string_features: bool = True
+ Whether autosklearn should process string features. By default the
+ textpreprocessing is enabled.
Attributes
----------
-
- cv_results\_ : dict of numpy (masked) ndarrays
+ cv_results_ : dict of numpy (masked) ndarrays
A dict with keys as column headers and values as columns, that can be
imported into a pandas ``DataFrame``.
Not all keys returned by scikit-learn are supported yet.
- performance_over_time\_ : pandas.core.frame.DataFrame
+ performance_over_time_ : pandas.core.frame.DataFrame
A ``DataFrame`` containing the models performance over time data. Can be
used for plotting directly. Please refer to the example
:ref:`Train and Test Inputs `.
@@ -242,12 +390,56 @@ def __init__(
""" # noqa (links are too long)
# Raise error if the given total time budget is less than 30 seconds.
if time_left_for_this_task < 30:
- raise ValueError("Time left for this task must be at least "
- "30 seconds. ")
+ raise ValueError("Time left for this task must be at least " "30 seconds. ")
self.time_left_for_this_task = time_left_for_this_task
self.per_run_time_limit = per_run_time_limit
- self.initial_configurations_via_metalearning = initial_configurations_via_metalearning
- self.ensemble_size = ensemble_size
+ self.initial_configurations_via_metalearning = (
+ initial_configurations_via_metalearning
+ )
+ self.ensemble_class = ensemble_class
+
+ # User specified `ensemble_size` explicitly, warn them about deprecation
+ if ensemble_size is not None:
+ # Keep consistent behaviour
+ message = (
+ "`ensemble_size` has been deprecated, please use `ensemble_kwargs = "
+ "{'ensemble_size': %d}`. Inserting `ensemble_size` into "
+ "`ensemble_kwargs` for now. `ensemble_size` will be removed in "
+ "auto-sklearn 0.16."
+ ) % ensemble_size
+ if ensemble_class == EnsembleSelection:
+ if ensemble_kwargs is None:
+ ensemble_kwargs = {"ensemble_size": ensemble_size}
+ warnings.warn(message, DeprecationWarning, stacklevel=2)
+ elif "ensemble_size" not in ensemble_kwargs:
+ ensemble_kwargs["ensemble_size"] = ensemble_size
+ warnings.warn(message, DeprecationWarning, stacklevel=2)
+ else:
+ warnings.warn(
+ "Deprecated argument `ensemble_size` is both provided "
+ "as an argument to the constructor and passed inside "
+ "`ensemble_kwargs`. Will ignore the argument and use "
+ "the value given in `ensemble_kwargs` (%d). `ensemble_size` "
+ "will be removed in auto-sklearn 0.16."
+ % ensemble_kwargs["ensemble_size"],
+ DeprecationWarning,
+ stacklevel=2,
+ )
+ else:
+ warnings.warn(
+ "`ensemble_size` has been deprecated, please use "
+ "`ensemble_kwargs = {'ensemble_size': %d} if this "
+ "was intended. Ignoring `ensemble_size` because "
+ "`ensemble_class` != EnsembleSelection. "
+ "`ensemble_size` will be removed in auto-sklearn 0.16."
+ % ensemble_size,
+ DeprecationWarning,
+ stacklevel=2,
+ )
+ self.ensemble_size = (
+ ensemble_size # Otherwise sklean.base.get_params() will complain
+ )
+ self.ensemble_kwargs = ensemble_kwargs
self.ensemble_nbest = ensemble_nbest
self.max_models_on_disc = max_models_on_disc
self.seed = seed
@@ -269,6 +461,8 @@ def __init__(
self.scoring_functions = scoring_functions
self.load_models = load_models
self.get_trials_callback = get_trials_callback
+ self.dataset_compression = dataset_compression
+ self.allow_string_features = allow_string_features
self.automl_ = None # type: Optional[AutoML]
@@ -290,13 +484,15 @@ def __getstate__(self):
def build_automl(self):
+ initial_configs = self.initial_configurations_via_metalearning
automl = self._get_automl_class()(
temporary_directory=self.tmp_folder,
delete_tmp_folder_after_terminate=self.delete_tmp_folder_after_terminate,
time_left_for_this_task=self.time_left_for_this_task,
per_run_time_limit=self.per_run_time_limit,
- initial_configurations_via_metalearning=self.initial_configurations_via_metalearning,
- ensemble_size=self.ensemble_size,
+ initial_configurations_via_metalearning=initial_configs,
+ ensemble_class=self.ensemble_class,
+ ensemble_kwargs=self.ensemble_kwargs,
ensemble_nbest=self.ensemble_nbest,
max_models_on_disc=self.max_models_on_disc,
seed=self.seed,
@@ -312,9 +508,11 @@ def build_automl(self):
smac_scenario_args=self.smac_scenario_args,
logging_config=self.logging_config,
metadata_directory=self.metadata_directory,
- metric=self.metric,
+ metrics=[self.metric] if isinstance(self.metric, Scorer) else self.metric,
scoring_functions=self.scoring_functions,
- get_trials_callback=self.get_trials_callback
+ get_trials_callback=self.get_trials_callback,
+ dataset_compression=self.dataset_compression,
+ allow_string_features=self.allow_string_features,
)
return automl
@@ -335,7 +533,7 @@ def fit_pipeline(
self,
X: SUPPORTED_FEAT_TYPES,
y: Union[SUPPORTED_TARGET_TYPES, spmatrix],
- config: Union[Configuration, Dict[str, Union[str, float, int]]],
+ config: Union[Configuration, Dict[str, Union[str, float, int]]],
dataset_name: Optional[str] = None,
X_test: Optional[SUPPORTED_FEAT_TYPES] = None,
y_test: Optional[Union[SUPPORTED_TARGET_TYPES, spmatrix]] = None,
@@ -343,7 +541,7 @@ def fit_pipeline(
*args,
**kwargs: Dict,
) -> Tuple[Optional[BasePipeline], RunInfo, RunValue]:
- """ Fits and individual pipeline configuration and returns
+ """Fits and individual pipeline configuration and returns
the result to the user.
The Estimator constraints are honored, for example the resampling
@@ -352,7 +550,8 @@ def fit_pipeline(
arguments are redirected to the TAE evaluation function, which allows for
further customization while building a pipeline.
- Any additional argument provided is directly passed to the worker exercising the run.
+ Any additional argument provided is directly passed to the
+ worker exercising the run.
Parameters
----------
@@ -366,7 +565,7 @@ def fit_pipeline(
If provided, the testing performance will be tracked on this labels
config: Union[Configuration, Dict[str, Union[str, float, int]]]
A configuration object used to define the pipeline steps.
- If a dictionary is passed, a configuration is created based on this dictionary.
+ If a dict is passed, a configuration is created based on this dict.
dataset_name: Optional[str]
Name that will be used to tag the Auto-Sklearn run and identify the
Auto-Sklearn run
@@ -390,24 +589,35 @@ def fit_pipeline(
"""
if self.automl_ is None:
self.automl_ = self.build_automl()
- return self.automl_.fit_pipeline(X=X, y=y,
- dataset_name=dataset_name,
- config=config,
- feat_type=feat_type,
- X_test=X_test, y_test=y_test,
- *args, **kwargs)
-
- def fit_ensemble(self, y, task=None, precision=32,
- dataset_name=None, ensemble_nbest=None,
- ensemble_size=None):
+ return self.automl_.fit_pipeline(
+ X=X,
+ y=y,
+ dataset_name=dataset_name,
+ config=config,
+ feat_type=feat_type,
+ X_test=X_test,
+ y_test=y_test,
+ *args,
+ **kwargs,
+ )
+
+ def fit_ensemble(
+ self,
+ y,
+ task: int = None,
+ precision: Literal[16, 21, 64] = 32,
+ dataset_name: Optional[str] = None,
+ ensemble_size: int | None = None,
+ ensemble_kwargs: Optional[Dict[str, Any]] = None,
+ ensemble_nbest: Optional[int] = None,
+ ensemble_class: Optional[AbstractEnsemble] = EnsembleSelection,
+ metrics: Scorer | Sequence[Scorer] | None = None,
+ ):
"""Fit an ensemble to models trained during an optimization process.
All parameters are ``None`` by default. If no other value is given,
the default values which were set in a call to ``fit()`` are used.
- Calling this function is only necessary if ``ensemble_size==0``, for
- example when executing *auto-sklearn* in parallel.
-
Parameters
----------
y : array-like
@@ -418,26 +628,85 @@ def fit_ensemble(self, y, task=None, precision=32,
the task type (binary classification, multiclass classification,
multilabel classification or regression).
- precision : str
+ precision : int
Numeric precision used when loading ensemble data. Can be either
- ``'16'``, ``'32'`` or ``'64'``.
+ ``16``, ``32`` or ``64``.
dataset_name : str
Name of the current data set.
+ ensemble_size : int, optional
+ Number of models added to the ensemble built by *Ensemble
+ selection from libraries of models*. Models are drawn with
+ replacement. If set to ``0`` no ensemble is fit.
+
+ Deprecated - will be removed in Auto-sklearn 0.16. Please pass
+ this argument via ``ensemble_kwargs={"ensemble_size": int}``
+ if you want to change the ensemble size for ensemble selection.
+
+ ensemble_kwargs : Dict, optional
+ Keyword arguments that are passed to the ensemble class upon
+ initialization.
+
ensemble_nbest : int
- Determines how many models should be considered from the ensemble
- building. This is inspired by a concept called library pruning
- introduced in `Getting Most out of Ensemble Selection`.
+ Only consider the ``ensemble_nbest`` models when building an
+ ensemble. This is inspired by a concept called library pruning
+ introduced in `Getting Most out of Ensemble Selection`. This
+ is independent of the ``ensemble_class`` argument and this
+ pruning step is done prior to constructing an ensemble.
+
+ ensemble_class : Type[AbstractEnsemble], optional (default=EnsembleSelection)
+ Class implementing the post-hoc ensemble algorithm. Set to
+ ``None`` to disable ensemble building or use ``SingleBest``
+ to obtain only use the single best model instead of an
+ ensemble.
- ensemble_size : int
- Size of the ensemble built by `Ensemble Selection`.
+ metrics: Scorer | Sequence[Scorer] | None = None
+ A metric or list of metrics to score the ensemble with
Returns
-------
self
-
"""
+ # User specified `ensemble_size` explicitly, warn them about deprecation
+ if ensemble_size is not None:
+ # Keep consistent behaviour
+ message = (
+ "`ensemble_size` has been deprecated, please use `ensemble_kwargs = "
+ "{'ensemble_size': %d}`. Inserting `ensemble_size` into "
+ "`ensemble_kwargs` for now. `ensemble_size` will be removed in "
+ "auto-sklearn 0.16."
+ ) % ensemble_size
+ if ensemble_class == EnsembleSelection:
+ if ensemble_kwargs is None:
+ ensemble_kwargs = {"ensemble_size": ensemble_size}
+ warnings.warn(message, DeprecationWarning, stacklevel=2)
+ elif "ensemble_size" not in ensemble_kwargs:
+ ensemble_kwargs["ensemble_size"] = ensemble_size
+ warnings.warn(message, DeprecationWarning, stacklevel=2)
+ else:
+ warnings.warn(
+ "Deprecated argument `ensemble_size` is both provided "
+ "as an argument to the constructor and passed inside "
+ "`ensemble_kwargs`. Will ignore the argument and use "
+ "the value given in `ensemble_kwargs` (%d). `ensemble_size` "
+ "will be removed in auto-sklearn 0.16."
+ % ensemble_kwargs["ensemble_size"],
+ DeprecationWarning,
+ stacklevel=2,
+ )
+ else:
+ warnings.warn(
+ "`ensemble_size` has been deprecated, please use "
+ "`ensemble_kwargs = {'ensemble_size': %d} if this "
+ "was intended. Ignoring `ensemble_size` because "
+ "`ensemble_class` != EnsembleSelection. "
+ "`ensemble_size` will be removed in auto-sklearn 0.16."
+ % ensemble_size,
+ DeprecationWarning,
+ stacklevel=2,
+ )
+
if self.automl_ is None:
# Build a dummy automl object to call fit_ensemble
# The ensemble size is honored in the .automl_.fit_ensemble
@@ -449,7 +718,9 @@ def fit_ensemble(self, y, task=None, precision=32,
precision=precision,
dataset_name=dataset_name,
ensemble_nbest=ensemble_nbest,
- ensemble_size=ensemble_size,
+ ensemble_class=ensemble_class,
+ ensemble_kwargs=ensemble_kwargs,
+ metrics=metrics,
)
return self
@@ -485,20 +756,93 @@ def predict(self, X, batch_size=None, n_jobs=1):
return self.automl_.predict(X, batch_size=batch_size, n_jobs=n_jobs)
def predict_proba(self, X, batch_size=None, n_jobs=1):
- return self.automl_.predict_proba(
- X, batch_size=batch_size, n_jobs=n_jobs)
+ return self.automl_.predict_proba(X, batch_size=batch_size, n_jobs=n_jobs)
def score(self, X, y):
return self.automl_.score(X, y)
def show_models(self):
- """Return a representation of the final ensemble found by auto-sklearn.
+ """Returns a dictionary containing dictionaries of ensemble models.
+
+ Each model in the ensemble can be accessed by giving its ``model_id`` as key.
+
+ A model dictionary contains the following:
+
+ * ``"model_id"`` - The id given to a model by ``autosklearn``.
+
+ * ``"rank"`` - The rank of the model based on it's ``"cost"``.
+
+ * ``"cost"`` - The loss of the model on the validation set.
+
+ * ``"ensemble_weight"`` - The weight given to the model in the ensemble.
+
+ * ``"voting_model"`` - The ``cv_voting_ensemble`` model (for 'cv' resampling).
+
+ * ``"estimators"`` - List of models (dicts) in ``cv_voting_ensemble``
+ ('cv' resampling).
+
+ * ``"data_preprocessor"`` - The preprocessor used on the data.
+
+ * ``"balancing"`` - The balancing used on the data (for classification).
+
+ * ``"feature_preprocessor"`` - The preprocessor for features types.
+
+ * ``"classifier"`` / ``"regressor"``
+ - The autosklearn wrapped classifier or regressor.
+
+ * ``"sklearn_classifier"`` or ``"sklearn_regressor"``
+ - The sklearn classifier or regressor.
+
+ **Example**
+
+ .. code-block:: python
+
+ import sklearn.datasets
+ import sklearn.metrics
+ import autosklearn.regression
+
+ X, y = sklearn.datasets.load_diabetes(return_X_y=True)
+
+ automl = autosklearn.regression.AutoSklearnRegressor(
+ time_left_for_this_task=120
+ )
+ automl.fit(X_train, y_train, dataset_name='diabetes')
+
+ ensemble_dict = automl.show_models()
+ print(ensemble_dict)
+
+ Output:
+
+ .. code-block:: text
+
+ {
+ 25: {'model_id': 25.0,
+ 'rank': 1,
+ 'cost': 0.43667876507897496,
+ 'ensemble_weight': 0.38,
+ 'data_preprocessor': ,
+ 'feature_preprocessor': ,
+ 'regressor': ,
+ 'sklearn_regressor': SGDRegressor(alpha=0.0006517033225329654,...)
+ },
+ 6: {'model_id': 6.0,
+ 'rank': 2,
+ 'cost': 0.4550418898836528,
+ 'ensemble_weight': 0.3,
+ 'data_preprocessor': ,
+ 'feature_preprocessor': ,
+ 'regressor': ,
+ 'sklearn_regressor': ARDRegression(alpha_1=0.0003701926442639788,...)
+ }...
+ }
Returns
-------
- str
+ Dict(int, Any) : dictionary of length = number of models in the ensemble
+ A dictionary of models in the ensemble, where ``model_id`` is the key.
+
+ """ # noqa: E501
- """
return self.automl_.show_models()
def get_models_with_weights(self):
@@ -549,17 +893,17 @@ def leaderboard(
self,
detailed: bool = False,
ensemble_only: bool = True,
- top_k: Union[int, Literal['all']] = 'all',
- sort_by: str = 'cost',
- sort_order: Literal['auto', 'ascending', 'descending'] = 'auto',
- include: Optional[Union[str, Iterable[str]]] = None
+ top_k: Union[int, Literal["all"]] = "all",
+ sort_by: str = "cost",
+ sort_order: Literal["auto", "ascending", "descending"] = "auto",
+ include: Optional[Union[str, Iterable[str]]] = None,
) -> pd.DataFrame:
- """ Returns a pandas table of results for all evaluated models.
+ """Returns a pandas table of results for all evaluated models.
Gives an overview of all models trained during the search process along
with various statistics about their training.
- The availble statistics are:
+ The available statistics are:
**Simple**:
@@ -600,6 +944,9 @@ def leaderboard(
What column to sort by. If that column is not present, the
sorting defaults to the ``"model_id"`` index column.
+ Defaults to the metric optimized. Sort by the first objective
+ in case of a multi-objective optimization problem
+
sort_order: "auto" or "ascending" or "descending" = "auto"
Which sort order to apply to the ``sort_by`` column. If left
as ``"auto"``, it will sort by a sensible default where "better" is
@@ -624,81 +971,121 @@ def leaderboard(
# TODO validate that `self` is fitted. This is required for
# self.ensemble_ to get the identifiers of models it will generate
# weights for.
- column_types = AutoSklearnEstimator._leaderboard_columns()
+ num_metrics = (
+ 1
+ if self.metric is None or isinstance(self.metric, Scorer)
+ else len(self.metric)
+ )
+ column_types = AutoSklearnEstimator._leaderboard_columns(num_metrics)
+ if num_metrics == 1:
+ multi_objective_cost_names = []
+ else:
+ multi_objective_cost_names = [f"cost_{i}" for i in range(num_metrics)]
# Validation of top_k
if (
not (isinstance(top_k, str) or isinstance(top_k, int))
- or (isinstance(top_k, str) and top_k != 'all')
+ or (isinstance(top_k, str) and top_k != "all")
or (isinstance(top_k, int) and top_k <= 0)
):
- raise ValueError(f"top_k={top_k} must be a positive integer or pass"
- " `top_k`='all' to view results for all models")
+ raise ValueError(
+ f"top_k={top_k} must be a positive integer or pass"
+ " `top_k`='all' to view results for all models"
+ )
# Validate columns to include
if isinstance(include, str):
include = [include]
- if include == ['model_id']:
- raise ValueError('Must provide more than just `model_id`')
+ if include == ["model_id"]:
+ raise ValueError("Must provide more than just `model_id`")
if include is not None:
columns = [*include]
# 'model_id' should always be present as it is the unique index
# used for pandas
- if 'model_id' not in columns:
- columns.append('model_id')
+ if "model_id" not in columns:
+ columns.append("model_id")
- invalid_include_items = set(columns) - set(column_types['all'])
+ invalid_include_items = set(columns) - set(column_types["all"])
if len(invalid_include_items) != 0:
- raise ValueError(f"Values {invalid_include_items} are not known"
- f" columns to include, must be contained in "
- f"{column_types['all']}")
+ raise ValueError(
+ f"Values {invalid_include_items} are not known"
+ f" columns to include, must be contained in "
+ f"{column_types['all']}"
+ )
elif detailed:
- columns = column_types['all']
+ columns = column_types["all"]
else:
- columns = column_types['simple']
+ columns = column_types["simple"]
# Validation of sorting
- if sort_by not in column_types['all']:
- raise ValueError(f"sort_by='{sort_by}' must be one of included "
- f"columns {set(column_types['all'])}")
-
- valid_sort_orders = ['auto', 'ascending', 'descending']
+ if sort_by == "cost":
+ sort_by_cost = True
+ if num_metrics == 1:
+ sort_by = ["cost", "model_id"]
+ else:
+ sort_by = multi_objective_cost_names + ["model_id"]
+ else:
+ sort_by_cost = False
+ if isinstance(sort_by, str):
+ if sort_by not in column_types["all"]:
+ raise ValueError(
+ f"sort_by='{sort_by}' must be one of included "
+ f"columns {set(column_types['all'])}"
+ )
+ elif len(set(sort_by) - set(column_types["all"])) > 0:
+ too_much = set(sort_by) - set(column_types["all"])
+ raise ValueError(
+ f"sort_by='{too_much}' must be in the included columns "
+ f"{set(column_types['all'])}"
+ )
+
+ valid_sort_orders = ["auto", "ascending", "descending"]
if not (isinstance(sort_order, str) and sort_order in valid_sort_orders):
- raise ValueError(f"`sort_order` = {sort_order} must be a str in "
- f"{valid_sort_orders}")
+ raise ValueError(
+ f"`sort_order` = {sort_order} must be a str in " f"{valid_sort_orders}"
+ )
# To get all the models that were optmized, we collect what we can from
# runhistory first.
- def has_key(rv, key):
+ def additional_info_has_key(rv, key):
return rv.additional_info and key in rv.additional_info
- model_runs = {
- rval.additional_info['num_run']: {
- 'model_id': rval.additional_info['num_run'],
- 'seed': rkey.seed,
- 'budget': rkey.budget,
- 'duration': rval.time,
- 'config_id': rkey.config_id,
- 'start_time': rval.starttime,
- 'end_time': rval.endtime,
- 'status': str(rval.status),
- 'cost': rval.cost,
- 'train_loss': rval.additional_info['train_loss']
- if has_key(rval, 'train_loss') else None,
- 'config_origin': rval.additional_info['configuration_origin']
- if has_key(rval, 'configuration_origin') else None
- }
- for rkey, rval in self.automl_.runhistory_.data.items()
- if has_key(rval, 'num_run')
- }
+ model_runs = {}
+ for run_key, run_val in self.automl_.runhistory_.data.items():
+ if not additional_info_has_key(run_val, "num_run"):
+ continue
+ else:
+ model_key = run_val.additional_info["num_run"]
+ model_run = {
+ "model_id": run_val.additional_info["num_run"],
+ "seed": run_key.seed,
+ "budget": run_key.budget,
+ "duration": run_val.time,
+ "config_id": run_key.config_id,
+ "start_time": run_val.starttime,
+ "end_time": run_val.endtime,
+ "status": str(run_val.status),
+ "train_loss": run_val.additional_info["train_loss"]
+ if additional_info_has_key(run_val, "train_loss")
+ else None,
+ "config_origin": run_val.additional_info["configuration_origin"]
+ if additional_info_has_key(run_val, "configuration_origin")
+ else None,
+ }
+ if num_metrics == 1:
+ model_run["cost"] = run_val.cost
+ else:
+ for cost_idx, cost in enumerate(run_val.cost):
+ model_run[f"cost_{cost_idx}"] = cost
+ model_runs[model_key] = model_run
# Next we get some info about the model itself
model_class_strings = {
- AutoMLClassifier: 'classifier',
- AutoMLRegressor: 'regressor'
+ AutoMLClassifier: "classifier",
+ AutoMLRegressor: "regressor",
}
model_type = model_class_strings.get(self._get_automl_class(), None)
if model_type is None:
@@ -708,31 +1095,58 @@ def has_key(rv, key):
configurations = self.automl_.runhistory_.ids_config
for model_id, run_info in model_runs.items():
- config_id = run_info['config_id']
+ config_id = run_info["config_id"]
run_config = configurations[config_id]._values
- run_info.update({
- 'balancing_strategy': run_config.get('balancing:strategy', None),
- 'type': run_config[f'{model_type}:__choice__'],
- 'data_preprocessors': [
- value for key, value in run_config.items()
- if 'data_preprocessing' in key and '__choice__' in key
- ],
- 'feature_preprocessors': [
- value for key, value in run_config.items()
- if 'feature_preprocessor' in key and '__choice__' in key
- ]
- })
+ run_info.update(
+ {
+ "balancing_strategy": run_config.get("balancing:strategy", None),
+ "type": run_config[f"{model_type}:__choice__"],
+ "data_preprocessors": [
+ value
+ for key, value in run_config.items()
+ if "data_preprocessing" in key and "__choice__" in key
+ ],
+ "feature_preprocessors": [
+ value
+ for key, value in run_config.items()
+ if "feature_preprocessor" in key and "__choice__" in key
+ ],
+ }
+ )
# Get the models ensemble weight if it has one
- # TODO both implementing classes of AbstractEnsemble have a property
- # `identifiers_` and `weights_`, might be good to put it as an
- # abstract property
- # TODO `ensemble_.identifiers_` and `ensemble_.weights_` are loosely
- # tied together by ordering, might be better to store as tuple
- for i, weight in enumerate(self.automl_.ensemble_.weights_):
- (_, model_id, _) = self.automl_.ensemble_.identifiers_[i]
- model_runs[model_id]['ensemble_weight'] = weight
+ for (
+ _,
+ model_id,
+ _,
+ ), weight in self.automl_.ensemble_.get_identifiers_with_weights():
+
+ # We had issues where the model's in the ensembles are not in the runhistory
+ # collected. I have no clue why this is but to prevent failures, we fill
+ # the values with NaN
+ if model_id not in model_runs:
+ model_run = {
+ "model_id": model_id,
+ "seed": pd.NA,
+ "budget": pd.NA,
+ "duration": pd.NA,
+ "config_id": pd.NA,
+ "start_time": pd.NA,
+ "end_time": pd.NA,
+ "status": pd.NA,
+ "train_loss": pd.NA,
+ "config_origin": pd.NA,
+ "type": pd.NA,
+ }
+ if num_metrics == 1:
+ model_run["cost"] = pd.NA
+ else:
+ for cost_idx in range(num_metrics):
+ model_run[f"cost_{cost_idx}"] = pd.NA
+ model_runs[model_id] = model_run
+
+ model_runs[model_id]["ensemble_weight"] = weight
# Filter out non-ensemble members if needed, else fill in a default
# value of 0 if it's missing
@@ -740,65 +1154,97 @@ def has_key(rv, key):
model_runs = {
model_id: info
for model_id, info in model_runs.items()
- if ('ensemble_weight' in info and info['ensemble_weight'] > 0)
+ if ("ensemble_weight" in info and info["ensemble_weight"] > 0)
}
else:
for model_id, info in model_runs.items():
- if 'ensemble_weight' not in info:
- info['ensemble_weight'] = 0
+ if "ensemble_weight" not in info:
+ info["ensemble_weight"] = 0
# `rank` relies on `cost` so we include `cost`
# We drop it later if it's not requested
- if 'rank' in columns and 'cost' not in columns:
- columns = [*columns, 'cost']
+ if "rank" in columns:
+ if num_metrics == 1 and "cost" not in columns:
+ columns = [*columns, "cost"]
+ elif num_metrics > 1 and any(
+ cost_name not in columns for cost_name in multi_objective_cost_names
+ ):
+ columns = columns + list(multi_objective_cost_names)
# Finally, convert into a tabular format by converting the dict into
# column wise orientation.
- dataframe = pd.DataFrame({
- col: [run_info[col] for run_info in model_runs.values()]
- for col in columns if col != 'rank'
- })
+ dataframe = pd.DataFrame(
+ {
+ col: [run_info[col] for run_info in model_runs.values()]
+ for col in columns
+ if col != "rank"
+ }
+ )
# Give it an index, even if not in the `include`
- dataframe.set_index('model_id', inplace=True)
+ dataframe.set_index("model_id", inplace=True)
- # Add the `rank` column if needed, dropping `cost` if it's not
+ # Add the `rank` column if needed
# requested by the user
- if 'rank' in columns:
- dataframe.sort_values(by='cost', ascending=True, inplace=True)
- dataframe.insert(column='rank',
- value=range(1, len(dataframe) + 1),
- loc=list(columns).index('rank') - 1) # account for `model_id`
-
- if 'cost' not in columns:
- dataframe.drop('cost', inplace=True)
+ if "rank" in columns:
+ if num_metrics == 1:
+ dataframe.sort_values(by="cost", ascending=True, inplace=True)
+ else:
+ dataframe.sort_values(by="cost_0", ascending=True, inplace=True)
+ dataframe.insert(
+ column="rank",
+ value=range(1, len(dataframe) + 1),
+ loc=list(columns).index("rank") - 1,
+ ) # account for `model_id`
# Decide on the sort order depending on what it gets sorted by
- descending_columns = ['ensemble_weight', 'duration']
- if sort_order == 'auto':
- ascending_param = False if sort_by in descending_columns else True
+ descending_columns = ["ensemble_weight", "duration"]
+ if sort_order == "auto":
+ ascending_param = [
+ False if sby in descending_columns else True for sby in sort_by
+ ]
else:
- ascending_param = False if sort_order == 'descending' else True
+ ascending_param = False if sort_order == "descending" else True
# Sort by the given column name, defaulting to 'model_id' if not present
- if sort_by not in dataframe.columns:
- self.automl_._logger.warning(f"sort_by = '{sort_by}' was not present"
- ", defaulting to sort on the index "
- "'model_id'")
- sort_by = 'model_id'
-
- # Cost can be the same but leave rank all over the place
- if 'rank' in columns and sort_by == 'cost':
- dataframe.sort_values(by=[sort_by, 'rank'],
- ascending=[ascending_param, True],
- inplace=True)
+ if (
+ (not sort_by_cost and len(set(sort_by) - set(dataframe.columns)) > 0)
+ or (sort_by_cost and "cost" not in dataframe.columns)
+ or (
+ sort_by_cost
+ and any(
+ cost_name not in dataframe.columns
+ for cost_name in multi_objective_cost_names
+ )
+ )
+ ):
+ self.automl_._logger.warning(
+ f"sort_by = '{sort_by}' was not present"
+ ", defaulting to sort on the index "
+ "'model_id'"
+ )
+ sort_by = "model_id"
+ sort_by_cost = False
+ ascending_param = True
+
+ # Single objective
+ if sort_by_cost:
+ dataframe.sort_values(
+ by=sort_by, ascending=[True] * len(sort_by), inplace=True
+ )
+ else:
+ dataframe.sort_values(by=sort_by, ascending=ascending_param, inplace=True)
+
+ if num_metrics == 1:
+ if "cost" not in columns and "cost" in dataframe.columns:
+ dataframe.drop("cost", inplace=True)
else:
- dataframe.sort_values(by=sort_by,
- ascending=ascending_param,
- inplace=True)
+ for cost_name in multi_objective_cost_names:
+ if cost_name not in columns and cost_name in dataframe.columns:
+ dataframe.drop(cost_name, inplace=True)
# Lastly, just grab the top_k
- if top_k == 'all' or top_k >= len(dataframe):
+ if top_k == "all" or top_k >= len(dataframe):
top_k = len(dataframe)
dataframe = dataframe.head(top_k)
@@ -806,18 +1252,41 @@ def has_key(rv, key):
return dataframe
@staticmethod
- def _leaderboard_columns() -> Dict[Literal['all', 'simple', 'detailed'], List[str]]:
- all = [
- "model_id", "rank", "ensemble_weight", "type", "cost", "duration",
- "config_id", "train_loss", "seed", "start_time", "end_time",
- "budget", "status", "data_preprocessors", "feature_preprocessors",
- "balancing_strategy", "config_origin"
- ]
- simple = [
- "model_id", "rank", "ensemble_weight", "type", "cost", "duration"
- ]
+ def _leaderboard_columns(
+ num_metrics: int,
+ ) -> Dict[Literal["all", "simple", "detailed"], List[str]]:
+ if num_metrics == 1:
+ cost_list = ["cost"]
+ else:
+ cost_list = [f"cost_{i}" for i in range(num_metrics)]
+ all = (
+ [
+ "model_id",
+ "rank",
+ "ensemble_weight",
+ "type",
+ ]
+ + cost_list
+ + [
+ "duration",
+ "config_id",
+ "train_loss",
+ "seed",
+ "start_time",
+ "end_time",
+ "budget",
+ "status",
+ "data_preprocessors",
+ "feature_preprocessors",
+ "balancing_strategy",
+ "config_origin",
+ ]
+ )
+ simple = (
+ ["model_id", "rank", "ensemble_weight", "type"] + cost_list + ["duration"]
+ )
detailed = all
- return {'all': all, 'detailed': detailed, 'simple': simple}
+ return {"all": all, "detailed": detailed, "simple": simple}
def _get_automl_class(self):
raise NotImplementedError()
@@ -852,34 +1321,35 @@ def get_configuration_space(
if self.automl_ is None:
self.automl_ = self.build_automl()
- return self.automl_.fit(
- X, y,
- X_test=X_test, y_test=y_test,
- dataset_name=dataset_name,
- feat_type=feat_type,
- only_return_configuration_space=True,
- ) if self.automl_.configuration_space is None else self.automl_.configuration_space
+ return (
+ self.automl_.fit(
+ X,
+ y,
+ X_test=X_test,
+ y_test=y_test,
+ dataset_name=dataset_name,
+ feat_type=feat_type,
+ only_return_configuration_space=True,
+ )
+ if self.automl_.configuration_space is None
+ else self.automl_.configuration_space
+ )
+ def get_pareto_set(self) -> Sequence[VotingClassifier | VotingRegressor]:
+ return self.automl_._load_pareto_set()
-class AutoSklearnClassifier(AutoSklearnEstimator, ClassifierMixin):
- """
- This class implements the classification task.
- """
+class AutoSklearnClassifier(AutoSklearnEstimator, ClassifierMixin):
+ """This class implements the classification task."""
- def fit(self, X, y,
- X_test=None,
- y_test=None,
- feat_type=None,
- dataset_name=None):
+ def fit(self, X, y, X_test=None, y_test=None, feat_type=None, dataset_name=None):
"""Fit *auto-sklearn* to given training set (X, y).
Fit both optimizes the machine learning models and builds an ensemble
- out of them. To disable ensembling, set ``ensemble_size==0``.
+ out of them.
Parameters
----------
-
X : array-like or sparse matrix of shape = [n_samples, n_features]
The training input samples.
@@ -911,7 +1381,6 @@ def fit(self, X, y,
Returns
-------
self
-
"""
# AutoSklearn does not handle sparse y for now
y = convert_if_sparse(y)
@@ -920,18 +1389,16 @@ def fit(self, X, y,
# type of data is compatible with auto-sklearn. Legal target
# types are: binary, multiclass, multilabel-indicator.
target_type = type_of_target(y)
- supported_types = ['binary', 'multiclass', 'multilabel-indicator']
+ supported_types = ["binary", "multiclass", "multilabel-indicator"]
if target_type not in supported_types:
- raise ValueError("Classification with data of type {} is "
- "not supported. Supported types are {}. "
- "You can find more information about scikit-learn "
- "data types in: "
- "https://scikit-learn.org/stable/modules/multiclass.html"
- "".format(
- target_type,
- supported_types
- )
- )
+ raise ValueError(
+ "Classification with data of type {} is "
+ "not supported. Supported types are {}. "
+ "You can find more information about scikit-learn "
+ "data types in: "
+ "https://scikit-learn.org/stable/modules/multiclass.html"
+ "".format(target_type, supported_types)
+ )
# remember target type for using in predict_proba later.
self.target_type = target_type
@@ -963,12 +1430,10 @@ def predict(self, X, batch_size=None, n_jobs=1):
-------
y : array of shape = [n_samples] or [n_samples, n_labels]
The predicted classes.
-
"""
return super().predict(X, batch_size=batch_size, n_jobs=n_jobs)
def predict_proba(self, X, batch_size=None, n_jobs=1):
-
"""Predict probabilities of classes for all samples X.
Parameters
@@ -984,24 +1449,20 @@ def predict_proba(self, X, batch_size=None, n_jobs=1):
-------
y : array of shape = [n_samples, n_classes] or [n_samples, n_labels]
The predicted class probabilities.
-
"""
- pred_proba = super().predict_proba(
- X, batch_size=batch_size, n_jobs=n_jobs)
+ pred_proba = super().predict_proba(X, batch_size=batch_size, n_jobs=n_jobs)
# Check if all probabilities sum up to 1.
# Assert only if target type is not multilabel-indicator.
- if self.target_type not in ['multilabel-indicator']:
- assert(
- np.allclose(
- np.sum(pred_proba, axis=1),
- np.ones_like(pred_proba[:, 0]))
+ if self.target_type not in ["multilabel-indicator"]:
+ assert np.allclose(
+ np.sum(pred_proba, axis=1), np.ones_like(pred_proba[:, 0])
), "prediction probability does not sum up to 1!"
# Check that all probability values lie between 0 and 1.
- assert(
- (pred_proba >= 0).all() and (pred_proba <= 1).all()
- ), "found prediction probability value outside of [0, 1]!"
+ assert (pred_proba >= 0).all() and (
+ pred_proba <= 1
+ ).all(), "found prediction probability value outside of [0, 1]!"
return pred_proba
@@ -1015,15 +1476,11 @@ class AutoSklearnRegressor(AutoSklearnEstimator, RegressorMixin):
"""
- def fit(self, X, y,
- X_test=None,
- y_test=None,
- feat_type=None,
- dataset_name=None):
+ def fit(self, X, y, X_test=None, y_test=None, feat_type=None, dataset_name=None):
"""Fit *Auto-sklearn* to given training set (X, y).
Fit both optimizes the machine learning models and builds an ensemble
- out of them. To disable ensembling, set ``ensemble_size==0``.
+ out of them.
Parameters
----------
@@ -1067,18 +1524,21 @@ def fit(self, X, y,
y = convert_if_sparse(y)
target_type = type_of_target(y)
- supported_types = ['continuous', 'binary', 'multiclass', 'continuous-multioutput']
+ supported_types = [
+ "continuous",
+ "binary",
+ "multiclass",
+ "continuous-multioutput",
+ ]
if target_type not in supported_types:
- raise ValueError("Regression with data of type {} is "
- "not supported. Supported types are {}. "
- "You can find more information about scikit-learn "
- "data types in: "
- "https://scikit-learn.org/stable/modules/multiclass.html"
- "".format(
- target_type,
- supported_types
- )
- )
+ raise ValueError(
+ "Regression with data of type {} is "
+ "not supported. Supported types are {}. "
+ "You can find more information about scikit-learn "
+ "data types in: "
+ "https://scikit-learn.org/stable/modules/multiclass.html"
+ "".format(target_type, supported_types)
+ )
# Fit is supposed to be idempotent!
# But not if we use share_mode.
diff --git a/autosklearn/evaluation/__init__.py b/autosklearn/evaluation/__init__.py
index 589535d085..ba17513ae0 100644
--- a/autosklearn/evaluation/__init__.py
+++ b/autosklearn/evaluation/__init__.py
@@ -1,41 +1,55 @@
# -*- encoding: utf-8 -*-
+from __future__ import annotations
+
+from typing import (
+ Any,
+ Callable,
+ Dict,
+ List,
+ Optional,
+ Sequence,
+ Tuple,
+ Type,
+ Union,
+ cast,
+)
+
import functools
-import logging
import json
+import logging
import math
import multiprocessing
-from queue import Empty
import time
import traceback
-from typing import Any, Callable, Dict, List, Optional, Tuple, Type, Union, cast
+from queue import Empty
-from ConfigSpace import Configuration
import numpy as np
import pynisher
+from ConfigSpace import Configuration
+from sklearn.model_selection._split import (
+ BaseCrossValidator,
+ BaseShuffleSplit,
+ _RepeatedSplits,
+)
from smac.runhistory.runhistory import RunInfo, RunValue
from smac.stats.stats import Stats
from smac.tae import StatusType, TAEAbortException
from smac.tae.execute_func import AbstractTAFunc
-from sklearn.model_selection._split import _RepeatedSplits, BaseShuffleSplit,\
- BaseCrossValidator
-from autosklearn.metrics import Scorer
-
-import autosklearn.evaluation.train_evaluator
import autosklearn.evaluation.test_evaluator
+import autosklearn.evaluation.train_evaluator
import autosklearn.evaluation.util
import autosklearn.pipeline.components
+from autosklearn.automl_common.common.utils.backend import Backend
from autosklearn.evaluation.train_evaluator import TYPE_ADDITIONAL_INFO
-from autosklearn.util.backend import Backend
+from autosklearn.metrics import Scorer
from autosklearn.util.logging_ import PickableLoggerAdapter, get_named_client_logger
from autosklearn.util.parallel import preload_modules
def fit_predict_try_except_decorator(
- ta: Callable,
- queue: multiprocessing.Queue,
- cost_for_crash: float,
- **kwargs: Any) -> None:
+ ta: Callable, queue: multiprocessing.Queue, cost_for_crash: float, **kwargs: Any
+) -> None:
try:
return ta(queue=queue, **kwargs)
@@ -47,7 +61,8 @@ def fit_predict_try_except_decorator(
exception_traceback = traceback.format_exc()
error_message = repr(e)
- # Printing stuff to stdout just in case the queue doesn't work, which happened with the
+ # Printing stuff to stdout just in case the queue doesn't work,
+ # which happened with the
# following traceback:
# File "auto-sklearn/autosklearn/evaluation/__init__.py", line 29, in fit_predict_try_except_decorator # noqa E501
# return ta(queue=queue, **kwargs)
@@ -56,45 +71,62 @@ def fit_predict_try_except_decorator(
# File "auto-sklearn/autosklearn/evaluation/train_evaluator.py", line 616, in fit_predict_and_loss, # noqa E501
# status=status
# File "auto-sklearn/autosklearn/evaluation/abstract_evaluator.py", line 320, in finish_up # noqa E501
- # self.queue.put(rval_dict)
+ # self.queue.put(return_value_dict)
# File "miniconda/3-4.5.4/envs/autosklearn/lib/python3.7/multiprocessing/queues.py", line 87, in put # noqa E501
# self._start_thread()
# File "miniconda/3-4.5.4/envs/autosklearn/lib/python3.7/multiprocessing/queues.py", line 170, in _start_thread # noqa E501
# self._thread.start()
# File "miniconda/3-4.5.4/envs/autosklearn/lib/python3.7/threading.py", line 847, in start # noqa E501
# RuntimeError: can't start new thread
- print("Exception handling in `fit_predict_try_except_decorator`: "
- "traceback: %s \nerror message: %s" % (exception_traceback, error_message))
-
- queue.put({'loss': cost_for_crash,
- 'additional_run_info': {'traceback': exception_traceback,
- 'error': error_message},
- 'status': StatusType.CRASHED,
- 'final_queue_element': True}, block=True)
- queue.close()
-
+ print(
+ "Exception handling in `fit_predict_try_except_decorator`: "
+ "traceback: %s \nerror message: %s" % (exception_traceback, error_message)
+ )
-def get_cost_of_crash(metric: Scorer) -> float:
+ queue.put(
+ {
+ "loss": cost_for_crash,
+ "additional_run_info": {
+ "traceback": exception_traceback,
+ "error": error_message,
+ },
+ "status": StatusType.CRASHED,
+ "final_queue_element": True,
+ },
+ block=True,
+ )
+ queue.close()
- # The metric must always be defined to extract optimum/worst
- if not isinstance(metric, Scorer):
- raise ValueError("The metric must be stricly be an instance of Scorer")
- # Autosklearn optimizes the err. This function translates
- # worst_possible_result to be a minimization problem.
- # For metrics like accuracy that are bounded to [0,1]
- # metric.optimum==1 is the worst cost.
- # A simple guide is to use greater_is_better embedded as sign
- if metric._sign < 0:
- worst_possible_result = metric._worst_possible_result
- else:
- worst_possible_result = metric._optimum - metric._worst_possible_result
+def get_cost_of_crash(metrics: Sequence[Scorer]) -> List[float] | float:
+ """Return the cost of crash.
+
+ Return value can be either a list (multi-objective optimization) or a
+ raw float (single objective) because SMAC assumes different types in the
+ two different cases.
+ """
+ costs = []
+ for metric in metrics:
+ if not isinstance(metric, Scorer):
+ raise ValueError("The metric {metric} must be an instance of Scorer")
+
+ # Autosklearn optimizes the err. This function translates
+ # worst_possible_result to be a minimization problem.
+ # For metrics like accuracy that are bounded to [0,1]
+ # metric.optimum==1 is the worst cost.
+ # A simple guide is to use greater_is_better embedded as sign
+ if metric._sign < 0:
+ worst_possible_result = metric._worst_possible_result
+ else:
+ worst_possible_result = metric._optimum - metric._worst_possible_result
+ costs.append(worst_possible_result)
- return worst_possible_result
+ return costs if len(costs) > 1 else costs[0]
-def _encode_exit_status(exit_status: Union[str, int, Type[BaseException]]
- ) -> Union[str, int]:
+def _encode_exit_status(
+ exit_status: Union[str, int, Type[BaseException]]
+) -> Union[str, int]:
try:
# If it can be dumped, then it is int
exit_status = cast(int, exit_status)
@@ -107,20 +139,22 @@ def _encode_exit_status(exit_status: Union[str, int, Type[BaseException]]
# TODO potentially log all inputs to this class to pickle them in order to do
# easier debugging of potential crashes
class ExecuteTaFuncWithQueue(AbstractTAFunc):
-
def __init__(
self,
backend: Backend,
autosklearn_seed: int,
- resampling_strategy: Union[str, BaseCrossValidator, _RepeatedSplits, BaseShuffleSplit],
- metric: Scorer,
+ resampling_strategy: Union[
+ str, BaseCrossValidator, _RepeatedSplits, BaseShuffleSplit
+ ],
+ metrics: Sequence[Scorer],
cost_for_crash: float,
abort_on_first_run_crash: bool,
port: int,
pynisher_context: str,
+ multi_objectives: List[str],
initial_num_run: int = 1,
stats: Optional[Stats] = None,
- run_obj: str = 'quality',
+ run_obj: str = "quality",
par_factor: int = 1,
scoring_functions: Optional[List[Scorer]] = None,
output_y_hat_optimization: bool = True,
@@ -130,30 +164,32 @@ def __init__(
disable_file_output: bool = False,
init_params: Optional[Dict[str, Any]] = None,
budget_type: Optional[str] = None,
- ta: Optional[Callable] = None,
+ ta: Optional[Callable] = None, # Required by SMAC's parent class
**resampling_strategy_args: Any,
):
-
- if resampling_strategy == 'holdout':
+ if resampling_strategy == "holdout":
eval_function = autosklearn.evaluation.train_evaluator.eval_holdout
- elif resampling_strategy == 'holdout-iterative-fit':
- eval_function = autosklearn.evaluation.train_evaluator.eval_iterative_holdout
- elif resampling_strategy == 'cv-iterative-fit':
+ elif resampling_strategy == "holdout-iterative-fit":
+ eval_function = (
+ autosklearn.evaluation.train_evaluator.eval_iterative_holdout
+ )
+ elif resampling_strategy == "cv-iterative-fit":
eval_function = autosklearn.evaluation.train_evaluator.eval_iterative_cv
- elif resampling_strategy == 'cv' or isinstance(resampling_strategy, (
- BaseCrossValidator, _RepeatedSplits, BaseShuffleSplit)
+ elif resampling_strategy == "cv" or isinstance(
+ resampling_strategy, (BaseCrossValidator, _RepeatedSplits, BaseShuffleSplit)
):
eval_function = autosklearn.evaluation.train_evaluator.eval_cv
- elif resampling_strategy == 'partial-cv':
+ elif resampling_strategy == "partial-cv":
eval_function = autosklearn.evaluation.train_evaluator.eval_partial_cv
- elif resampling_strategy == 'partial-cv-iterative-fit':
- eval_function = autosklearn.evaluation.train_evaluator.eval_partial_cv_iterative
- elif resampling_strategy == 'test':
+ elif resampling_strategy == "partial-cv-iterative-fit":
+ eval_function = (
+ autosklearn.evaluation.train_evaluator.eval_partial_cv_iterative
+ )
+ elif resampling_strategy == "test":
eval_function = autosklearn.evaluation.test_evaluator.eval_t
output_y_hat_optimization = False
else:
- raise ValueError('Unknown resampling strategy %s' %
- resampling_strategy)
+ raise ValueError("Unknown resampling strategy %s" % resampling_strategy)
self.worst_possible_result = cost_for_crash
@@ -170,17 +206,18 @@ def __init__(
par_factor=par_factor,
cost_for_crash=self.worst_possible_result,
abort_on_first_run_crash=abort_on_first_run_crash,
+ multi_objectives=multi_objectives,
)
self.backend = backend
self.autosklearn_seed = autosklearn_seed
self.resampling_strategy = resampling_strategy
self.initial_num_run = initial_num_run
- self.metric = metric
+ self.metrics = metrics
self.resampling_strategy = resampling_strategy
self.resampling_strategy_args = resampling_strategy_args
self.scoring_functions = scoring_functions
- # TODO deactivate output_y_hat_optimization and let the respective evaluator decide
+ # TODO deactivate output_y_hat_optimization and let respective evaluator decide
self.output_y_hat_optimization = output_y_hat_optimization
self.include = include
self.exclude = exclude
@@ -193,19 +230,14 @@ def __init__(
self.memory_limit = memory_limit
dm = self.backend.load_datamanager()
- if 'X_valid' in dm.data and 'Y_valid' in dm.data:
- self._get_validation_loss = True
- else:
- self._get_validation_loss = False
- if 'X_test' in dm.data and 'Y_test' in dm.data:
- self._get_test_loss = True
- else:
- self._get_test_loss = False
+ self._get_test_loss = "X_test" in dm.data and "Y_test" in dm.data
self.port = port
self.pynisher_context = pynisher_context
if self.port is None:
- self.logger: Union[logging.Logger, PickableLoggerAdapter] = logging.getLogger("TAE")
+ self.logger: Union[
+ logging.Logger, PickableLoggerAdapter
+ ] = logging.getLogger("TAE")
else:
self.logger = get_named_client_logger(
name="TAE",
@@ -235,18 +267,23 @@ def run_wrapper(
if self.budget_type is None:
if run_info.budget != 0:
raise ValueError(
- 'If budget_type is None, budget must be.0, but is %f' % run_info.budget
+ "If budget_type is None, budget must be.0, but is %f"
+ % run_info.budget
)
else:
if run_info.budget == 0:
run_info = run_info._replace(budget=100)
elif run_info.budget <= 0 or run_info.budget > 100:
- raise ValueError('Illegal value for budget, must be >0 and <=100, but is %f' %
- run_info.budget)
- if self.budget_type not in ('subsample', 'iterations', 'mixed'):
- raise ValueError("Illegal value for budget type, must be one of "
- "('subsample', 'iterations', 'mixed'), but is : %s" %
- self.budget_type)
+ raise ValueError(
+ "Illegal value for budget, must be >0 and <=100, but is %f"
+ % run_info.budget
+ )
+ if self.budget_type not in ("subsample", "iterations", "mixed"):
+ raise ValueError(
+ "Illegal value for budget type, must be one of "
+ "('subsample', 'iterations', 'mixed'), but is : %s"
+ % self.budget_type
+ )
remaining_time = self.stats.get_remaing_time_budget()
@@ -254,11 +291,15 @@ def run_wrapper(
run_info = run_info._replace(cutoff=int(remaining_time - 5))
config_id = (
- run_info.config if isinstance(run_info.config, int) else run_info.config.config_id
+ run_info.config
+ if isinstance(run_info.config, int)
+ else run_info.config.config_id
)
if run_info.cutoff < 1.0:
- self.logger.info("Not starting configuration %d because time is up" % config_id)
+ self.logger.info(
+ "Not starting configuration %d because time is up" % config_id
+ )
return run_info, RunValue(
status=StatusType.STOP,
cost=self.worst_possible_result,
@@ -267,9 +308,8 @@ def run_wrapper(
starttime=time.time(),
endtime=time.time(),
)
- elif (
- run_info.cutoff != int(np.ceil(run_info.cutoff))
- and not isinstance(run_info.cutoff, int)
+ elif run_info.cutoff != int(np.ceil(run_info.cutoff)) and not isinstance(
+ run_info.cutoff, int
):
run_info = run_info._replace(cutoff=int(np.ceil(run_info.cutoff)))
@@ -284,7 +324,12 @@ def run(
seed: int = 12345,
budget: float = 0.0,
instance_specific: Optional[str] = None,
- ) -> Tuple[StatusType, float, float, Dict[str, Union[int, float, str, Dict, List, Tuple]]]:
+ ) -> Tuple[
+ StatusType,
+ float | list[float],
+ float,
+ Dict[str, Union[int, float, str, Dict, List, Tuple]],
+ ]:
# Additional information of each of the tae executions
# Defined upfront for mypy
@@ -294,14 +339,17 @@ def run(
preload_modules(context)
queue = context.Queue()
- if not (instance_specific is None or instance_specific == '0'):
+ if not (instance_specific is None or instance_specific == "0"):
raise ValueError(instance_specific)
- init_params = {'instance': instance}
+
+ init_params = {"instance": instance}
if self.init_params is not None:
init_params.update(self.init_params)
if self.port is None:
- logger: Union[logging.Logger, PickableLoggerAdapter] = logging.getLogger("pynisher")
+ logger: Union[logging.Logger, PickableLoggerAdapter] = logging.getLogger(
+ "pynisher"
+ )
else:
logger = get_named_client_logger(
name="pynisher",
@@ -325,7 +373,7 @@ def run(
config=config,
backend=self.backend,
port=self.port,
- metric=self.metric,
+ metrics=self.metrics,
seed=self.autosklearn_seed,
num_run=num_run,
scoring_functions=self.scoring_functions,
@@ -340,9 +388,9 @@ def run(
additional_components=autosklearn.pipeline.components.base._addons,
)
- if self.resampling_strategy != 'test':
- obj_kwargs['resampling_strategy'] = self.resampling_strategy
- obj_kwargs['resampling_strategy_args'] = self.resampling_strategy_args
+ if self.resampling_strategy != "test":
+ obj_kwargs["resampling_strategy"] = self.resampling_strategy
+ obj_kwargs["resampling_strategy_args"] = self.resampling_strategy_args
try:
obj = pynisher.enforce_limits(**arguments)(self.ta)
@@ -350,31 +398,38 @@ def run(
except Exception as e:
exception_traceback = traceback.format_exc()
error_message = repr(e)
- additional_run_info.update({
- 'traceback': exception_traceback,
- 'error': error_message
- })
- return StatusType.CRASHED, self.worst_possible_result, 0.0, additional_run_info
+ additional_run_info.update(
+ {"traceback": exception_traceback, "error": error_message}
+ )
+ return (
+ StatusType.CRASHED,
+ self.worst_possible_result,
+ 0.0,
+ additional_run_info,
+ )
- if obj.exit_status in (pynisher.TimeoutException, pynisher.MemorylimitException):
+ if obj.exit_status in (
+ pynisher.TimeoutException,
+ pynisher.MemorylimitException,
+ ):
# Even if the pynisher thinks that a timeout or memout occured,
# it can be that the target algorithm wrote something into the queue
# - then we treat it as a succesful run
try:
info = autosklearn.evaluation.util.read_queue(queue)
- result = info[-1]['loss']
- status = info[-1]['status']
- additional_run_info = info[-1]['additional_run_info']
+ result = info[-1]["loss"]
+ status = info[-1]["status"]
+ additional_run_info = info[-1]["additional_run_info"]
if obj.stdout:
- additional_run_info['subprocess_stdout'] = obj.stdout
+ additional_run_info["subprocess_stdout"] = obj.stdout
if obj.stderr:
- additional_run_info['subprocess_stderr'] = obj.stderr
+ additional_run_info["subprocess_stderr"] = obj.stderr
if obj.exit_status is pynisher.TimeoutException:
- additional_run_info['info'] = 'Run stopped because of timeout.'
+ additional_run_info["info"] = "Run stopped because of timeout."
elif obj.exit_status is pynisher.MemorylimitException:
- additional_run_info['info'] = 'Run stopped because of memout.'
+ additional_run_info["info"] = "Run stopped because of memout."
if status in [StatusType.SUCCESS, StatusType.DONOTADVANCE]:
cost = result
@@ -385,11 +440,13 @@ def run(
info = None
if obj.exit_status is pynisher.TimeoutException:
status = StatusType.TIMEOUT
- additional_run_info = {'error': 'Timeout'}
+ additional_run_info = {"error": "Timeout"}
elif obj.exit_status is pynisher.MemorylimitException:
status = StatusType.MEMOUT
additional_run_info = {
- "error": "Memout (used more than {} MB).".format(self.memory_limit)
+ "error": "Memout (used more than {} MB).".format(
+ self.memory_limit
+ )
}
else:
raise ValueError(obj.exit_status)
@@ -399,102 +456,129 @@ def run(
info = None
status = StatusType.ABORT
cost = self.worst_possible_result
- additional_run_info = {'error': 'Your configuration of '
- 'auto-sklearn does not work!',
- 'exit_status': _encode_exit_status(obj.exit_status),
- 'subprocess_stdout': obj.stdout,
- 'subprocess_stderr': obj.stderr,
- }
+ additional_run_info = {
+ "error": "Your configuration of " "auto-sklearn does not work!",
+ "exit_status": _encode_exit_status(obj.exit_status),
+ "subprocess_stdout": obj.stdout,
+ "subprocess_stderr": obj.stderr,
+ }
else:
try:
info = autosklearn.evaluation.util.read_queue(queue)
- result = info[-1]['loss']
- status = info[-1]['status']
- additional_run_info = info[-1]['additional_run_info']
+ result = info[-1]["loss"]
+ status = info[-1]["status"]
+ additional_run_info = info[-1]["additional_run_info"]
if obj.exit_status == 0:
cost = result
else:
status = StatusType.CRASHED
cost = self.worst_possible_result
- additional_run_info['info'] = 'Run treated as crashed ' \
- 'because the pynisher exit ' \
- 'status %s is unknown.' % \
- str(obj.exit_status)
- additional_run_info['exit_status'] = _encode_exit_status(obj.exit_status)
- additional_run_info['subprocess_stdout'] = obj.stdout
- additional_run_info['subprocess_stderr'] = obj.stderr
+ additional_run_info["info"] = (
+ "Run treated as crashed "
+ "because the pynisher exit "
+ "status %s is unknown." % str(obj.exit_status)
+ )
+ additional_run_info["exit_status"] = _encode_exit_status(
+ obj.exit_status
+ )
+ additional_run_info["subprocess_stdout"] = obj.stdout
+ additional_run_info["subprocess_stderr"] = obj.stderr
except Empty:
info = None
additional_run_info = {
- 'error': 'Result queue is empty',
- 'exit_status': _encode_exit_status(obj.exit_status),
- 'subprocess_stdout': obj.stdout,
- 'subprocess_stderr': obj.stderr,
- 'exitcode': obj.exitcode
+ "error": "Result queue is empty",
+ "exit_status": _encode_exit_status(obj.exit_status),
+ "subprocess_stdout": obj.stdout,
+ "subprocess_stderr": obj.stderr,
+ "exitcode": obj.exitcode,
}
status = StatusType.CRASHED
cost = self.worst_possible_result
if (
- (self.budget_type is None or budget == 0)
- and status == StatusType.DONOTADVANCE
- ):
+ self.budget_type is None or budget == 0
+ ) and status == StatusType.DONOTADVANCE:
status = StatusType.SUCCESS
if not isinstance(additional_run_info, dict):
- additional_run_info = {'message': additional_run_info}
+ additional_run_info = {"message": additional_run_info}
if (
info is not None
- and self.resampling_strategy in ('holdout-iterative-fit', 'cv-iterative-fit')
+ and self.resampling_strategy
+ in ("holdout-iterative-fit", "cv-iterative-fit")
and status != StatusType.CRASHED
):
learning_curve = autosklearn.evaluation.util.extract_learning_curve(info)
learning_curve_runtime = autosklearn.evaluation.util.extract_learning_curve(
- info, 'duration'
+ info, "duration"
)
if len(learning_curve) > 1:
- additional_run_info['learning_curve'] = learning_curve
- additional_run_info['learning_curve_runtime'] = learning_curve_runtime
+ additional_run_info["learning_curve"] = learning_curve
+ additional_run_info["learning_curve_runtime"] = learning_curve_runtime
train_learning_curve = autosklearn.evaluation.util.extract_learning_curve(
- info, 'train_loss'
+ info, "train_loss"
)
if len(train_learning_curve) > 1:
- additional_run_info['train_learning_curve'] = train_learning_curve
- additional_run_info['learning_curve_runtime'] = learning_curve_runtime
-
- if self._get_validation_loss:
- validation_learning_curve = autosklearn.evaluation.util.extract_learning_curve(
- info, 'validation_loss',
- )
- if len(validation_learning_curve) > 1:
- additional_run_info['validation_learning_curve'] = \
- validation_learning_curve
- additional_run_info[
- 'learning_curve_runtime'] = learning_curve_runtime
+ additional_run_info["train_learning_curve"] = train_learning_curve
+ additional_run_info["learning_curve_runtime"] = learning_curve_runtime
if self._get_test_loss:
- test_learning_curve = autosklearn.evaluation.util.extract_learning_curve(
- info, 'test_loss',
+ test_learning_curve = (
+ autosklearn.evaluation.util.extract_learning_curve(
+ info,
+ "test_loss",
+ )
)
if len(test_learning_curve) > 1:
- additional_run_info['test_learning_curve'] = test_learning_curve
+ additional_run_info["test_learning_curve"] = test_learning_curve
additional_run_info[
- 'learning_curve_runtime'] = learning_curve_runtime
+ "learning_curve_runtime"
+ ] = learning_curve_runtime
if isinstance(config, int):
- origin = 'DUMMY'
+ origin = "DUMMY"
config_id = config
else:
- origin = getattr(config, 'origin', 'UNKNOWN')
+ origin = getattr(config, "origin", "UNKNOWN")
config_id = config.config_id
- additional_run_info['configuration_origin'] = origin
+
+ additional_run_info["configuration_origin"] = origin
runtime = float(obj.wall_clock_time)
autosklearn.evaluation.util.empty_queue(queue)
self.logger.info("Finished evaluating configuration %d" % config_id)
+
+ # Do some sanity checking (for multi objective)
+ if len(self.multi_objectives) > 1:
+ error = (
+ f"Returned costs {cost} does not match the number of objectives"
+ f" {len(self.multi_objectives)}."
+ )
+
+ # If dict convert to array
+ # Make sure the ordering is correct
+ if isinstance(cost, dict):
+ ordered_cost = []
+ for name in self.multi_objectives:
+ if name not in cost:
+ raise RuntimeError(
+ f"Objective {name} was not found "
+ f"in the returned costs ({cost})"
+ )
+
+ ordered_cost.append(cost[name])
+ cost = ordered_cost
+
+ if isinstance(cost, list):
+ if len(cost) != len(self.multi_objectives):
+ raise RuntimeError(error)
+
+ if isinstance(cost, float):
+ raise RuntimeError(error)
+
return status, cost, runtime, additional_run_info
diff --git a/autosklearn/evaluation/abstract_evaluator.py b/autosklearn/evaluation/abstract_evaluator.py
index 2e398b00ae..b97f588a45 100644
--- a/autosklearn/evaluation/abstract_evaluator.py
+++ b/autosklearn/evaluation/abstract_evaluator.py
@@ -1,44 +1,41 @@
+from __future__ import annotations
+
+from typing import Any, Dict, List, Optional, Sequence, TextIO, Tuple, Type, Union, cast
+
import logging
import multiprocessing
import time
import warnings
-from typing import Any, Dict, List, Optional, TextIO, Tuple, Type, Union, cast
import numpy as np
-
+from ConfigSpace import Configuration
from sklearn.base import BaseEstimator
from sklearn.dummy import DummyClassifier, DummyRegressor
from sklearn.ensemble import VotingClassifier, VotingRegressor
-
from smac.tae import StatusType
-
from threadpoolctl import threadpool_limits
import autosklearn.pipeline.classification
import autosklearn.pipeline.regression
-from autosklearn.pipeline.components.base import ThirdPartyComponents, _addons
+from autosklearn.automl_common.common.utils.backend import Backend
from autosklearn.constants import (
CLASSIFICATION_TASKS,
- REGRESSION_TASKS,
- MULTILABEL_CLASSIFICATION,
MULTICLASS_CLASSIFICATION,
- MULTIOUTPUT_REGRESSION
+ MULTILABEL_CLASSIFICATION,
+ MULTIOUTPUT_REGRESSION,
+ REGRESSION_TASKS,
)
+from autosklearn.data.target_validator import (
+ SUPPORTED_TARGET_TYPES,
+ SUPPORTED_XDATA_TYPES,
+)
+from autosklearn.metrics import Scorer, calculate_losses
+from autosklearn.pipeline.components.base import ThirdPartyComponents, _addons
from autosklearn.pipeline.implementations.util import (
- convert_multioutput_multiclass_to_multilabel
+ convert_multioutput_multiclass_to_multilabel,
)
-from autosklearn.metrics import calculate_loss, Scorer
-from autosklearn.util.backend import Backend
from autosklearn.util.logging_ import PicklableClientLogger, get_named_client_logger
-from ConfigSpace import Configuration
-
-
-__all__ = [
- 'AbstractEvaluator'
-]
-
-
# General TYPE definitions for numpy
TYPE_ADDITIONAL_INFO = Dict[str, Union[int, float, str, Dict, List, Tuple]]
@@ -55,9 +52,10 @@ def __init__(
):
self.config = config
if config == 1:
- super(MyDummyClassifier, self).__init__(strategy="uniform")
+ super().__init__(strategy="uniform")
else:
- super(MyDummyClassifier, self).__init__(strategy="most_frequent")
+ super().__init__(strategy="most_frequent")
+
self.random_state = random_state
self.init_params = init_params
self.dataset_properties = dataset_properties
@@ -68,34 +66,40 @@ def pre_transform(
self,
X: np.ndarray,
y: np.ndarray,
- fit_params: Optional[Dict[str, Any]] = None
- ) -> Tuple[np.ndarray, Dict[str, Any]]: # pylint: disable=R0201
+ fit_params: Optional[Dict[str, Any]] = None,
+ ) -> Tuple[np.ndarray, Dict[str, Any]]:
if fit_params is None:
fit_params = {}
return X, fit_params
- def fit(self, X: np.ndarray, y: np.ndarray,
- sample_weight: Optional[Union[np.ndarray, List]] = None
- ) -> DummyClassifier:
- return super(MyDummyClassifier, self).fit(np.ones((X.shape[0], 1)), y,
- sample_weight=sample_weight)
+ def fit(
+ self,
+ X: np.ndarray,
+ y: np.ndarray,
+ sample_weight: Optional[Union[np.ndarray, List]] = None,
+ ) -> DummyClassifier:
+ return super(MyDummyClassifier, self).fit(
+ np.ones((X.shape[0], 1)), y, sample_weight=sample_weight
+ )
- def fit_estimator(self, X: np.ndarray, y: np.ndarray,
- fit_params: Optional[Dict[str, Any]] = None) -> DummyClassifier:
+ def fit_estimator(
+ self,
+ X: np.ndarray,
+ y: np.ndarray,
+ fit_params: Optional[Dict[str, Any]] = None,
+ ) -> DummyClassifier:
return self.fit(X, y)
- def predict_proba(self, X: np.ndarray, batch_size: int = 1000
- ) -> np.ndarray:
+ def predict_proba(self, X: np.ndarray, batch_size: int = 1000) -> np.ndarray:
new_X = np.ones((X.shape[0], 1))
- probas = super(MyDummyClassifier, self).predict_proba(new_X)
- probas = convert_multioutput_multiclass_to_multilabel(probas).astype(
- np.float32)
+ probas = super().predict_proba(new_X)
+ probas = convert_multioutput_multiclass_to_multilabel(probas).astype(np.float32)
return probas
- def estimator_supports_iterative_fit(self) -> bool: # pylint: disable=R0201
+ def estimator_supports_iterative_fit(self) -> bool:
return False
- def get_additional_run_info(self) -> Optional[TYPE_ADDITIONAL_INFO]: # pylint: disable=R0201
+ def get_additional_run_info(self) -> Optional[TYPE_ADDITIONAL_INFO]:
return None
@@ -111,9 +115,9 @@ def __init__(
):
self.config = config
if config == 1:
- super(MyDummyRegressor, self).__init__(strategy='mean')
+ super().__init__(strategy="mean")
else:
- super(MyDummyRegressor, self).__init__(strategy='median')
+ super().__init__(strategy="median")
self.random_state = random_state
self.init_params = init_params
self.dataset_properties = dataset_properties
@@ -124,30 +128,36 @@ def pre_transform(
self,
X: np.ndarray,
y: np.ndarray,
- fit_params: Optional[Dict[str, Any]] = None
- ) -> Tuple[np.ndarray, Dict[str, Any]]: # pylint: disable=R0201
+ fit_params: Optional[Dict[str, Any]] = None,
+ ) -> Tuple[np.ndarray, Dict[str, Any]]:
if fit_params is None:
fit_params = {}
return X, fit_params
- def fit(self, X: np.ndarray, y: np.ndarray,
- sample_weight: Optional[Union[np.ndarray, List]] = None
- ) -> DummyRegressor:
- return super(MyDummyRegressor, self).fit(np.ones((X.shape[0], 1)), y,
- sample_weight=sample_weight)
+ def fit(
+ self,
+ X: np.ndarray,
+ y: np.ndarray,
+ sample_weight: Optional[Union[np.ndarray, List]] = None,
+ ) -> DummyRegressor:
+ return super().fit(np.ones((X.shape[0], 1)), y, sample_weight=sample_weight)
- def fit_estimator(self, X: np.ndarray, y: np.ndarray,
- fit_params: Optional[Dict[str, Any]] = None) -> DummyRegressor:
+ def fit_estimator(
+ self,
+ X: np.ndarray,
+ y: np.ndarray,
+ fit_params: Optional[Dict[str, Any]] = None,
+ ) -> DummyRegressor:
return self.fit(X, y)
def predict(self, X: np.ndarray, batch_size: int = 1000) -> np.ndarray:
new_X = np.ones((X.shape[0], 1))
- return super(MyDummyRegressor, self).predict(new_X).astype(np.float32)
+ return super().predict(new_X).astype(np.float32)
- def estimator_supports_iterative_fit(self) -> bool: # pylint: disable=R0201
+ def estimator_supports_iterative_fit(self) -> bool:
return False
- def get_additional_run_info(self) -> Optional[TYPE_ADDITIONAL_INFO]: # pylint: disable=R0201
+ def get_additional_run_info(self) -> Optional[TYPE_ADDITIONAL_INFO]:
return None
@@ -155,7 +165,7 @@ def _fit_and_suppress_warnings(
logger: Union[logging.Logger, PicklableClientLogger],
model: BaseEstimator,
X: np.ndarray,
- y: np.ndarray
+ y: np.ndarray,
) -> BaseEstimator:
def send_warnings_to_log(
message: Union[Warning, str],
@@ -165,8 +175,7 @@ def send_warnings_to_log(
file: Optional[TextIO] = None,
line: Optional[str] = None,
) -> None:
- logger.debug('%s:%s: %s:%s' %
- (filename, lineno, str(category), message))
+ logger.debug("%s:%s: %s:%s" % (filename, lineno, str(category), message))
return
with warnings.catch_warnings():
@@ -181,7 +190,7 @@ def __init__(
self,
backend: Backend,
queue: multiprocessing.Queue,
- metric: Scorer,
+ metrics: Sequence[Scorer],
additional_components: Dict[str, ThirdPartyComponents],
port: Optional[int],
configuration: Optional[Union[int, Configuration]] = None,
@@ -211,40 +220,39 @@ def __init__(
self.include = include
self.exclude = exclude
- self.X_valid = self.datamanager.data.get('X_valid')
- self.y_valid = self.datamanager.data.get('Y_valid')
- self.X_test = self.datamanager.data.get('X_test')
- self.y_test = self.datamanager.data.get('Y_test')
+ self.X_test = self.datamanager.data.get("X_test")
+ self.y_test = self.datamanager.data.get("Y_test")
- self.metric = metric
- self.task_type = self.datamanager.info['task']
+ self.metrics = metrics
+ self.task_type = self.datamanager.info["task"]
self.seed = seed
self.output_y_hat_optimization = output_y_hat_optimization
- self.scoring_functions = scoring_functions
+ self.scoring_functions = scoring_functions if scoring_functions else []
if isinstance(disable_file_output, (bool, list)):
self.disable_file_output: Union[bool, List[str]] = disable_file_output
else:
- raise ValueError('disable_file_output should be either a bool or a list')
+ raise ValueError("disable_file_output should be either a bool or a list")
if self.task_type in REGRESSION_TASKS:
if not isinstance(self.configuration, Configuration):
self.model_class = MyDummyRegressor
else:
- self.model_class = \
+ self.model_class = (
autosklearn.pipeline.regression.SimpleRegressionPipeline
+ )
self.predict_function = self._predict_regression
else:
if not isinstance(self.configuration, Configuration):
self.model_class = MyDummyClassifier
else:
- self.model_class = autosklearn.pipeline.classification.SimpleClassificationPipeline
+ self.model_class = (
+ autosklearn.pipeline.classification.SimpleClassificationPipeline
+ )
self.predict_function = self._predict_proba
- self._init_params = {
- 'data_preprocessor:feat_type': self.datamanager.feat_type
- }
+ self._init_params = {"data_preprocessor:feat_type": self.datamanager.feat_type}
if init_params is not None:
self._init_params.update(init_params)
@@ -253,8 +261,11 @@ def __init__(
num_run = 0
self.num_run = num_run
- logger_name = '%s(%d):%s' % (self.__class__.__name__.split('.')[-1],
- self.seed, self.datamanager.name)
+ logger_name = "%s(%d):%s" % (
+ self.__class__.__name__.split(".")[-1],
+ self.seed,
+ self.datamanager.name,
+ )
if self.port is None:
self.logger = logging.getLogger(__name__)
@@ -264,18 +275,21 @@ def __init__(
port=self.port,
)
- self.Y_optimization: Optional[Union[List, np.ndarray]] = None
+ self.X_optimization: Optional[SUPPORTED_XDATA_TYPES] = None
+ self.Y_optimization: Optional[SUPPORTED_TARGET_TYPES] = None
self.Y_actual_train = None
self.budget = budget
self.budget_type = budget_type
- # Add 3rd-party components to the list of 3rd-party components in case this wasn't done
- # before (this happens if we run in parallel and the components are only passed to the
- # AbstractEvaluator via the TAE and are not there yet because the worker is in its own
- # process).
+ # Add 3rd-party components to the list of 3rd-party components in case this
+ # wasn't done before (this happens if we run in parallel and the components
+ # are only passed to the AbstractEvaluator via the TAE and are not there
+ # yet because the worker is in its own process).
for key in additional_components:
- for component_name, component in additional_components[key].components.items():
+ for component_name, component in additional_components[
+ key
+ ].components.items():
if component_name not in _addons[key].components:
_addons[key].add_component(component)
@@ -284,34 +298,41 @@ def __init__(
def _get_model(self) -> BaseEstimator:
if not isinstance(self.configuration, Configuration):
- model = self.model_class(config=self.configuration,
- random_state=self.seed,
- init_params=self._init_params)
+ model = self.model_class(
+ config=self.configuration,
+ random_state=self.seed,
+ init_params=self._init_params,
+ )
else:
if self.task_type in REGRESSION_TASKS:
dataset_properties = {
- 'task': self.task_type,
- 'sparse': self.datamanager.info['is_sparse'] == 1,
- 'multioutput': self.task_type == MULTIOUTPUT_REGRESSION,
+ "task": self.task_type,
+ "sparse": self.datamanager.info["is_sparse"] == 1,
+ "multioutput": self.task_type == MULTIOUTPUT_REGRESSION,
}
else:
dataset_properties = {
- 'task': self.task_type,
- 'sparse': self.datamanager.info['is_sparse'] == 1,
- 'multilabel': self.task_type == MULTILABEL_CLASSIFICATION,
- 'multiclass': self.task_type == MULTICLASS_CLASSIFICATION,
+ "task": self.task_type,
+ "sparse": self.datamanager.info["is_sparse"] == 1,
+ "multilabel": self.task_type == MULTILABEL_CLASSIFICATION,
+ "multiclass": self.task_type == MULTICLASS_CLASSIFICATION,
}
- model = self.model_class(config=self.configuration,
- dataset_properties=dataset_properties,
- random_state=self.seed,
- include=self.include,
- exclude=self.exclude,
- init_params=self._init_params)
+ model = self.model_class(
+ config=self.configuration,
+ dataset_properties=dataset_properties,
+ random_state=self.seed,
+ include=self.include,
+ exclude=self.exclude,
+ init_params=self._init_params,
+ )
return model
- def _loss(self, y_true: np.ndarray, y_hat: np.ndarray,
- scoring_functions: Optional[List[Scorer]] = None
- ) -> Union[float, Dict[str, float]]:
+ def _loss(
+ self,
+ y_true: np.ndarray,
+ y_hat: np.ndarray,
+ X_data: Optional[SUPPORTED_XDATA_TYPES] = None,
+ ) -> Dict[str, float]:
"""Auto-sklearn follows a minimization goal.
The calculate_loss internally translate a score function to
a minimization problem.
@@ -322,123 +343,109 @@ def _loss(self, y_true: np.ndarray, y_hat: np.ndarray,
----------
y_true
"""
- scoring_functions = (
- self.scoring_functions
- if scoring_functions is None
- else scoring_functions
+ return calculate_losses(
+ y_true,
+ y_hat,
+ self.task_type,
+ self.metrics,
+ X_data=X_data,
+ scoring_functions=self.scoring_functions,
)
- if not isinstance(self.configuration, Configuration):
- if scoring_functions:
- return {self.metric.name: self.metric._worst_possible_result}
- else:
- return self.metric._worst_possible_result
-
- return calculate_loss(
- y_true, y_hat, self.task_type, self.metric,
- scoring_functions=scoring_functions)
def finish_up(
self,
loss: Union[Dict[str, float], float],
- train_loss: Optional[Union[float, Dict[str, float]]],
+ train_loss: Optional[Dict[str, float]],
opt_pred: np.ndarray,
- valid_pred: np.ndarray,
test_pred: np.ndarray,
additional_run_info: Optional[TYPE_ADDITIONAL_INFO],
file_output: bool,
final_call: bool,
status: StatusType,
- ) -> Tuple[float, Union[float, Dict[str, float]], int,
- Dict[str, Union[str, int, float, Dict, List, Tuple]]]:
- """This function does everything necessary after the fitting is done:
+ ) -> Tuple[
+ float,
+ Union[float, Dict[str, float]],
+ int,
+ Dict[str, Union[str, int, float, Dict, List, Tuple]],
+ ]:
+ """Do everything necessary after the fitting is done:
* predicting
* saving the files for the ensembles_statistics
* generate output for SMAC
We use it as the signal handler so we can recycle the code for the
- normal usecase and when the runsolver kills us here :)"""
-
+ normal usecase and when the runsolver kills us here :)
+ """
self.duration = time.time() - self.starttime
if file_output:
- file_out_loss, additional_run_info_ = self.file_output(
- opt_pred, valid_pred, test_pred,
- )
+ file_out_loss, additional_run_info_ = self.file_output(opt_pred, test_pred)
else:
file_out_loss = None
additional_run_info_ = {}
- validation_loss, test_loss = self.calculate_auxiliary_losses(
- valid_pred, test_pred,
- )
+ test_loss = self.calculate_auxiliary_losses(test_pred)
if file_out_loss is not None:
return self.duration, file_out_loss, self.seed, additional_run_info_
- if isinstance(loss, dict):
- loss_ = loss
- loss = loss_[self.metric.name]
+ loss_ = loss
+ for metric in self.metrics:
+ if metric.name not in loss_:
+ raise ValueError(
+ f"Unable to compute optimization metric {metric.name}. Are you "
+ f"sure {metric.name} is applicable for the given task type?"
+ )
+ if len(self.metrics) == 1:
+ loss = loss_[self.metrics[0].name]
else:
- loss_ = {}
-
- additional_run_info = (
- {} if additional_run_info is None else additional_run_info
- )
- for metric_name, value in loss_.items():
- additional_run_info[metric_name] = value
- additional_run_info['duration'] = self.duration
- additional_run_info['num_run'] = self.num_run
+ loss = {metric.name: loss_[metric.name] for metric in self.metrics}
+
+ additional_run_info = {} if additional_run_info is None else additional_run_info
+ for metric in self.scoring_functions:
+ if metric.name in loss_:
+ additional_run_info[metric.name] = loss_[metric.name]
+ additional_run_info["duration"] = self.duration
+ additional_run_info["num_run"] = self.num_run
if train_loss is not None:
- additional_run_info['train_loss'] = train_loss
- if validation_loss is not None:
- additional_run_info['validation_loss'] = validation_loss
+ if len(self.metrics) == 1:
+ additional_run_info["train_loss"] = train_loss[self.metrics[0].name]
+ else:
+ additional_run_info["train_loss"] = [
+ train_loss[metric.name] for metric in self.metrics
+ ]
if test_loss is not None:
- additional_run_info['test_loss'] = test_loss
+ additional_run_info["test_loss"] = test_loss
- rval_dict = {'loss': loss,
- 'additional_run_info': additional_run_info,
- 'status': status}
+ return_value_dict = {
+ "loss": loss,
+ "additional_run_info": additional_run_info,
+ "status": status,
+ }
if final_call:
- rval_dict['final_queue_element'] = True
+ return_value_dict["final_queue_element"] = True
- self.queue.put(rval_dict)
+ self.queue.put(return_value_dict)
return self.duration, loss_, self.seed, additional_run_info_
def calculate_auxiliary_losses(
self,
- Y_valid_pred: np.ndarray,
- Y_test_pred: np.ndarray,
- ) -> Tuple[Optional[float], Optional[float]]:
- if Y_valid_pred is not None:
- if self.y_valid is not None:
- validation_loss: Optional[Union[float, Dict[str, float]]] = self._loss(
- self.y_valid, Y_valid_pred)
- if isinstance(validation_loss, dict):
- validation_loss = validation_loss[self.metric.name]
- else:
- validation_loss = None
- else:
- validation_loss = None
-
- if Y_test_pred is not None:
- if self.y_test is not None:
- test_loss: Optional[Union[float, Dict[str, float]]] = self._loss(
- self.y_test, Y_test_pred)
- if isinstance(test_loss, dict):
- test_loss = test_loss[self.metric.name]
- else:
- test_loss = None
- else:
- test_loss = None
+ Y_test_pred: np.ndarray | None,
+ ) -> float | dict[str, float] | None:
+ if Y_test_pred is None or self.y_test is None:
+ return None
+
+ test_loss = self._loss(self.y_test, Y_test_pred)
+ if len(self.metrics) == 1:
+ test_loss = test_loss[self.metrics[0].name]
- return validation_loss, test_loss
+ return test_loss
def file_output(
self,
Y_optimization_pred: np.ndarray,
- Y_valid_pred: np.ndarray,
Y_test_pred: np.ndarray,
- ) -> Tuple[Optional[float], Dict[str, Union[str, int, float, List, Dict, Tuple]]]:
+ ) -> tuple[float | None, dict[str, Any]]:
# Abort if self.Y_optimization is None
# self.Y_optimization can be None if we use partial-cv, then,
# obviously no output should be saved.
@@ -450,27 +457,19 @@ def file_output(
return (
1.0,
{
- 'error':
- "Targets %s and prediction %s don't have "
- "the same length. Probably training didn't "
- "finish" % (np.shape(self.Y_optimization), Y_optimization_pred.shape)
- },
+ "error": "Targets %s and prediction %s don't have "
+ "the same length. Probably training didn't "
+ "finish"
+ % (np.shape(self.Y_optimization), Y_optimization_pred.shape)
+ },
)
# Abort if predictions contain NaNs
- for y, s in [
- # Y_train_pred deleted here. Fix unittest accordingly.
- [Y_optimization_pred, 'optimization'],
- [Y_valid_pred, 'validation'],
- [Y_test_pred, 'test']
- ]:
+ for y, s in [(Y_optimization_pred, "optimization"), (Y_test_pred, "test")]:
if y is not None and not np.all(np.isfinite(y)):
return (
1.0,
- {
- 'error':
- 'Model predictions for %s set contains NaNs.' % s
- },
+ {"error": "Model predictions for %s set contains NaNs." % s},
)
# Abort if we don't want to output anything.
@@ -488,17 +487,25 @@ def file_output(
self.disable_file_output = cast(List, self.disable_file_output)
# This file can be written independently of the others down bellow
- if ('y_optimization' not in self.disable_file_output):
+ if "y_optimization" not in self.disable_file_output:
if self.output_y_hat_optimization:
- self.backend.save_targets_ensemble(self.Y_optimization)
+ self.backend.save_additional_data(
+ self.Y_optimization, what="targets_ensemble"
+ )
+ self.backend.save_additional_data(
+ self.X_optimization, what="input_ensemble"
+ )
models: Optional[BaseEstimator] = None
- if hasattr(self, 'models'):
- if len(self.models) > 0 and self.models[0] is not None: # type: ignore[attr-defined]
- if ('models' not in self.disable_file_output):
+ if hasattr(self, "models"):
+ if len(self.models) > 0 and self.models[0] is not None:
+ if "models" not in self.disable_file_output:
if self.task_type in CLASSIFICATION_TASKS:
- models = VotingClassifier(estimators=None, voting='soft', )
+ models = VotingClassifier(
+ estimators=None,
+ voting="soft",
+ )
else:
models = VotingRegressor(estimators=None)
# Mypy cannot understand hasattr yet
@@ -508,24 +515,29 @@ def file_output(
seed=self.seed,
idx=self.num_run,
budget=self.budget,
- model=self.model if 'model' not in self.disable_file_output else None,
- cv_model=models if 'cv_model' not in self.disable_file_output else None,
+ model=self.model if "model" not in self.disable_file_output else None,
+ cv_model=models if "cv_model" not in self.disable_file_output else None,
+ # TODO: below line needs to be deleted once backend is updated
+ valid_predictions=None,
ensemble_predictions=(
- Y_optimization_pred if 'y_optimization' not in self.disable_file_output else None
- ),
- valid_predictions=(
- Y_valid_pred if 'y_valid' not in self.disable_file_output else None
+ Y_optimization_pred
+ if "y_optimization" not in self.disable_file_output
+ else None
),
test_predictions=(
- Y_test_pred if 'y_test' not in self.disable_file_output else None
+ Y_test_pred if "y_test" not in self.disable_file_output else None
),
)
return None, {}
- def _predict_proba(self, X: np.ndarray, model: BaseEstimator,
- task_type: int, Y_train: Optional[np.ndarray] = None,
- ) -> np.ndarray:
+ def _predict_proba(
+ self,
+ X: np.ndarray,
+ model: BaseEstimator,
+ task_type: int,
+ Y_train: Optional[np.ndarray] = None,
+ ) -> np.ndarray:
def send_warnings_to_log(
message: Union[Warning, str],
category: Type[Warning],
@@ -534,8 +546,9 @@ def send_warnings_to_log(
file: Optional[TextIO] = None,
line: Optional[str] = None,
) -> None:
- self.logger.debug('%s:%s: %s:%s' %
- (filename, lineno, str(category), message))
+ self.logger.debug(
+ "%s:%s: %s:%s" % (filename, lineno, str(category), message)
+ )
return
with warnings.catch_warnings():
@@ -548,8 +561,13 @@ def send_warnings_to_log(
Y_pred = self._ensure_prediction_array_sizes(Y_pred, Y_train)
return Y_pred
- def _predict_regression(self, X: np.ndarray, model: BaseEstimator,
- task_type: int, Y_train: Optional[np.ndarray] = None) -> np.ndarray:
+ def _predict_regression(
+ self,
+ X: np.ndarray,
+ model: BaseEstimator,
+ task_type: int,
+ Y_train: Optional[np.ndarray] = None,
+ ) -> np.ndarray:
def send_warnings_to_log(
message: Union[Warning, str],
category: Type[Warning],
@@ -558,8 +576,9 @@ def send_warnings_to_log(
file: Optional[TextIO] = None,
line: Optional[str] = None,
) -> None:
- self.logger.debug('%s:%s: %s:%s' %
- (filename, lineno, str(category), message))
+ self.logger.debug(
+ "%s:%s: %s:%s" % (filename, lineno, str(category), message)
+ )
return
with warnings.catch_warnings():
@@ -571,14 +590,17 @@ def send_warnings_to_log(
return Y_pred
- def _ensure_prediction_array_sizes(self, prediction: np.ndarray, Y_train: np.ndarray
- ) -> np.ndarray:
- num_classes = self.datamanager.info['label_num']
+ def _ensure_prediction_array_sizes(
+ self, prediction: np.ndarray, Y_train: np.ndarray
+ ) -> np.ndarray:
+ num_classes = self.datamanager.info["label_num"]
- if self.task_type == MULTICLASS_CLASSIFICATION and \
- prediction.shape[1] < num_classes:
+ if (
+ self.task_type == MULTICLASS_CLASSIFICATION
+ and prediction.shape[1] < num_classes
+ ):
if Y_train is None:
- raise ValueError('Y_train must not be None!')
+ raise ValueError("Y_train must not be None!")
classes = list(np.unique(Y_train))
mapping = dict()
@@ -586,8 +608,9 @@ def _ensure_prediction_array_sizes(self, prediction: np.ndarray, Y_train: np.nda
if class_number in classes:
index = classes.index(class_number)
mapping[index] = class_number
- new_predictions = np.zeros((prediction.shape[0], num_classes),
- dtype=np.float32)
+ new_predictions = np.zeros(
+ (prediction.shape[0], num_classes), dtype=np.float32
+ )
for index in mapping:
class_index = mapping[index]
diff --git a/autosklearn/evaluation/splitter.py b/autosklearn/evaluation/splitter.py
index e717a3bcbe..586d92c88f 100644
--- a/autosklearn/evaluation/splitter.py
+++ b/autosklearn/evaluation/splitter.py
@@ -1,49 +1,52 @@
-import numpy as np
+import warnings
-from sklearn.model_selection import StratifiedShuffleSplit, StratifiedKFold
+import numpy as np
+from sklearn.model_selection import StratifiedKFold, StratifiedShuffleSplit
from sklearn.model_selection._split import _validate_shuffle_split
-from sklearn.utils import indexable, check_random_state
-from sklearn.utils import _approximate_mode
-from sklearn.utils.validation import _num_samples, column_or_1d
-from sklearn.utils.validation import check_array
+from sklearn.utils import _approximate_mode, check_random_state, indexable
from sklearn.utils.multiclass import type_of_target
+from sklearn.utils.validation import _num_samples, check_array, column_or_1d
class CustomStratifiedShuffleSplit(StratifiedShuffleSplit):
- """Stratified ShuffleSplit cross-validator that deals with classes with too few samples
- """
+ """Splitter that deals with classes with too few samples"""
def _iter_indices(self, X, y, groups=None): # type: ignore
n_samples = _num_samples(X)
y = check_array(y, ensure_2d=False, dtype=None)
n_train, n_test = _validate_shuffle_split(
- n_samples, self.test_size, self.train_size,
- default_test_size=self._default_test_size)
+ n_samples,
+ self.test_size,
+ self.train_size,
+ default_test_size=self._default_test_size,
+ )
if y.ndim == 2:
# for multi-label y, map each distinct row to a string repr
# using join because str(row) uses an ellipsis if len(row) > 1000
- y = np.array([' '.join(row.astype('str')) for row in y])
+ y = np.array([" ".join(row.astype("str")) for row in y])
classes, y_indices = np.unique(y, return_inverse=True)
n_classes = classes.shape[0]
class_counts = np.bincount(y_indices)
- # print(class_counts)
if n_train < n_classes:
- raise ValueError('The train_size = %d should be greater or '
- 'equal to the number of classes = %d' %
- (n_train, n_classes))
+ raise ValueError(
+ "The train_size = %d should be greater or "
+ "equal to the number of classes = %d" % (n_train, n_classes)
+ )
if n_test < n_classes:
- raise ValueError('The test_size = %d should be greater or '
- 'equal to the number of classes = %d' %
- (n_test, n_classes))
+ raise ValueError(
+ "The test_size = %d should be greater or "
+ "equal to the number of classes = %d" % (n_test, n_classes)
+ )
# Find the sorted list of instances for each class:
# (np.unique above performs a sort, so code is O(n logn) already)
- class_indices = np.split(np.argsort(y_indices, kind='mergesort'),
- np.cumsum(class_counts)[:-1])
+ class_indices = np.split(
+ np.argsort(y_indices, kind="mergesort"), np.cumsum(class_counts)[:-1]
+ )
rng = check_random_state(self.random_state)
@@ -56,18 +59,46 @@ def _iter_indices(self, X, y, groups=None): # type: ignore
train = []
test = []
+ # NOTE: Adapting for unique instances
+ #
+ # Each list n_i, t_i represent the list of class in the
+ # training_set and test_set resepectively.
+ #
+ # n_i = [100, 100, 0, 3] # 100 of class '0', 0 of class '2'
+ # t_i = [300, 300, 1, 3] # 300 of class '0', 1 of class '2'
+ #
+ # To support unique labels such as class '2', which only has one sample
+ # between both n_i and t_i, we need to make sure that n_i has at least
+ # one sample of all classes. There is also the extra check to ensure
+ # that the sizes stay the same.
+ #
+ # n_i = [ 99, 100, 1, 3] # 100 of class '0', 0 of class '2'
+ # | ^
+ # v |
+ # t_i = [301, 300, 0, 3] # 300 of class '0', 1 of class '2'
+ #
+ for i, class_count in enumerate(n_i):
+ if class_count == 0:
+ t_i[i] -= 1
+ n_i[i] += 1
+
+ j = np.argmax(n_i)
+ if n_i[j] == 1:
+ warnings.warn(
+ "Can't respect size requirements for split.",
+ " The training set must contain all of the unique"
+ " labels that exist in the dataset.",
+ )
+ else:
+ n_i[j] -= 1
+ t_i[j] += 1
+
for i in range(n_classes):
- # print("Before", i, class_counts[i], n_i[i], t_i[i])
permutation = rng.permutation(class_counts[i])
- perm_indices_class_i = class_indices[i].take(permutation,
- mode='clip')
- if n_i[i] == 0:
- n_i[i] = 1
- t_i[i] = t_i[i] - 1
+ perm_indices_class_i = class_indices[i].take(permutation, mode="clip")
- # print("After", i, class_counts[i], n_i[i], t_i[i])
- train.extend(perm_indices_class_i[:n_i[i]])
- test.extend(perm_indices_class_i[n_i[i]:n_i[i] + t_i[i]])
+ train.extend(perm_indices_class_i[: n_i[i]])
+ test.extend(perm_indices_class_i[n_i[i] : n_i[i] + t_i[i]])
train = rng.permutation(train)
test = rng.permutation(test)
@@ -84,11 +115,13 @@ def _make_test_folds(self, X, y=None): # type: ignore
rng = check_random_state(self.random_state)
y = np.asarray(y)
type_of_target_y = type_of_target(y)
- allowed_target_types = ('binary', 'multiclass')
+ allowed_target_types = ("binary", "multiclass")
if type_of_target_y not in allowed_target_types:
raise ValueError(
- 'Supported target types are: {}. Got {!r} instead.'.format(
- allowed_target_types, type_of_target_y))
+ "Supported target types are: {}. Got {!r} instead.".format(
+ allowed_target_types, type_of_target_y
+ )
+ )
y = column_or_1d(y)
@@ -106,13 +139,16 @@ def _make_test_folds(self, X, y=None): # type: ignore
# counts, but that code is unreadable.)
y_order = np.sort(y_encoded)
allocation = np.asarray(
- [np.bincount(y_order[i::self.n_splits], minlength=n_classes)
- for i in range(self.n_splits)])
+ [
+ np.bincount(y_order[i :: self.n_splits], minlength=n_classes)
+ for i in range(self.n_splits)
+ ]
+ )
# To maintain the data order dependencies as best as possible within
# the stratification constraint, we assign samples from each class in
# blocks (and then mess that up when shuffle=True).
- test_folds = np.empty(len(y), dtype='i')
+ test_folds = np.empty(len(y), dtype="i")
for k in range(n_classes):
# since the kth column of allocation stores the number of samples
# of class k in each test set, this generates blocks of fold
@@ -129,12 +165,11 @@ def split(self, X, y=None, groups=None): # type: ignore
n_samples = _num_samples(X)
if self.n_splits > n_samples:
raise ValueError(
- ("Cannot have number of splits n_splits={0} greater"
- " than the number of samples: n_samples={1}.")
- .format(self.n_splits, n_samples))
+ f"Cannot have number of splits n_splits={self.n_splits} greater"
+ f" than the number of samples: n_samples={n_samples}."
+ )
for train, test in super().split(X, y, groups):
- # print(len(np.unique(y)), len(np.unique(y[train])), len(np.unique(y[test])))
all_classes = np.unique(y)
train_classes = np.unique(y[train])
train = list(train)
@@ -151,11 +186,5 @@ def split(self, X, y=None, groups=None): # type: ignore
# print(len(train), len(test))
train = np.array(train, dtype=int)
test = np.array(test, dtype=int)
- # print(
- # len(np.unique(y)),
- # len(np.unique(y[train])),
- # len(np.unique(y[test])),
- # len(train), len(test),
- # )
yield train, test
diff --git a/autosklearn/evaluation/test_evaluator.py b/autosklearn/evaluation/test_evaluator.py
index e83edb0682..d624c1a44d 100644
--- a/autosklearn/evaluation/test_evaluator.py
+++ b/autosklearn/evaluation/test_evaluator.py
@@ -1,35 +1,29 @@
# -*- encoding: utf-8 -*-
-import multiprocessing
-from typing import Any, Dict, List, Optional, Tuple, Union
+from typing import Any, Dict, List, Optional, Sequence, Tuple, Union
-from ConfigSpace import Configuration
+import multiprocessing
import numpy as np
-
+from ConfigSpace import Configuration
from smac.tae import StatusType
+from autosklearn.automl_common.common.utils.backend import Backend
from autosklearn.evaluation.abstract_evaluator import (
AbstractEvaluator,
_fit_and_suppress_warnings,
)
+from autosklearn.metrics import Scorer, calculate_losses
from autosklearn.pipeline.components.base import ThirdPartyComponents
-from autosklearn.metrics import calculate_loss, Scorer
-from autosklearn.util.backend import Backend
-
-__all__ = [
- 'eval_t',
- 'TestEvaluator'
-]
+__all__ = ["eval_t", "TestEvaluator"]
class TestEvaluator(AbstractEvaluator):
-
def __init__(
self,
backend: Backend,
queue: multiprocessing.Queue,
- metric: Scorer,
+ metrics: Sequence[Scorer],
additional_components: Dict[str, ThirdPartyComponents],
port: Optional[int],
configuration: Optional[Union[int, Configuration]] = None,
@@ -45,7 +39,7 @@ def __init__(
queue=queue,
port=port,
configuration=configuration,
- metric=metric,
+ metrics=metrics,
additional_components=additional_components,
scoring_functions=scoring_functions,
seed=seed,
@@ -54,15 +48,15 @@ def __init__(
include=include,
exclude=exclude,
disable_file_output=disable_file_output,
- init_params=init_params
+ init_params=init_params,
)
self.configuration = configuration
- self.X_train = self.datamanager.data['X_train']
- self.Y_train = self.datamanager.data['Y_train']
+ self.X_train = self.datamanager.data["X_train"]
+ self.Y_train = self.datamanager.data["Y_train"]
- self.X_test = self.datamanager.data.get('X_test')
- self.Y_test = self.datamanager.data.get('Y_test')
+ self.X_test = self.datamanager.data.get("X_test")
+ self.Y_test = self.datamanager.data.get("Y_test")
self.model = self._get_model()
@@ -73,7 +67,6 @@ def fit_predict_and_loss(self) -> None:
loss=loss,
train_loss=None,
opt_pred=Y_pred,
- valid_pred=None,
test_pred=None,
file_output=False,
final_call=True,
@@ -84,25 +77,28 @@ def fit_predict_and_loss(self) -> None:
def predict_and_loss(
self, train: bool = False
) -> Tuple[Union[Dict[str, float], float], np.array, Any, Any]:
-
if train:
- Y_pred = self.predict_function(self.X_train, self.model,
- self.task_type, self.Y_train)
- err = calculate_loss(
+ Y_pred = self.predict_function(
+ self.X_train, self.model, self.task_type, self.Y_train
+ )
+ err = calculate_losses(
solution=self.Y_train,
prediction=Y_pred,
task_type=self.task_type,
- metric=self.metric,
- scoring_functions=self.scoring_functions)
+ metrics=self.metrics,
+ scoring_functions=self.scoring_functions,
+ )
else:
- Y_pred = self.predict_function(self.X_test, self.model,
- self.task_type, self.Y_train)
- err = calculate_loss(
+ Y_pred = self.predict_function(
+ self.X_test, self.model, self.task_type, self.Y_train
+ )
+ err = calculate_losses(
solution=self.Y_test,
prediction=Y_pred,
task_type=self.task_type,
- metric=self.metric,
- scoring_functions=self.scoring_functions)
+ metrics=self.metrics,
+ scoring_functions=self.scoring_functions,
+ )
return err, Y_pred, None, None
@@ -113,7 +109,7 @@ def eval_t(
queue: multiprocessing.Queue,
config: Union[int, Configuration],
backend: Backend,
- metric: Scorer,
+ metrics: Sequence[Scorer],
seed: int,
num_run: int,
instance: Dict[str, Any],
@@ -128,14 +124,19 @@ def eval_t(
budget: Optional[float] = None,
budget_type: Optional[str] = None,
) -> None:
- evaluator = TestEvaluator(configuration=config,
- backend=backend, metric=metric, seed=seed,
- port=port,
- queue=queue,
- scoring_functions=scoring_functions,
- include=include, exclude=exclude,
- disable_file_output=disable_file_output,
- additional_components=additional_components,
- init_params=init_params,)
+ evaluator = TestEvaluator(
+ configuration=config,
+ backend=backend,
+ metrics=metrics,
+ seed=seed,
+ port=port,
+ queue=queue,
+ scoring_functions=scoring_functions,
+ include=include,
+ exclude=exclude,
+ disable_file_output=disable_file_output,
+ additional_components=additional_components,
+ init_params=init_params,
+ )
evaluator.fit_predict_and_loss()
diff --git a/autosklearn/evaluation/train_evaluator.py b/autosklearn/evaluation/train_evaluator.py
index 51b433153d..f19db473bf 100644
--- a/autosklearn/evaluation/train_evaluator.py
+++ b/autosklearn/evaluation/train_evaluator.py
@@ -1,53 +1,67 @@
-import logging
-import multiprocessing
-import warnings
-from typing import Any, Dict, List, Optional, Tuple, Union, cast
+from typing import Any, Dict, List, Optional, Sequence, Tuple, Union, cast
import copy
import json
-
-from ConfigSpace import Configuration
+import logging
+import multiprocessing
+import warnings
import numpy as np
-
-from smac.tae import TAEAbortException, StatusType
-
+import pandas
+import pandas as pd
+import scipy.sparse
+from ConfigSpace import Configuration
from sklearn.base import BaseEstimator
-from sklearn.model_selection import ShuffleSplit, StratifiedShuffleSplit, KFold, \
- StratifiedKFold, train_test_split, BaseCrossValidator, PredefinedSplit
-from sklearn.model_selection._split import _RepeatedSplits, BaseShuffleSplit
-
-from autosklearn.evaluation.abstract_evaluator import (
- AbstractEvaluator,
- TYPE_ADDITIONAL_INFO,
- _fit_and_suppress_warnings,
+from sklearn.model_selection import (
+ BaseCrossValidator,
+ KFold,
+ PredefinedSplit,
+ ShuffleSplit,
+ StratifiedKFold,
+ StratifiedShuffleSplit,
+ train_test_split,
)
-from autosklearn.evaluation.splitter import CustomStratifiedShuffleSplit, CustomStratifiedKFold
-from autosklearn.data.abstract_data_manager import AbstractDataManager
+from sklearn.model_selection._split import BaseShuffleSplit, _RepeatedSplits
+from smac.tae import StatusType, TAEAbortException
+
+from autosklearn.automl_common.common.utils.backend import Backend
from autosklearn.constants import (
CLASSIFICATION_TASKS,
MULTILABEL_CLASSIFICATION,
+ MULTIOUTPUT_REGRESSION,
REGRESSION_TASKS,
- MULTIOUTPUT_REGRESSION
)
-from autosklearn.data.validation import (
- SUPPORTED_FEAT_TYPES,
- SUPPORTED_TARGET_TYPES,
- )
-from autosklearn.pipeline.base import PIPELINE_DATA_DTYPE
-from autosklearn.pipeline.components.base import IterativeComponent, ThirdPartyComponents
+from autosklearn.data.abstract_data_manager import AbstractDataManager
+from autosklearn.data.validation import SUPPORTED_FEAT_TYPES, SUPPORTED_TARGET_TYPES
+from autosklearn.evaluation.abstract_evaluator import (
+ TYPE_ADDITIONAL_INFO,
+ AbstractEvaluator,
+ _fit_and_suppress_warnings,
+)
+from autosklearn.evaluation.splitter import (
+ CustomStratifiedKFold,
+ CustomStratifiedShuffleSplit,
+)
from autosklearn.metrics import Scorer
-from autosklearn.util.backend import Backend
+from autosklearn.pipeline.base import PIPELINE_DATA_DTYPE
+from autosklearn.pipeline.components.base import (
+ IterativeComponent,
+ ThirdPartyComponents,
+)
from autosklearn.util.logging_ import PicklableClientLogger
-
-__all__ = ['TrainEvaluator', 'eval_holdout', 'eval_iterative_holdout',
- 'eval_cv', 'eval_partial_cv', 'eval_partial_cv_iterative']
+__all__ = [
+ "TrainEvaluator",
+ "eval_holdout",
+ "eval_iterative_holdout",
+ "eval_cv",
+ "eval_partial_cv",
+ "eval_partial_cv_iterative",
+]
def _get_y_array(y: SUPPORTED_TARGET_TYPES, task_type: int) -> SUPPORTED_TARGET_TYPES:
- if task_type in CLASSIFICATION_TASKS and task_type != \
- MULTILABEL_CLASSIFICATION:
+ if task_type in CLASSIFICATION_TASKS and task_type != MULTILABEL_CLASSIFICATION:
return y.ravel()
else:
return y
@@ -57,29 +71,26 @@ def subsample_indices(
train_indices: List[int],
subsample: Optional[float],
task_type: int,
- Y_train: SUPPORTED_TARGET_TYPES
+ Y_train: SUPPORTED_TARGET_TYPES,
) -> List[int]:
if not isinstance(subsample, float):
raise ValueError(
- 'Subsample must be of type float, but is of type %s'
- % type(subsample)
+ "Subsample must be of type float, but is of type %s" % type(subsample)
)
elif subsample > 1:
- raise ValueError(
- 'Subsample must not be larger than 1, but is %f'
- % subsample
- )
+ raise ValueError("Subsample must not be larger than 1, but is %f" % subsample)
if subsample is not None and subsample < 1:
# Only subsample if there are more indices given to this method than
# required to subsample because otherwise scikit-learn will complain
if task_type in CLASSIFICATION_TASKS and task_type != MULTILABEL_CLASSIFICATION:
- stratify: Optional[
- SUPPORTED_TARGET_TYPES
- ] = Y_train.iloc[train_indices] if hasattr(
- Y_train, 'iloc') else Y_train[train_indices]
+ stratify: Optional[SUPPORTED_TARGET_TYPES] = (
+ Y_train.iloc[train_indices]
+ if hasattr(Y_train, "iloc")
+ else Y_train[train_indices]
+ )
else:
stratify = None
@@ -108,40 +119,55 @@ def _fit_with_budget(
task_type: int,
) -> None:
if (
- budget_type == 'iterations'
- or budget_type == 'mixed' and model.estimator_supports_iterative_fit()
+ budget_type == "iterations"
+ or budget_type == "mixed"
+ and model.estimator_supports_iterative_fit()
):
if model.estimator_supports_iterative_fit():
budget_factor = model.get_max_iter()
Xt, fit_params = model.fit_transformer(
- X_train.iloc[train_indices] if hasattr(X_train, 'iloc') else X_train[train_indices],
- Y_train.iloc[train_indices] if hasattr(Y_train, 'iloc') else Y_train[train_indices],
+ X_train.iloc[train_indices]
+ if hasattr(X_train, "iloc")
+ else X_train[train_indices],
+ Y_train.iloc[train_indices]
+ if hasattr(Y_train, "iloc")
+ else Y_train[train_indices],
)
n_iter = int(np.ceil(budget / 100 * budget_factor))
model.iterative_fit(
Xt,
- Y_train.iloc[train_indices] if hasattr(Y_train, 'iloc') else Y_train[train_indices],
+ Y_train.iloc[train_indices]
+ if hasattr(Y_train, "iloc")
+ else Y_train[train_indices],
n_iter=n_iter,
refit=True,
- **fit_params
+ **fit_params,
)
else:
_fit_and_suppress_warnings(
logger,
model,
- X_train.iloc[train_indices] if hasattr(X_train, 'iloc') else X_train[train_indices],
- Y_train.iloc[train_indices] if hasattr(Y_train, 'iloc') else Y_train[train_indices],
+ X_train.iloc[train_indices]
+ if hasattr(X_train, "iloc")
+ else X_train[train_indices],
+ Y_train.iloc[train_indices]
+ if hasattr(Y_train, "iloc")
+ else Y_train[train_indices],
)
elif (
- budget_type == 'subsample'
- or budget_type == 'mixed' and not model.estimator_supports_iterative_fit()
+ budget_type == "subsample"
+ or budget_type == "mixed"
+ and not model.estimator_supports_iterative_fit()
):
subsample = budget / 100
train_indices_subset = subsample_indices(
- train_indices, subsample, task_type, Y_train,
+ train_indices,
+ subsample,
+ task_type,
+ Y_train,
)
_fit_and_suppress_warnings(
logger,
@@ -154,21 +180,44 @@ def _fit_with_budget(
raise ValueError(budget_type)
+def concat_data(
+ data: List[Any], num_cv_folds: int
+) -> Union[np.ndarray, pd.DataFrame, scipy.sparse.csr_matrix]:
+ if isinstance(data[0], np.ndarray):
+ return np.concatenate(
+ [data[i] for i in range(num_cv_folds) if data[i] is not None]
+ )
+ elif isinstance(data[0], scipy.sparse.spmatrix):
+ return scipy.sparse.vstack(
+ [data[i] for i in range(num_cv_folds) if data[i] is not None]
+ )
+ elif isinstance(data[0], pd.DataFrame):
+ return pd.concat(
+ [data[i] for i in range(num_cv_folds) if data[i] is not None],
+ axis=0,
+ )
+ else:
+ raise ValueError(f"Unknown datatype {type(data[0])}")
+
+
class TrainEvaluator(AbstractEvaluator):
def __init__(
self,
backend: Backend,
queue: multiprocessing.Queue,
- metric: Scorer,
+ metrics: Sequence[Scorer],
additional_components: Dict[str, ThirdPartyComponents],
port: Optional[int],
configuration: Optional[Union[int, Configuration]] = None,
scoring_functions: Optional[List[Scorer]] = None,
seed: int = 1,
output_y_hat_optimization: bool = True,
- resampling_strategy: Optional[Union[str, BaseCrossValidator,
- _RepeatedSplits, BaseShuffleSplit]] = None,
- resampling_strategy_args: Optional[Dict[str, Optional[Union[float, int, str]]]] = None,
+ resampling_strategy: Optional[
+ Union[str, BaseCrossValidator, _RepeatedSplits, BaseShuffleSplit]
+ ] = None,
+ resampling_strategy_args: Optional[
+ Dict[str, Optional[Union[float, int, str]]]
+ ] = None,
num_run: Optional[int] = None,
budget: Optional[float] = None,
budget_type: Optional[str] = None,
@@ -184,7 +233,7 @@ def __init__(
queue=queue,
port=port,
configuration=configuration,
- metric=metric,
+ metrics=metrics,
additional_components=additional_components,
scoring_functions=scoring_functions,
seed=seed,
@@ -205,15 +254,17 @@ def __init__(
self.resampling_strategy_args = resampling_strategy_args
self.splitter = self.get_splitter(self.datamanager)
self.num_cv_folds = self.splitter.get_n_splits(
- groups=self.resampling_strategy_args.get('groups')
+ groups=self.resampling_strategy_args.get("groups")
)
- self.X_train = self.datamanager.data['X_train']
- self.Y_train = self.datamanager.data['Y_train']
- self.Y_optimization: Optional[SUPPORTED_TARGET_TYPES] = None
+ self.X_train = self.datamanager.data["X_train"]
+ self.Y_train = self.datamanager.data["Y_train"]
+ self.X_targets = [None] * self.num_cv_folds
self.Y_targets = [None] * self.num_cv_folds
self.Y_train_targets = np.ones(self.Y_train.shape) * np.NaN
self.models = [None] * self.num_cv_folds
- self.indices: List[Optional[Tuple[List[int], List[int]]]] = [None] * self.num_cv_folds
+ self.indices: List[Optional[Tuple[List[int], List[int]]]] = [
+ None
+ ] * self.num_cv_folds
# Necessary for full CV. Makes full CV not write predictions if only
# a subset of folds is evaluated but time is up. Complicated, because
@@ -224,8 +275,8 @@ def __init__(
def fit_predict_and_loss(self, iterative: bool = False) -> None:
"""Fit, predict and compute the loss for cross-validation and
- holdout (both iterative and non-iterative)"""
-
+ holdout (both iterative and non-iterative)
+ """
# Define beforehand for mypy
additional_run_info: Optional[TYPE_ADDITIONAL_INFO] = None
@@ -233,14 +284,23 @@ def fit_predict_and_loss(self, iterative: bool = False) -> None:
if self.num_cv_folds == 1:
for train_split, test_split in self.splitter.split(
- self.X_train, self.Y_train,
- groups=self.resampling_strategy_args.get('groups')
+ self.X_train,
+ self.Y_train,
+ groups=self.resampling_strategy_args.get("groups"),
):
+ self.X_optimization = (
+ self.X_train.iloc[test_split]
+ if hasattr(self.X_train, "iloc")
+ else self.X_train[test_split]
+ )
self.Y_optimization = self.Y_train[test_split]
self.Y_actual_train = self.Y_train[train_split]
- self._partial_fit_and_predict_iterative(0, train_indices=train_split,
- test_indices=test_split,
- add_model_to_self=True)
+ self._partial_fit_and_predict_iterative(
+ 0,
+ train_indices=train_split,
+ test_indices=test_split,
+ add_model_to_self=True,
+ )
else:
# Test if the model allows for an iterative fit, if not,
@@ -256,7 +316,6 @@ def fit_predict_and_loss(self, iterative: bool = False) -> None:
Y_train_pred = [None] * self.num_cv_folds
Y_optimization_pred = [None] * self.num_cv_folds
- Y_valid_pred = [None] * self.num_cv_folds
Y_test_pred = [None] * self.num_cv_folds
train_splits = [None] * self.num_cv_folds
@@ -265,16 +324,24 @@ def fit_predict_and_loss(self, iterative: bool = False) -> None:
total_n_iterations = [0] * self.num_cv_folds
# model.estimator_supports_iterative_fit -> true
# After the if above, we know estimator support iterative fit
- model_max_iter = [cast(IterativeComponent, model).get_max_iter()
- for model in self.models]
-
- if self.budget_type in ['iterations', 'mixed'] and self.budget is None:
- raise ValueError(f"When budget type is {self.budget_type} the budget "
- "can not be None")
+ model_max_iter = [
+ cast(IterativeComponent, model).get_max_iter()
+ for model in self.models
+ ]
+
+ if self.budget_type in ["iterations", "mixed"] and self.budget is None:
+ raise ValueError(
+ f"When budget type is {self.budget_type} the budget "
+ "can not be None"
+ )
- if self.budget_type in ['iterations', 'mixed'] and cast(float, self.budget) > 0:
+ if (
+ self.budget_type in ["iterations", "mixed"]
+ and cast(float, self.budget) > 0
+ ):
max_n_iter_budget = int(
- np.ceil(cast(float, self.budget) / 100 * model_max_iter[0]))
+ np.ceil(cast(float, self.budget) / 100 * model_max_iter[0])
+ )
max_iter = min(model_max_iter[0], max_n_iter_budget)
else:
max_iter = model_max_iter[0]
@@ -282,12 +349,14 @@ def fit_predict_and_loss(self, iterative: bool = False) -> None:
models_current_iters = [0] * self.num_cv_folds
Xt_array = [None] * self.num_cv_folds
- fit_params_array = [{}] * self.num_cv_folds # type: List[Dict[str, Any]]
+ fit_params_array = [
+ {}
+ ] * self.num_cv_folds # type: List[Dict[str, Any]]
y = _get_y_array(self.Y_train, self.task_type)
- # stores train loss of each fold.
- train_losses = [np.NaN] * self.num_cv_folds
+ # stores train loss(es) of each fold.
+ train_losses = [dict()] * self.num_cv_folds
# used as weights when averaging train losses.
train_fold_weights = [np.NaN] * self.num_cv_folds
# stores opt (validation) loss of each fold.
@@ -299,26 +368,39 @@ def fit_predict_and_loss(self, iterative: bool = False) -> None:
splitter = self.get_splitter(self.datamanager)
- for i, (train_indices, test_indices) in enumerate(splitter.split(
- self.X_train, y,
- groups=self.resampling_strategy_args.get('groups')
- )):
+ for i, (train_indices, test_indices) in enumerate(
+ splitter.split(
+ self.X_train,
+ y,
+ groups=self.resampling_strategy_args.get("groups"),
+ )
+ ):
if converged[i]:
continue
model = self.models[i]
if iterations[i] == 1:
- self.Y_train_targets[train_indices] = \
- self.Y_train.iloc[train_indices] if hasattr(
- self.Y_train, 'iloc') else self.Y_train[train_indices]
+ self.Y_train_targets[train_indices] = (
+ self.Y_train.iloc[train_indices]
+ if hasattr(self.Y_train, "iloc")
+ else self.Y_train[train_indices]
+ )
+ self.X_targets[i] = (
+ self.X_train.iloc[test_indices]
+ if hasattr(self.X_train, "iloc")
+ else self.X_train[train_indices]
+ )
+
self.Y_targets[i] = self.Y_train[test_indices]
Xt, fit_params = model.fit_transformer(
- self.X_train.iloc[train_indices] if hasattr(
- self.X_train, 'iloc') else self.X_train[train_indices],
- self.Y_train.iloc[train_indices] if hasattr(
- self.Y_train, 'iloc') else self.Y_train[train_indices],
+ self.X_train.iloc[train_indices]
+ if hasattr(self.X_train, "iloc")
+ else self.X_train[train_indices],
+ self.Y_train.iloc[train_indices]
+ if hasattr(self.Y_train, "iloc")
+ else self.Y_train[train_indices],
)
Xt_array[i] = Xt
fit_params_array[i] = fit_params
@@ -327,17 +409,14 @@ def fit_predict_and_loss(self, iterative: bool = False) -> None:
model.iterative_fit(
Xt_array[i],
- self.Y_train.iloc[train_indices] if hasattr(
- self.Y_train, 'iloc') else self.Y_train[train_indices],
- n_iter=n_iter, **fit_params_array[i]
+ self.Y_train.iloc[train_indices]
+ if hasattr(self.Y_train, "iloc")
+ else self.Y_train[train_indices],
+ n_iter=n_iter,
+ **fit_params_array[i],
)
- (
- train_pred,
- opt_pred,
- valid_pred,
- test_pred
- ) = self._predict(
+ (train_pred, opt_pred, test_pred) = self._predict(
model,
train_indices=train_indices,
test_indices=test_indices,
@@ -345,26 +424,24 @@ def fit_predict_and_loss(self, iterative: bool = False) -> None:
Y_train_pred[i] = train_pred
Y_optimization_pred[i] = opt_pred
- Y_valid_pred[i] = valid_pred
Y_test_pred[i] = test_pred
train_splits[i] = train_indices
- # Compute train loss of this fold and store it. train_loss could
- # either be a scalar or a dict of scalars with metrics as keys.
train_loss = self._loss(
- self.Y_train.iloc[train_indices] if hasattr(
- self.Y_train, 'iloc') else self.Y_train[train_indices],
+ self.Y_train.iloc[train_indices]
+ if hasattr(self.Y_train, "iloc")
+ else self.Y_train[train_indices],
train_pred,
+ X_data=Xt_array[i],
)
train_losses[i] = train_loss
- # number of training data points for this fold. Used for weighting
- # the average.
+ # Number of training data points for this fold.
+ # Used for weighting the average.
train_fold_weights[i] = len(train_indices)
# Compute validation loss of this fold and store it.
optimization_loss = self._loss(
- self.Y_targets[i],
- opt_pred,
+ self.Y_targets[i], opt_pred, X_data=self.X_targets[i]
)
opt_losses[i] = optimization_loss
# number of optimization data points for this fold.
@@ -381,8 +458,8 @@ def fit_predict_and_loss(self, iterative: bool = False) -> None:
iterations[i] = iterations[i] + 1
- # Compute weights of each fold based on the number of samples in each
- # fold.
+ # Compute weights of each fold based on the number of samples
+ # in each fold.
train_fold_weights_percentage = [
w / sum(train_fold_weights) for w in train_fold_weights
]
@@ -390,76 +467,67 @@ def fit_predict_and_loss(self, iterative: bool = False) -> None:
w / sum(opt_fold_weights) for w in opt_fold_weights
]
- # train_losses is a list of either scalars or dicts. If it contains
- # dicts, then train_loss is computed using the target metric
- # (self.metric).
- if all(isinstance(elem, dict) for elem in train_losses):
- train_loss = np.average([train_losses[i][str(self.metric)]
- for i in range(self.num_cv_folds)],
- weights=train_fold_weights_percentage,
- )
- else:
- train_loss = np.average(train_losses, weights=train_fold_weights_percentage)
+ train_loss = {
+ metric.name: np.average(
+ [
+ train_losses[i][str(metric)]
+ for i in range(self.num_cv_folds)
+ ],
+ weights=train_fold_weights_percentage,
+ )
+ for metric in self.metrics
+ }
# if all_scoring_function is true, return a dict of opt_loss.
# Otherwise, return a scalar.
- if self.scoring_functions:
- opt_loss = {}
- for metric in opt_losses[0].keys():
- opt_loss[metric] = np.average(
- [
- opt_losses[i][metric]
- for i in range(self.num_cv_folds)
- ],
- weights=opt_fold_weights_percentage,
- )
- else:
- opt_loss = np.average(opt_losses, weights=opt_fold_weights_percentage)
+ opt_loss = {}
+ for metric in opt_losses[0].keys():
+ opt_loss[metric] = np.average(
+ [opt_losses[i][metric] for i in range(self.num_cv_folds)],
+ weights=opt_fold_weights_percentage,
+ )
+ X_targets = self.X_targets
Y_targets = self.Y_targets
Y_train_targets = self.Y_train_targets
- Y_optimization_preds = np.concatenate(
- [Y_optimization_pred[i] for i in range(self.num_cv_folds)
- if Y_optimization_pred[i] is not None])
- Y_targets = np.concatenate([
- Y_targets[i] for i in range(self.num_cv_folds)
- if Y_targets[i] is not None
- ])
-
- if self.X_valid is not None:
- Y_valid_preds = np.array([Y_valid_pred[i]
- for i in range(self.num_cv_folds)
- if Y_valid_pred[i] is not None])
- # Average the predictions of several models
- if len(Y_valid_preds.shape) == 3:
- Y_valid_preds = np.nanmean(Y_valid_preds, axis=0)
- else:
- Y_valid_preds = None
+ Y_optimization_pred_concat = concat_data(
+ Y_optimization_pred, num_cv_folds=self.num_cv_folds
+ )
+ X_targets = concat_data(X_targets, num_cv_folds=self.num_cv_folds)
+ Y_targets = concat_data(Y_targets, num_cv_folds=self.num_cv_folds)
if self.X_test is not None:
- Y_test_preds = np.array([Y_test_pred[i]
- for i in range(self.num_cv_folds)
- if Y_test_pred[i] is not None])
+ Y_test_preds = np.array(
+ [
+ Y_test_pred[i]
+ for i in range(self.num_cv_folds)
+ if Y_test_pred[i] is not None
+ ]
+ )
# Average the predictions of several models
if len(Y_test_preds.shape) == 3:
Y_test_preds = np.nanmean(Y_test_preds, axis=0)
else:
Y_test_preds = None
+ self.X_optimization = X_targets
self.Y_optimization = Y_targets
self.Y_actual_train = Y_train_targets
self.model = self._get_model()
status = StatusType.DONOTADVANCE
- if any([model_current_iter == max_iter
- for model_current_iter in models_current_iters]):
+ if any(
+ [
+ model_current_iter == max_iter
+ for model_current_iter in models_current_iters
+ ]
+ ):
status = StatusType.SUCCESS
self.finish_up(
loss=opt_loss,
train_loss=train_loss,
- opt_pred=Y_optimization_preds,
- valid_pred=Y_valid_preds,
+ opt_pred=Y_optimization_pred_concat,
test_pred=Y_test_preds,
additional_run_info=additional_run_info,
file_output=True,
@@ -473,7 +541,6 @@ def fit_predict_and_loss(self, iterative: bool = False) -> None:
Y_train_pred = [None] * self.num_cv_folds
Y_optimization_pred = [None] * self.num_cv_folds
- Y_valid_pred = [None] * self.num_cv_folds
Y_test_pred = [None] * self.num_cv_folds
train_splits = [None] * self.num_cv_folds
@@ -487,10 +554,11 @@ def fit_predict_and_loss(self, iterative: bool = False) -> None:
# TODO: mention that no additional run info is possible in this
# case! -> maybe remove full CV from the train evaluator anyway and
# make the user implement this!
- for i, (train_split, test_split) in enumerate(self.splitter.split(
- self.X_train, y,
- groups=self.resampling_strategy_args.get('groups')
- )):
+ for i, (train_split, test_split) in enumerate(
+ self.splitter.split(
+ self.X_train, y, groups=self.resampling_strategy_args.get("groups")
+ )
+ ):
# TODO add check that split is actually an integer array,
# not a boolean array (to allow indexed assignement of
@@ -500,27 +568,25 @@ def fit_predict_and_loss(self, iterative: bool = False) -> None:
(
train_pred,
opt_pred,
- valid_pred,
test_pred,
additional_run_info,
- ) = (
- self._partial_fit_and_predict_standard(
- i, train_indices=train_split, test_indices=test_split,
- add_model_to_self=self.num_cv_folds == 1,
- )
+ ) = self._partial_fit_and_predict_standard(
+ i,
+ train_indices=train_split,
+ test_indices=test_split,
+ add_model_to_self=self.num_cv_folds == 1,
)
else:
(
train_pred,
opt_pred,
- valid_pred,
test_pred,
additional_run_info,
- ) = (
- self._partial_fit_and_predict_budget(
- i, train_indices=train_split, test_indices=test_split,
- add_model_to_self=self.num_cv_folds == 1,
- )
+ ) = self._partial_fit_and_predict_budget(
+ i,
+ train_indices=train_split,
+ test_indices=test_split,
+ add_model_to_self=self.num_cv_folds == 1,
)
if (
@@ -530,21 +596,21 @@ def fit_predict_and_loss(self, iterative: bool = False) -> None:
):
raise TAEAbortException(
'Found additional run info "%s" in fold %d, '
- 'but cannot handle additional run info if fold >= 1.' %
- (additional_run_info, i)
+ "but cannot handle additional run info if fold >= 1."
+ % (additional_run_info, i)
)
Y_train_pred[i] = train_pred
Y_optimization_pred[i] = opt_pred
- Y_valid_pred[i] = valid_pred
Y_test_pred[i] = test_pred
train_splits[i] = train_split
- # Compute train loss of this fold and store it. train_loss could
- # either be a scalar or a dict of scalars with metrics as keys.
train_loss = self._loss(
self.Y_train_targets[train_split],
train_pred,
+ X_data=self.X_train.iloc[train_split]
+ if hasattr(self.X_train, "iloc")
+ else self.X_train[train_split],
)
train_losses.append(train_loss)
# number of training data points for this fold. Used for weighting
@@ -553,8 +619,7 @@ def fit_predict_and_loss(self, iterative: bool = False) -> None:
# Compute validation loss of this fold and store it.
optimization_loss = self._loss(
- self.Y_targets[i],
- opt_pred,
+ self.Y_targets[i], opt_pred, X_data=self.X_targets[i]
)
opt_losses.append(optimization_loss)
# number of optimization data points for this fold. Used for weighting
@@ -563,56 +628,53 @@ def fit_predict_and_loss(self, iterative: bool = False) -> None:
# Compute weights of each fold based on the number of samples in each
# fold.
- train_fold_weights = [w / sum(train_fold_weights) for w in train_fold_weights]
+ train_fold_weights = [
+ w / sum(train_fold_weights) for w in train_fold_weights
+ ]
opt_fold_weights = [w / sum(opt_fold_weights) for w in opt_fold_weights]
- # train_losses is a list of either scalars or dicts. If it contains dicts,
- # then train_loss is computed using the target metric (self.metric).
- if all(isinstance(elem, dict) for elem in train_losses):
- train_loss = np.average([train_losses[i][str(self.metric)]
- for i in range(self.num_cv_folds)],
- weights=train_fold_weights,
- )
- else:
- train_loss = np.average(train_losses, weights=train_fold_weights)
+ train_loss = {
+ metric.name: np.average(
+ [train_losses[i][str(metric)] for i in range(self.num_cv_folds)],
+ weights=train_fold_weights,
+ )
+ for metric in self.metrics
+ }
# if all_scoring_function is true, return a dict of opt_loss. Otherwise,
# return a scalar.
- if self.scoring_functions:
- opt_loss = {}
- for metric in opt_losses[0].keys():
- opt_loss[metric] = np.average([opt_losses[i][metric]
- for i in range(self.num_cv_folds)],
- weights=opt_fold_weights,
- )
- else:
- opt_loss = np.average(opt_losses, weights=opt_fold_weights)
+ opt_loss = {}
+ for metric_name in list(opt_losses[0].keys()) + [
+ metric.name for metric in self.metrics
+ ]:
+ opt_loss[metric_name] = np.average(
+ [opt_losses[i][metric_name] for i in range(self.num_cv_folds)],
+ weights=opt_fold_weights,
+ )
+ X_targets = self.X_targets
Y_targets = self.Y_targets
Y_train_targets = self.Y_train_targets
- Y_optimization_pred = np.concatenate(
- [Y_optimization_pred[i] for i in range(self.num_cv_folds)
- if Y_optimization_pred[i] is not None])
- Y_targets = np.concatenate([Y_targets[i] for i in range(self.num_cv_folds)
- if Y_targets[i] is not None])
-
- if self.X_valid is not None:
- Y_valid_pred = np.array([Y_valid_pred[i]
- for i in range(self.num_cv_folds)
- if Y_valid_pred[i] is not None])
- # Average the predictions of several models
- if len(np.shape(Y_valid_pred)) == 3:
- Y_valid_pred = np.nanmean(Y_valid_pred, axis=0)
+ Y_optimization_pred = concat_data(
+ Y_optimization_pred, num_cv_folds=self.num_cv_folds
+ )
+ X_targets = concat_data(X_targets, num_cv_folds=self.num_cv_folds)
+ Y_targets = concat_data(Y_targets, num_cv_folds=self.num_cv_folds)
if self.X_test is not None:
- Y_test_pred = np.array([Y_test_pred[i]
- for i in range(self.num_cv_folds)
- if Y_test_pred[i] is not None])
+ Y_test_pred = np.array(
+ [
+ Y_test_pred[i]
+ for i in range(self.num_cv_folds)
+ if Y_test_pred[i] is not None
+ ]
+ )
# Average the predictions of several models
if len(np.shape(Y_test_pred)) == 3:
Y_test_pred = np.nanmean(Y_test_pred, axis=0)
+ self.X_optimization = X_targets
self.Y_optimization = Y_targets
self.Y_actual_train = Y_train_targets
@@ -624,8 +686,8 @@ def fit_predict_and_loss(self, iterative: bool = False) -> None:
# TODO check if there might be reasons for do-not-advance here!
status = StatusType.SUCCESS
elif (
- self.budget_type == 'iterations'
- or self.budget_type == 'mixed'
+ self.budget_type == "iterations"
+ or self.budget_type == "mixed"
and self.model.estimator_supports_iterative_fit()
):
budget_factor = self.model.get_max_iter()
@@ -651,7 +713,6 @@ def fit_predict_and_loss(self, iterative: bool = False) -> None:
loss=opt_loss,
train_loss=train_loss,
opt_pred=Y_optimization_pred,
- valid_pred=Y_valid_pred if self.X_valid is not None else None,
test_pred=Y_test_pred if self.X_test is not None else None,
additional_run_info=additional_run_info,
file_output=True,
@@ -660,42 +721,51 @@ def fit_predict_and_loss(self, iterative: bool = False) -> None:
)
def partial_fit_predict_and_loss(self, fold: int, iterative: bool = False) -> None:
- """Fit, predict and compute the loss for eval_partial_cv (both iterative and normal)"""
-
+ """Fit, predict and get loss for eval_partial_cv (iterative and normal)"""
if fold > self.num_cv_folds:
- raise ValueError('Cannot evaluate a fold %d which is higher than '
- 'the number of folds %d.' % (fold, self.num_cv_folds))
+ raise ValueError(
+ "Cannot evaluate a fold %d which is higher than "
+ "the number of folds %d." % (fold, self.num_cv_folds)
+ )
if self.budget_type is not None:
raise NotImplementedError()
y = _get_y_array(self.Y_train, self.task_type)
- for i, (train_split, test_split) in enumerate(self.splitter.split(
- self.X_train, y,
- groups=self.resampling_strategy_args.get('groups')
- )):
+ for i, (train_split, test_split) in enumerate(
+ self.splitter.split(
+ self.X_train, y, groups=self.resampling_strategy_args.get("groups")
+ )
+ ):
if i != fold:
continue
else:
break
if self.num_cv_folds > 1:
+ self.X_optimization = self.X_train[test_split]
self.Y_optimization = self.Y_train[test_split]
self.Y_actual_train = self.Y_train[train_split]
if iterative:
self._partial_fit_and_predict_iterative(
- fold, train_indices=train_split, test_indices=test_split,
- add_model_to_self=True)
+ fold,
+ train_indices=train_split,
+ test_indices=test_split,
+ add_model_to_self=True,
+ )
elif self.budget_type is not None:
raise NotImplementedError()
else:
- train_pred, opt_pred, valid_pred, test_pred, additional_run_info = (
- self._partial_fit_and_predict_standard(
- fold,
- train_indices=train_split,
- test_indices=test_split,
- add_model_to_self=True,
- )
+ (
+ train_pred,
+ opt_pred,
+ test_pred,
+ additional_run_info,
+ ) = self._partial_fit_and_predict_standard(
+ fold,
+ train_indices=train_split,
+ test_indices=test_split,
+ add_model_to_self=True,
)
train_loss = self._loss(self.Y_actual_train, train_pred)
loss = self._loss(self.Y_targets[fold], opt_pred)
@@ -714,20 +784,23 @@ def partial_fit_predict_and_loss(self, fold: int, iterative: bool = False) -> No
loss=loss,
train_loss=train_loss,
opt_pred=opt_pred,
- valid_pred=valid_pred,
test_pred=test_pred,
file_output=False,
final_call=True,
additional_run_info=None,
- status=status
+ status=status,
)
- def _partial_fit_and_predict_iterative(self, fold: int, train_indices: List[int],
- test_indices: List[int],
- add_model_to_self: bool) -> None:
+ def _partial_fit_and_predict_iterative(
+ self,
+ fold: int,
+ train_indices: List[int],
+ test_indices: List[int],
+ add_model_to_self: bool,
+ ) -> None:
model = self._get_model()
- self.indices[fold] = ((train_indices, test_indices))
+ self.indices[fold] = (train_indices, test_indices)
# Do only output the files in the case of iterative holdout,
# In case of iterative partial cv, no file output is needed
@@ -736,14 +809,19 @@ def _partial_fit_and_predict_iterative(self, fold: int, train_indices: List[int]
if model.estimator_supports_iterative_fit():
Xt, fit_params = model.fit_transformer(
- self.X_train.iloc[train_indices] if hasattr(
- self.Y_train, 'iloc') else self.X_train[train_indices],
- self.Y_train.iloc[train_indices] if hasattr(
- self.Y_train, 'iloc') else self.Y_train[train_indices],
+ self.X_train.iloc[train_indices]
+ if hasattr(self.Y_train, "iloc")
+ else self.X_train[train_indices],
+ self.Y_train.iloc[train_indices]
+ if hasattr(self.Y_train, "iloc")
+ else self.Y_train[train_indices],
)
- self.Y_train_targets[train_indices] = self.Y_train.iloc[train_indices] if hasattr(
- self.Y_train, 'iloc') else self.Y_train[train_indices]
+ self.Y_train_targets[train_indices] = (
+ self.Y_train.iloc[train_indices]
+ if hasattr(self.Y_train, "iloc")
+ else self.Y_train[train_indices]
+ )
iteration = 1
total_n_iteration = 0
@@ -759,20 +837,17 @@ def _partial_fit_and_predict_iterative(self, fold: int, train_indices: List[int]
while (
not model.configuration_fully_fitted() and model_current_iter < max_iter
):
- n_iter = int(2**iteration/2) if iteration > 1 else 2
+ n_iter = int(2**iteration / 2) if iteration > 1 else 2
total_n_iteration += n_iter
model.iterative_fit(
Xt,
- self.Y_train.iloc[train_indices] if hasattr(
- self.Y_train, 'iloc') else self.Y_train[train_indices],
- n_iter=n_iter, **fit_params
+ self.Y_train.iloc[train_indices]
+ if hasattr(self.Y_train, "iloc")
+ else self.Y_train[train_indices],
+ n_iter=n_iter,
+ **fit_params,
)
- (
- Y_train_pred,
- Y_optimization_pred,
- Y_valid_pred,
- Y_test_pred
- ) = self._predict(
+ (Y_train_pred, Y_optimization_pred, Y_test_pred,) = self._predict(
model,
train_indices=train_indices,
test_indices=test_indices,
@@ -782,9 +857,10 @@ def _partial_fit_and_predict_iterative(self, fold: int, train_indices: List[int]
self.model = model
train_loss = self._loss(
- self.Y_train.iloc[train_indices] if hasattr(
- self.Y_train, 'iloc') else self.Y_train[train_indices],
- Y_train_pred
+ self.Y_train.iloc[train_indices]
+ if hasattr(self.Y_train, "iloc")
+ else self.Y_train[train_indices],
+ Y_train_pred,
)
loss = self._loss(self.Y_train[test_indices], Y_optimization_pred)
additional_run_info = model.get_additional_run_info()
@@ -804,7 +880,6 @@ def _partial_fit_and_predict_iterative(self, fold: int, train_indices: List[int]
loss=loss,
train_loss=train_loss,
opt_pred=Y_optimization_pred,
- valid_pred=Y_valid_pred,
test_pred=Y_test_pred,
additional_run_info=additional_run_info,
file_output=file_output,
@@ -819,15 +894,16 @@ def _partial_fit_and_predict_iterative(self, fold: int, train_indices: List[int]
(
Y_train_pred,
Y_optimization_pred,
- Y_valid_pred,
Y_test_pred,
- additional_run_info
- ) = self._partial_fit_and_predict_standard(fold, train_indices, test_indices,
- add_model_to_self)
+ additional_run_info,
+ ) = self._partial_fit_and_predict_standard(
+ fold, train_indices, test_indices, add_model_to_self
+ )
train_loss = self._loss(
- self.Y_train.iloc[train_indices] if hasattr(
- self.Y_train, 'iloc') else self.Y_train[train_indices],
- Y_train_pred
+ self.Y_train.iloc[train_indices]
+ if hasattr(self.Y_train, "iloc")
+ else self.Y_train[train_indices],
+ Y_train_pred,
)
loss = self._loss(self.Y_train[test_indices], Y_optimization_pred)
if self.model.estimator_supports_iterative_fit():
@@ -843,7 +919,6 @@ def _partial_fit_and_predict_iterative(self, fold: int, train_indices: List[int]
loss=loss,
train_loss=train_loss,
opt_pred=Y_optimization_pred,
- valid_pred=Y_valid_pred,
test_pred=Y_test_pred,
additional_run_info=additional_run_info,
file_output=file_output,
@@ -854,25 +929,29 @@ def _partial_fit_and_predict_iterative(self, fold: int, train_indices: List[int]
def _partial_fit_and_predict_standard(
self,
- fold: int, train_indices: List[int],
+ fold: int,
+ train_indices: List[int],
test_indices: List[int],
- add_model_to_self: bool = False
- ) -> Tuple[PIPELINE_DATA_DTYPE, # train_pred
- PIPELINE_DATA_DTYPE, # opt_pred
- PIPELINE_DATA_DTYPE, # valid_pred
- PIPELINE_DATA_DTYPE, # test_pred
- TYPE_ADDITIONAL_INFO]:
+ add_model_to_self: bool = False,
+ ) -> Tuple[
+ PIPELINE_DATA_DTYPE, # train_pred
+ PIPELINE_DATA_DTYPE, # opt_pred
+ PIPELINE_DATA_DTYPE, # test_pred
+ TYPE_ADDITIONAL_INFO,
+ ]:
model = self._get_model()
- self.indices[fold] = ((train_indices, test_indices))
+ self.indices[fold] = (train_indices, test_indices)
_fit_and_suppress_warnings(
self.logger,
model,
- self.X_train.iloc[train_indices] if hasattr(
- self.X_train, 'iloc') else self.X_train[train_indices],
- self.Y_train.iloc[train_indices] if hasattr(
- self.Y_train, 'iloc') else self.Y_train[train_indices],
+ self.X_train.iloc[train_indices]
+ if hasattr(self.X_train, "iloc")
+ else self.X_train[train_indices],
+ self.Y_train.iloc[train_indices]
+ if hasattr(self.Y_train, "iloc")
+ else self.Y_train[train_indices],
)
if add_model_to_self:
@@ -880,12 +959,23 @@ def _partial_fit_and_predict_standard(
else:
self.models[fold] = model
- self.Y_targets[fold] = self.Y_train.iloc[test_indices] if hasattr(
- self.Y_train, 'iloc') else self.Y_train[test_indices]
- self.Y_train_targets[train_indices] = self.Y_train.iloc[train_indices] if hasattr(
- self.Y_train, 'iloc') else self.Y_train[train_indices]
+ self.X_targets[fold] = (
+ self.X_train.iloc[test_indices]
+ if hasattr(self.X_train, "iloc")
+ else self.X_train[test_indices]
+ )
+ self.Y_targets[fold] = (
+ self.Y_train.iloc[test_indices]
+ if hasattr(self.Y_train, "iloc")
+ else self.Y_train[test_indices]
+ )
+ self.Y_train_targets[train_indices] = (
+ self.Y_train.iloc[train_indices]
+ if hasattr(self.Y_train, "iloc")
+ else self.Y_train[train_indices]
+ )
- train_pred, opt_pred, valid_pred, test_pred = self._predict(
+ train_pred, opt_pred, test_pred = self._predict(
model=model,
train_indices=train_indices,
test_indices=test_indices,
@@ -894,31 +984,36 @@ def _partial_fit_and_predict_standard(
return (
train_pred,
opt_pred,
- valid_pred,
test_pred,
additional_run_info,
)
def _partial_fit_and_predict_budget(
self,
- fold: int, train_indices: List[int],
+ fold: int,
+ train_indices: List[int],
test_indices: List[int],
add_model_to_self: bool = False,
- ) -> Tuple[PIPELINE_DATA_DTYPE, # train_pred
- PIPELINE_DATA_DTYPE, # opt_pred
- PIPELINE_DATA_DTYPE, # valid_pred
- PIPELINE_DATA_DTYPE, # test_pred
- TYPE_ADDITIONAL_INFO]:
+ ) -> Tuple[
+ PIPELINE_DATA_DTYPE, # train_pred
+ PIPELINE_DATA_DTYPE, # opt_pred
+ PIPELINE_DATA_DTYPE, # test_pred
+ TYPE_ADDITIONAL_INFO,
+ ]:
# This function is only called in the event budget is not None
# Add this statement for mypy
assert self.budget is not None
model = self._get_model()
- self.indices[fold] = ((train_indices, test_indices))
+ self.indices[fold] = (train_indices, test_indices)
+ self.X_targets[fold] = self.X_train[test_indices]
self.Y_targets[fold] = self.Y_train[test_indices]
- self.Y_train_targets[train_indices] = self.Y_train.iloc[train_indices] if hasattr(
- self.Y_train, 'iloc') else self.Y_train[train_indices],
+ self.Y_train_targets[train_indices] = (
+ self.Y_train.iloc[train_indices]
+ if hasattr(self.Y_train, "iloc")
+ else self.Y_train[train_indices],
+ )
_fit_with_budget(
X_train=self.X_train,
@@ -931,7 +1026,7 @@ def _partial_fit_and_predict_budget(
task_type=self.task_type,
)
- train_pred, opt_pred, valid_pred, test_pred = self._predict(
+ train_pred, opt_pred, test_pred = self._predict(
model,
train_indices=train_indices,
test_indices=test_indices,
@@ -946,98 +1041,101 @@ def _partial_fit_and_predict_budget(
return (
train_pred,
opt_pred,
- valid_pred,
test_pred,
additional_run_info,
)
- def _predict(self, model: BaseEstimator, test_indices: List[int],
- train_indices: List[int]) -> Tuple[PIPELINE_DATA_DTYPE,
- PIPELINE_DATA_DTYPE,
- PIPELINE_DATA_DTYPE,
- PIPELINE_DATA_DTYPE]:
+ def _predict(
+ self, model: BaseEstimator, test_indices: List[int], train_indices: List[int]
+ ) -> Tuple[PIPELINE_DATA_DTYPE, PIPELINE_DATA_DTYPE, PIPELINE_DATA_DTYPE]:
train_pred = self.predict_function(
- self.X_train.iloc[train_indices] if hasattr(
- self.X_train, 'iloc') else self.X_train[train_indices],
- model, self.task_type,
- self.Y_train.iloc[train_indices] if hasattr(
- self.Y_train, 'iloc') else self.Y_train[train_indices]
+ self.X_train.iloc[train_indices]
+ if hasattr(self.X_train, "iloc")
+ else self.X_train[train_indices],
+ model,
+ self.task_type,
+ self.Y_train.iloc[train_indices]
+ if hasattr(self.Y_train, "iloc")
+ else self.Y_train[train_indices],
)
opt_pred = self.predict_function(
- self.X_train.iloc[test_indices] if hasattr(
- self.X_train, 'iloc') else self.X_train[test_indices],
- model, self.task_type,
- self.Y_train.iloc[train_indices] if hasattr(
- self.Y_train, 'iloc') else self.Y_train[train_indices]
+ self.X_train.iloc[test_indices]
+ if hasattr(self.X_train, "iloc")
+ else self.X_train[test_indices],
+ model,
+ self.task_type,
+ self.Y_train.iloc[train_indices]
+ if hasattr(self.Y_train, "iloc")
+ else self.Y_train[train_indices],
)
- if self.X_valid is not None:
- X_valid = self.X_valid.copy()
- valid_pred = self.predict_function(X_valid, model,
- self.task_type,
- self.Y_train[train_indices])
- else:
- valid_pred = None
-
if self.X_test is not None:
X_test = self.X_test.copy()
test_pred = self.predict_function(
- X_test, model,
+ X_test,
+ model,
self.task_type,
- self.Y_train.iloc[train_indices] if hasattr(
- self.Y_train, 'iloc') else self.Y_train[train_indices]
+ self.Y_train.iloc[train_indices]
+ if hasattr(self.Y_train, "iloc")
+ else self.Y_train[train_indices],
)
else:
test_pred = None
- return train_pred, opt_pred, valid_pred, test_pred
+ return train_pred, opt_pred, test_pred
- def get_splitter(self, D: AbstractDataManager) -> Union[BaseCrossValidator, _RepeatedSplits,
- BaseShuffleSplit]:
+ def get_splitter(
+ self, D: AbstractDataManager
+ ) -> Union[BaseCrossValidator, _RepeatedSplits, BaseShuffleSplit]:
if self.resampling_strategy_args is None:
self.resampling_strategy_args = {}
- if (
- self.resampling_strategy is not None
- and not isinstance(self.resampling_strategy, str)
+ if self.resampling_strategy is not None and not isinstance(
+ self.resampling_strategy, str
):
- if 'groups' not in self.resampling_strategy_args:
- self.resampling_strategy_args['groups'] = None
+ if "groups" not in self.resampling_strategy_args:
+ self.resampling_strategy_args["groups"] = None
- if isinstance(self.resampling_strategy, (BaseCrossValidator,
- _RepeatedSplits,
- BaseShuffleSplit)):
+ if isinstance(
+ self.resampling_strategy,
+ (BaseCrossValidator, _RepeatedSplits, BaseShuffleSplit),
+ ):
self.check_splitter_resampling_strategy(
- X=D.data['X_train'], y=D.data['Y_train'],
- groups=self.resampling_strategy_args.get('groups'),
- task=D.info['task'],
+ X=D.data["X_train"],
+ y=D.data["Y_train"],
+ groups=self.resampling_strategy_args.get("groups"),
+ task=D.info["task"],
resampling_strategy=self.resampling_strategy,
)
return self.resampling_strategy
# If it got to this point, we are dealing with a non-supported
# re-sampling strategy
- raise ValueError("Unsupported resampling strategy {}/{} provided".format(
- self.resampling_strategy,
- type(self.resampling_strategy),
- ))
+ raise ValueError(
+ "Unsupported resampling strategy {}/{} provided".format(
+ self.resampling_strategy,
+ type(self.resampling_strategy),
+ )
+ )
- y = D.data['Y_train']
- shuffle = self.resampling_strategy_args.get('shuffle', True)
+ y = D.data["Y_train"]
+ shuffle = self.resampling_strategy_args.get("shuffle", True)
train_size = 0.67
if self.resampling_strategy_args:
- train_size_from_user = self.resampling_strategy_args.get('train_size')
+ train_size_from_user = self.resampling_strategy_args.get("train_size")
if train_size_from_user is not None:
train_size = float(train_size_from_user)
test_size = float("%.4f" % (1 - train_size))
- if D.info['task'] in CLASSIFICATION_TASKS and D.info['task'] != MULTILABEL_CLASSIFICATION:
+ if (
+ D.info["task"] in CLASSIFICATION_TASKS
+ and D.info["task"] != MULTILABEL_CLASSIFICATION
+ ):
y = y.ravel()
- if self.resampling_strategy in ['holdout',
- 'holdout-iterative-fit']:
+ if self.resampling_strategy in ["holdout", "holdout-iterative-fit"]:
if shuffle:
try:
@@ -1049,7 +1147,7 @@ def get_splitter(self, D: AbstractDataManager) -> Union[BaseCrossValidator, _Rep
test_cv = copy.deepcopy(cv)
next(test_cv.split(y, y))
except ValueError as e:
- if 'The least populated class in y has only' in e.args[0]:
+ if "The least populated class in y has only" in e.args[0]:
cv = CustomStratifiedShuffleSplit(
n_splits=1,
test_size=test_size,
@@ -1063,14 +1161,18 @@ def get_splitter(self, D: AbstractDataManager) -> Union[BaseCrossValidator, _Rep
test_fold[:tmp_train_size] = -1
cv = PredefinedSplit(test_fold=test_fold)
cv.n_splits = 1 # As sklearn is inconsistent here
- elif self.resampling_strategy in ['cv', 'cv-iterative-fit', 'partial-cv',
- 'partial-cv-iterative-fit']:
+ elif self.resampling_strategy in [
+ "cv",
+ "cv-iterative-fit",
+ "partial-cv",
+ "partial-cv-iterative-fit",
+ ]:
if shuffle:
try:
with warnings.catch_warnings():
- warnings.simplefilter('error')
+ warnings.simplefilter("error")
cv = StratifiedKFold(
- n_splits=self.resampling_strategy_args['folds'],
+ n_splits=self.resampling_strategy_args["folds"],
shuffle=shuffle,
random_state=1,
)
@@ -1078,37 +1180,39 @@ def get_splitter(self, D: AbstractDataManager) -> Union[BaseCrossValidator, _Rep
next(test_cv.split(y, y))
except UserWarning as e:
print(e)
- if 'The least populated class in y has only' in e.args[0]:
+ if "The least populated class in y has only" in e.args[0]:
cv = CustomStratifiedKFold(
- n_splits=self.resampling_strategy_args['folds'],
+ n_splits=self.resampling_strategy_args["folds"],
shuffle=shuffle,
random_state=1,
)
else:
raise e
else:
- cv = KFold(n_splits=self.resampling_strategy_args['folds'],
- shuffle=shuffle)
+ cv = KFold(
+ n_splits=self.resampling_strategy_args["folds"], shuffle=shuffle
+ )
else:
raise ValueError(self.resampling_strategy)
else:
- if self.resampling_strategy in ['holdout',
- 'holdout-iterative-fit']:
+ if self.resampling_strategy in ["holdout", "holdout-iterative-fit"]:
# TODO shuffle not taken into account for this
if shuffle:
- cv = ShuffleSplit(n_splits=1, test_size=test_size,
- random_state=1)
+ cv = ShuffleSplit(n_splits=1, test_size=test_size, random_state=1)
else:
tmp_train_size = int(np.floor(train_size * y.shape[0]))
test_fold = np.zeros(y.shape[0])
test_fold[:tmp_train_size] = -1
cv = PredefinedSplit(test_fold=test_fold)
cv.n_splits = 1 # As sklearn is inconsistent here
- elif self.resampling_strategy in ['cv', 'partial-cv',
- 'partial-cv-iterative-fit']:
+ elif self.resampling_strategy in [
+ "cv",
+ "partial-cv",
+ "partial-cv-iterative-fit",
+ ]:
random_state = 1 if shuffle else None
cv = KFold(
- n_splits=self.resampling_strategy_args['folds'],
+ n_splits=self.resampling_strategy_args["folds"],
shuffle=shuffle,
random_state=random_state,
)
@@ -1123,16 +1227,14 @@ def check_splitter_resampling_strategy(
y: np.ndarray,
task: int,
groups: Any,
- resampling_strategy: Union[BaseCrossValidator, _RepeatedSplits,
- BaseShuffleSplit],
+ resampling_strategy: Union[
+ BaseCrossValidator, _RepeatedSplits, BaseShuffleSplit
+ ],
) -> None:
if (
task in CLASSIFICATION_TASKS
and task != MULTILABEL_CLASSIFICATION
- or (
- task in REGRESSION_TASKS
- and task != MULTIOUTPUT_REGRESSION
- )
+ or (task in REGRESSION_TASKS and task != MULTIOUTPUT_REGRESSION)
):
y = y.ravel()
@@ -1140,12 +1242,14 @@ def check_splitter_resampling_strategy(
resampling_strategy.get_n_splits(X=X, y=y, groups=groups)
next(resampling_strategy.split(X=X, y=y, groups=groups))
except Exception as e:
- raise ValueError("Unsupported resampling strategy "
- "{}/{} cause exception: {}".format(
- resampling_strategy,
- groups,
- str(e),
- ))
+ raise ValueError(
+ "Unsupported resampling strategy "
+ "{}/{} cause exception: {}".format(
+ resampling_strategy,
+ groups,
+ str(e),
+ )
+ )
# create closure for evaluating an algorithm
@@ -1153,9 +1257,11 @@ def eval_holdout(
queue: multiprocessing.Queue,
config: Union[int, Configuration],
backend: Backend,
- resampling_strategy: Union[str, BaseCrossValidator, _RepeatedSplits, BaseShuffleSplit],
+ resampling_strategy: Union[
+ str, BaseCrossValidator, _RepeatedSplits, BaseShuffleSplit
+ ],
resampling_strategy_args: Dict[str, Optional[Union[float, int, str]]],
- metric: Scorer,
+ metrics: Sequence[Scorer],
seed: int,
num_run: int,
instance: str,
@@ -1177,7 +1283,7 @@ def eval_holdout(
queue=queue,
resampling_strategy=resampling_strategy,
resampling_strategy_args=resampling_strategy_args,
- metric=metric,
+ metrics=metrics,
configuration=config,
seed=seed,
num_run=num_run,
@@ -1198,9 +1304,11 @@ def eval_iterative_holdout(
queue: multiprocessing.Queue,
config: Union[int, Configuration],
backend: Backend,
- resampling_strategy: Union[str, BaseCrossValidator, _RepeatedSplits, BaseShuffleSplit],
+ resampling_strategy: Union[
+ str, BaseCrossValidator, _RepeatedSplits, BaseShuffleSplit
+ ],
resampling_strategy_args: Dict[str, Optional[Union[float, int, str]]],
- metric: Scorer,
+ metrics: Sequence[Scorer],
seed: int,
num_run: int,
instance: str,
@@ -1220,7 +1328,7 @@ def eval_iterative_holdout(
port=port,
config=config,
backend=backend,
- metric=metric,
+ metrics=metrics,
resampling_strategy=resampling_strategy,
resampling_strategy_args=resampling_strategy_args,
seed=seed,
@@ -1235,7 +1343,7 @@ def eval_iterative_holdout(
additional_components=additional_components,
init_params=init_params,
budget=budget,
- budget_type=budget_type
+ budget_type=budget_type,
)
@@ -1243,9 +1351,11 @@ def eval_partial_cv(
queue: multiprocessing.Queue,
config: Union[int, Configuration],
backend: Backend,
- resampling_strategy: Union[str, BaseCrossValidator, _RepeatedSplits, BaseShuffleSplit],
+ resampling_strategy: Union[
+ str, BaseCrossValidator, _RepeatedSplits, BaseShuffleSplit
+ ],
resampling_strategy_args: Dict[str, Optional[Union[float, int, str]]],
- metric: Scorer,
+ metrics: Sequence[Scorer],
seed: int,
num_run: int,
instance: str,
@@ -1264,13 +1374,13 @@ def eval_partial_cv(
if budget_type is not None:
raise NotImplementedError()
instance_dict: Dict[str, int] = json.loads(instance) if instance is not None else {}
- fold = instance_dict['fold']
+ fold = instance_dict["fold"]
evaluator = TrainEvaluator(
backend=backend,
port=port,
queue=queue,
- metric=metric,
+ metrics=metrics,
configuration=config,
resampling_strategy=resampling_strategy,
resampling_strategy_args=resampling_strategy_args,
@@ -1294,9 +1404,11 @@ def eval_partial_cv_iterative(
queue: multiprocessing.Queue,
config: Union[int, Configuration],
backend: Backend,
- resampling_strategy: Union[str, BaseCrossValidator, _RepeatedSplits, BaseShuffleSplit],
+ resampling_strategy: Union[
+ str, BaseCrossValidator, _RepeatedSplits, BaseShuffleSplit
+ ],
resampling_strategy_args: Dict[str, Optional[Union[float, int, str]]],
- metric: Scorer,
+ metrics: Sequence[Scorer],
seed: int,
num_run: int,
instance: str,
@@ -1317,7 +1429,7 @@ def eval_partial_cv_iterative(
queue=queue,
config=config,
backend=backend,
- metric=metric,
+ metrics=metrics,
resampling_strategy=resampling_strategy,
resampling_strategy_args=resampling_strategy_args,
seed=seed,
@@ -1340,9 +1452,11 @@ def eval_cv(
queue: multiprocessing.Queue,
config: Union[int, Configuration],
backend: Backend,
- resampling_strategy: Union[str, BaseCrossValidator, _RepeatedSplits, BaseShuffleSplit],
+ resampling_strategy: Union[
+ str, BaseCrossValidator, _RepeatedSplits, BaseShuffleSplit
+ ],
resampling_strategy_args: Dict[str, Optional[Union[float, int, str]]],
- metric: Scorer,
+ metrics: Sequence[Scorer],
seed: int,
num_run: int,
instance: str,
@@ -1362,7 +1476,7 @@ def eval_cv(
backend=backend,
port=port,
queue=queue,
- metric=metric,
+ metrics=metrics,
configuration=config,
seed=seed,
num_run=num_run,
@@ -1386,9 +1500,11 @@ def eval_iterative_cv(
queue: multiprocessing.Queue,
config: Union[int, Configuration],
backend: Backend,
- resampling_strategy: Union[str, BaseCrossValidator, _RepeatedSplits, BaseShuffleSplit],
+ resampling_strategy: Union[
+ str, BaseCrossValidator, _RepeatedSplits, BaseShuffleSplit
+ ],
resampling_strategy_args: Dict[str, Optional[Union[float, int, str]]],
- metric: Scorer,
+ metrics: Sequence[Scorer],
seed: int,
num_run: int,
instance: str,
@@ -1407,7 +1523,7 @@ def eval_iterative_cv(
eval_cv(
backend=backend,
queue=queue,
- metric=metric,
+ metrics=metrics,
config=config,
seed=seed,
num_run=num_run,
diff --git a/autosklearn/evaluation/util.py b/autosklearn/evaluation/util.py
index e7483cbd26..158825786b 100644
--- a/autosklearn/evaluation/util.py
+++ b/autosklearn/evaluation/util.py
@@ -3,30 +3,28 @@
import multiprocessing
import queue
+__all__ = ["read_queue"]
-__all__ = [
- 'read_queue'
-]
-
-def read_queue(queue_: multiprocessing.Queue
- ) -> List[Dict[str, Union[str, bool, int, float, List, Dict, Tuple]]]:
+def read_queue(
+ queue_: multiprocessing.Queue,
+) -> List[Dict[str, Union[str, bool, int, float, List, Dict, Tuple]]]:
stack = []
while True:
try:
- rval = queue_.get(timeout=1)
+ return_value = queue_.get(timeout=1)
except queue.Empty:
break
# Check if there is a special placeholder value which tells us that
# we don't have to wait until the queue times out in order to
# retrieve the final value!
- if 'final_queue_element' in rval:
- del rval['final_queue_element']
+ if "final_queue_element" in return_value:
+ del return_value["final_queue_element"]
do_break = True
else:
do_break = False
- stack.append(rval)
+ stack.append(return_value)
if do_break:
break
@@ -46,12 +44,13 @@ def empty_queue(queue_: multiprocessing.Queue) -> None:
queue_.close()
-def extract_learning_curve(stack: List[Dict[str, Any]],
- key: Optional[str] = None) -> List[float]:
+def extract_learning_curve(
+ stack: List[Dict[str, Any]], key: Optional[str] = None
+) -> List[float]:
learning_curve = []
for entry in stack:
if key:
- learning_curve.append(entry['additional_run_info'][key])
+ learning_curve.append(entry["additional_run_info"][key])
else:
- learning_curve.append(entry['loss'])
+ learning_curve.append(entry["loss"])
return list(learning_curve)
diff --git a/autosklearn/experimental/askl2.py b/autosklearn/experimental/askl2.py
index 7cbeebc9d0..4303c953dc 100644
--- a/autosklearn/experimental/askl2.py
+++ b/autosklearn/experimental/askl2.py
@@ -1,75 +1,28 @@
+from __future__ import annotations
+
+from typing import Any, Dict, List, Mapping, Optional, Union
+
import hashlib
import json
import os
import pathlib
import pickle
-from typing import Any, Dict, List, Optional, Union
import dask.distributed
-import scipy.sparse
-
-from ConfigSpace import Configuration
import numpy as np
import pandas as pd
+import scipy.sparse
import sklearn
+from ConfigSpace import Configuration
import autosklearn
-from autosklearn.classification import AutoSklearnClassifier
import autosklearn.experimental.selector
-from autosklearn.metrics import Scorer, balanced_accuracy, roc_auc, log_loss, accuracy
-
-metrics = (balanced_accuracy, roc_auc, log_loss)
-selector_files = {}
-this_directory = pathlib.Path(__file__).resolve().parent
-for metric in metrics:
- training_data_file = this_directory / metric.name / 'askl2_training_data.json'
- with open(training_data_file) as fh:
- training_data = json.load(fh)
- fh.seek(0)
- m = hashlib.md5()
- m.update(fh.read().encode('utf8'))
- training_data_hash = m.hexdigest()[:10]
- selector_filename = "askl2_selector_%s_%s_%s_%s.pkl" % (
- autosklearn.__version__,
- sklearn.__version__,
- metric.name,
- training_data_hash
- )
- selector_directory = os.environ.get('XDG_CACHE_HOME')
- if selector_directory is None:
- selector_directory = pathlib.Path.home()
- selector_directory = pathlib.Path(selector_directory).joinpath('auto-sklearn').expanduser()
- selector_files[metric.name] = selector_directory / selector_filename
- metafeatures = pd.DataFrame(training_data['metafeatures'])
- strategies = training_data['strategies']
- y_values = pd.DataFrame(training_data['y_values'], columns=strategies, index=metafeatures.index)
- minima_for_methods = training_data['minima_for_methods']
- maxima_for_methods = training_data['maxima_for_methods']
- default_strategies = training_data['tie_break_order']
- if not selector_files[metric.name].exists():
- selector = autosklearn.experimental.selector.OVORF(
- configuration=training_data['configuration'],
- random_state=np.random.RandomState(1),
- n_estimators=500,
- tie_break_order=default_strategies,
- )
- selector = autosklearn.experimental.selector.FallbackWrapper(selector, default_strategies)
- selector.fit(
- X=metafeatures,
- y=y_values,
- minima=minima_for_methods,
- maxima=maxima_for_methods,
- )
- selector_files[metric.name].parent.mkdir(exist_ok=True, parents=True)
+from autosklearn.classification import AutoSklearnClassifier
+from autosklearn.ensembles.abstract_ensemble import AbstractEnsemble
+from autosklearn.ensembles.ensemble_selection import EnsembleSelection
+from autosklearn.metrics import Scorer, accuracy, balanced_accuracy, log_loss, roc_auc
- try:
- with open(selector_files[metric.name], 'wb') as fh:
- pickle.dump(selector, fh)
- except Exception as e:
- print("AutoSklearn2Classifier needs to create a selector file under "
- "the user's home directory or XDG_CACHE_HOME. Nevertheless "
- "the path {} is not writable.".format(selector_files[metric.name]))
- raise e
+selector_metrics = (balanced_accuracy, roc_auc, log_loss)
class SmacObjectCallback:
@@ -85,11 +38,13 @@ def __call__(
metalearning_configurations,
n_jobs,
dask_client,
+ multi_objective_algorithm,
+ multi_objective_kwargs,
):
from smac.facade.smac_ac_facade import SMAC4AC
+ from smac.intensification.simple_intensifier import SimpleIntensifier
from smac.runhistory.runhistory2epm import RunHistory2EPM4LogCost
from smac.scenario.scenario import Scenario
- from smac.intensification.simple_intensifier import SimpleIntensifier
scenario = Scenario(scenario_dict)
@@ -114,6 +69,8 @@ def __call__(
run_id=seed,
n_jobs=n_jobs,
dask_client=dask_client,
+ multi_objective_algorithm=multi_objective_algorithm,
+ multi_objective_kwargs=multi_objective_kwargs,
)
@@ -133,6 +90,8 @@ def __call__(
metalearning_configurations,
n_jobs,
dask_client,
+ multi_objective_algorithm,
+ multi_objective_kwargs,
):
from smac.facade.smac_ac_facade import SMAC4AC
from smac.intensification.successive_halving import SuccessiveHalving
@@ -151,7 +110,7 @@ def __call__(
pass
rh2EPM = RunHistory2EPM4LogCost
- ta_kwargs['budget_type'] = self.budget_type
+ ta_kwargs["budget_type"] = self.budget_type
smac4ac = SMAC4AC(
scenario=scenario,
@@ -163,13 +122,15 @@ def __call__(
run_id=seed,
intensifier=SuccessiveHalving,
intensifier_kwargs={
- 'initial_budget': self.initial_budget,
- 'max_budget': 100,
- 'eta': self.eta,
- 'min_chall': 1,
+ "initial_budget": self.initial_budget,
+ "max_budget": 100,
+ "eta": self.eta,
+ "min_chall": 1,
},
dask_client=dask_client,
n_jobs=n_jobs,
+ multi_objective_algorithm=multi_objective_algorithm,
+ multi_objective_kwargs=multi_objective_kwargs,
)
smac4ac.solver.epm_chooser.min_samples_model = int(
len(scenario.cs.get_hyperparameters()) / 2
@@ -178,12 +139,13 @@ def __call__(
class AutoSklearn2Classifier(AutoSklearnClassifier):
-
def __init__(
self,
time_left_for_this_task: int = 3600,
per_run_time_limit=None,
- ensemble_size: int = 50,
+ ensemble_size: int | None = None,
+ ensemble_class: AbstractEnsemble | None = EnsembleSelection,
+ ensemble_kwargs: Dict[str, Any] | None = None,
ensemble_nbest: Union[float, int] = 50,
max_models_on_disc: int = 50,
seed: int = 1,
@@ -198,6 +160,8 @@ def __init__(
metric: Optional[Scorer] = None,
scoring_functions: Optional[List[Scorer]] = None,
load_models: bool = True,
+ dataset_compression: Union[bool, Mapping[str, Any]] = True,
+ allow_string_features: bool = True,
):
"""
@@ -215,15 +179,25 @@ def __init__(
that typical machine learning algorithms can be fit on the
training data.
- ensemble_size : int, optional (default=50)
+ ensemble_size : int, optional
Number of models added to the ensemble built by *Ensemble
selection from libraries of models*. Models are drawn with
- replacement.
+ replacement. If set to ``0`` no ensemble is fit.
- ensemble_nbest : int, optional (default=50)
- Only consider the ``ensemble_nbest`` models when building an
+ Deprecated - will be removed in Auto-sklearn 0.16. Please pass
+ this argument via ``ensemble_kwargs={"ensemble_size": int}``
+ if you want to change the ensemble size for ensemble selection.
+
+ ensemble_class : Type[AbstractEnsemble], optional (default=EnsembleSelection)
+ Class implementing the post-hoc ensemble algorithm. Set to
+ ``None`` to disable ensemble building or use ``SingleBest``
+ to obtain only use the single best model instead of an
ensemble.
+ ensemble_kwargs : Dict, optional
+ Keyword arguments that are passed to the ensemble class upon
+ initialization.
+
max_models_on_disc: int, optional (default=50),
Defines the maximum number of models that are kept in the disc.
The additional number of models are permanently deleted. Due to the
@@ -238,10 +212,14 @@ def __init__(
memory_limit : int, optional (3072)
Memory limit in MB for the machine learning algorithm.
`auto-sklearn` will stop fitting the machine learning algorithm if
- it tries to allocate more than `memory_limit` MB.
- If None is provided, no memory limit is set.
- In case of multi-processing, `memory_limit` will be per job.
- This memory limit also applies to the ensemble creation process.
+ it tries to allocate more than ``memory_limit`` MB.
+
+ **Important notes:**
+
+ * If ``None`` is provided, no memory limit is set.
+ * In case of multi-processing, ``memory_limit`` will be *per job*, so the total usage is
+ ``n_jobs x memory_limit``.
+ * The memory limit also applies to the ensemble creation process.
tmp_folder : string, optional (None)
folder to store configuration output and log files, if ``None``
@@ -253,13 +231,15 @@ def __init__(
n_jobs : int, optional, experimental
The number of jobs to run in parallel for ``fit()``. ``-1`` means
- using all processors. By default, Auto-sklearn uses a single core
- for fitting the machine learning model and a single core for fitting
- an ensemble. Ensemble building is not affected by ``n_jobs`` but
- can be controlled by the number of models in the ensemble. In
- contrast to most scikit-learn models, ``n_jobs`` given in the
- constructor is not applied to the ``predict()`` method. If
- ``dask_client`` is None, a new dask client is created.
+ using all processors.
+
+ **Important notes**:
+
+ * By default, Auto-sklearn uses one core.
+ * Ensemble building is not affected by ``n_jobs`` but can be controlled by the number
+ of models in the ensemble.
+ * ``predict()`` is not affected by ``n_jobs`` (in contrast to most scikit-learn models)
+ * If ``dask_client`` is ``None``, a new dask client is created.
dask_client : dask.distributed.Client, optional
User-created dask client, can be used to start a dask cluster and then
@@ -275,7 +255,7 @@ def __init__(
* ``'y_optimization'`` : do not save the predictions for the
optimization/validation set, which would later on be used to build
an ensemble.
- * ``'model'`` : do not save any model files
+ * ``model`` : do not save any model files
smac_scenario_args : dict, optional (None)
Additional arguments inserted into the scenario of SMAC. See the
@@ -310,18 +290,28 @@ def __init__(
Not all keys returned by scikit-learn are supported yet.
""" # noqa (links are too long)
-
+ self.required_training = False # Boolean to indicate if selectors were trained.
include_estimators = [
- 'extra_trees', 'passive_aggressive', 'random_forest', 'sgd', 'gradient_boosting', 'mlp',
+ "extra_trees",
+ "passive_aggressive",
+ "random_forest",
+ "sgd",
+ "gradient_boosting",
+ "mlp",
]
include_preprocessors = ["no_preprocessing"]
- include = {'classifier': include_estimators,
- 'feature_preprocessor': include_preprocessors}
+ include = {
+ "classifier": include_estimators,
+ "feature_preprocessor": include_preprocessors,
+ }
+ self.train_selectors(selected_metric=metric)
super().__init__(
time_left_for_this_task=time_left_for_this_task,
per_run_time_limit=per_run_time_limit,
initial_configurations_via_metalearning=0,
ensemble_size=ensemble_size,
+ ensemble_class=ensemble_class,
+ ensemble_kwargs=ensemble_kwargs,
ensemble_nbest=ensemble_nbest,
max_models_on_disc=max_models_on_disc,
seed=seed,
@@ -342,33 +332,120 @@ def __init__(
metric=metric,
scoring_functions=scoring_functions,
load_models=load_models,
+ allow_string_features=allow_string_features,
)
- def fit(self, X, y,
- X_test=None,
- y_test=None,
- metric=None,
- feat_type=None,
- dataset_name=None):
+ def train_selectors(self, selected_metric=None):
+ self.selector_metrics = (balanced_accuracy, roc_auc, log_loss)
+ self.selector_files = {}
+ self.this_directory = pathlib.Path(__file__).resolve().parent
+
+ if selected_metric is not None:
+ metric_list = [selected_metric]
+ else:
+ metric_list = self.selector_metrics
+
+ for metric in metric_list:
+ training_data_file = (
+ self.this_directory / metric.name / "askl2_training_data.json"
+ )
+ with open(training_data_file) as fh:
+ training_data = json.load(fh)
+ fh.seek(0)
+ m = hashlib.md5()
+ m.update(fh.read().encode("utf8"))
+ training_data_hash = m.hexdigest()[:10]
+ selector_filename = "askl2_selector_%s_%s_%s_%s.pkl" % (
+ autosklearn.__version__,
+ sklearn.__version__,
+ metric.name,
+ training_data_hash,
+ )
+ selector_directory = os.environ.get("XDG_CACHE_HOME")
+ if selector_directory is None:
+ selector_directory = pathlib.Path.home()
+ selector_directory = (
+ pathlib.Path(selector_directory).joinpath("auto-sklearn").expanduser()
+ )
+ self.selector_files[metric.name] = selector_directory / selector_filename
+ metafeatures = pd.DataFrame(training_data["metafeatures"])
+ self.strategies = training_data["strategies"]
+ y_values = pd.DataFrame(
+ training_data["y_values"],
+ columns=self.strategies,
+ index=metafeatures.index,
+ )
+ minima_for_methods = training_data["minima_for_methods"]
+ maxima_for_methods = training_data["maxima_for_methods"]
+ default_strategies = training_data["tie_break_order"]
+ if not self.selector_files[metric.name].exists():
+ self.required_training = True
+ selector = autosklearn.experimental.selector.OVORF(
+ configuration=training_data["configuration"],
+ random_state=np.random.RandomState(1),
+ n_estimators=500,
+ tie_break_order=default_strategies,
+ )
+ selector = autosklearn.experimental.selector.FallbackWrapper(
+ selector, default_strategies
+ )
+ selector.fit(
+ X=metafeatures,
+ y=y_values,
+ minima=minima_for_methods,
+ maxima=maxima_for_methods,
+ )
+ self.selector_files[metric.name].parent.mkdir(
+ exist_ok=True, parents=True
+ )
+
+ try:
+ with open(self.selector_files[metric.name], "wb") as fh:
+ pickle.dump(selector, fh)
+ except Exception as e:
+ print(
+ "AutoSklearn2Classifier needs to create a selector file under "
+ "the user's home directory or XDG_CACHE_HOME. Nevertheless "
+ "the path {} is not writable.".format(
+ self.selector_files[metric.name]
+ )
+ )
+ raise e
+
+ def fit(
+ self,
+ X,
+ y,
+ X_test=None,
+ y_test=None,
+ metric=None,
+ feat_type=None,
+ dataset_name=None,
+ ):
# TODO
- # regularly check https://github.com/scikit-learn/scikit-learn/issues/15336 whether
- # histogram gradient boosting in scikit-learn finally support sparse data
+ # regularly check https://github.com/scikit-learn/scikit-learn/issues/15336
+ # whether histogram gradient boosting in scikit-learn finally support
+ # sparse data
is_sparse = scipy.sparse.issparse(X)
if is_sparse:
include_estimators = [
- 'extra_trees', 'passive_aggressive', 'random_forest', 'sgd', 'mlp',
+ "extra_trees",
+ "passive_aggressive",
+ "random_forest",
+ "sgd",
+ "mlp",
]
else:
include_estimators = [
- 'extra_trees',
- 'passive_aggressive',
- 'random_forest',
- 'sgd',
- 'gradient_boosting',
- 'mlp',
+ "extra_trees",
+ "passive_aggressive",
+ "random_forest",
+ "sgd",
+ "gradient_boosting",
+ "mlp",
]
- self.include['classifier'] = include_estimators
+ self.include["classifier"] = include_estimators
if self.metric is None:
if len(y.shape) == 1 or y.shape[1] == 1:
@@ -376,75 +453,80 @@ def fit(self, X, y,
else:
self.metric = log_loss
- if self.metric in metrics:
+ if self.metric in self.selector_metrics:
metric_name = self.metric.name
- selector_file = selector_files[metric_name]
+ selector_file = self.selector_files[metric_name]
else:
- metric_name = 'balanced_accuracy'
- selector_file = selector_files[metric_name]
- with open(selector_file, 'rb') as fh:
+ metric_name = "balanced_accuracy"
+ selector_file = self.selector_files[metric_name]
+ with open(selector_file, "rb") as fh:
selector = pickle.load(fh)
- metafeatures = pd.DataFrame({dataset_name: [X.shape[1], X.shape[0]]}).transpose()
+ metafeatures = pd.DataFrame(
+ {dataset_name: [X.shape[1], X.shape[0]]}
+ ).transpose()
selection = np.argmax(selector.predict(metafeatures))
- automl_policy = strategies[selection]
+ automl_policy = self.strategies[selection]
setting = {
- 'RF_None_holdout_iterative_es_if': {
- 'resampling_strategy': 'holdout-iterative-fit',
- 'fidelity': None,
+ "RF_None_holdout_iterative_es_if": {
+ "resampling_strategy": "holdout-iterative-fit",
+ "fidelity": None,
+ },
+ "RF_None_3CV_iterative_es_if": {
+ "resampling_strategy": "cv-iterative-fit",
+ "folds": 3,
+ "fidelity": None,
},
- 'RF_None_3CV_iterative_es_if': {
- 'resampling_strategy': 'cv-iterative-fit',
- 'folds': 3,
- 'fidelity': None,
+ "RF_None_5CV_iterative_es_if": {
+ "resampling_strategy": "cv-iterative-fit",
+ "folds": 5,
+ "fidelity": None,
},
- 'RF_None_5CV_iterative_es_if': {
- 'resampling_strategy': 'cv-iterative-fit',
- 'folds': 5,
- 'fidelity': None,
+ "RF_None_10CV_iterative_es_if": {
+ "resampling_strategy": "cv-iterative-fit",
+ "folds": 10,
+ "fidelity": None,
},
- 'RF_None_10CV_iterative_es_if': {
- 'resampling_strategy': 'cv-iterative-fit',
- 'folds': 10,
- 'fidelity': None,
+ "RF_SH-eta4-i_holdout_iterative_es_if": {
+ "resampling_strategy": "holdout-iterative-fit",
+ "fidelity": "SH",
},
- 'RF_SH-eta4-i_holdout_iterative_es_if': {
- 'resampling_strategy': 'holdout-iterative-fit',
- 'fidelity': 'SH',
+ "RF_SH-eta4-i_3CV_iterative_es_if": {
+ "resampling_strategy": "cv-iterative-fit",
+ "folds": 3,
+ "fidelity": "SH",
},
- 'RF_SH-eta4-i_3CV_iterative_es_if': {
- 'resampling_strategy': 'cv-iterative-fit',
- 'folds': 3,
- 'fidelity': 'SH',
+ "RF_SH-eta4-i_5CV_iterative_es_if": {
+ "resampling_strategy": "cv-iterative-fit",
+ "folds": 5,
+ "fidelity": "SH",
},
- 'RF_SH-eta4-i_5CV_iterative_es_if': {
- 'resampling_strategy': 'cv-iterative-fit',
- 'folds': 5,
- 'fidelity': 'SH',
+ "RF_SH-eta4-i_10CV_iterative_es_if": {
+ "resampling_strategy": "cv-iterative-fit",
+ "folds": 10,
+ "fidelity": "SH",
},
- 'RF_SH-eta4-i_10CV_iterative_es_if': {
- 'resampling_strategy': 'cv-iterative-fit',
- 'folds': 10,
- 'fidelity': 'SH',
- }
}[automl_policy]
- resampling_strategy = setting['resampling_strategy']
- if resampling_strategy == 'cv-iterative-fit':
- resampling_strategy_kwargs = {'folds': setting['folds']}
+ resampling_strategy = setting["resampling_strategy"]
+ if resampling_strategy == "cv-iterative-fit":
+ resampling_strategy_kwargs = {"folds": setting["folds"]}
else:
resampling_strategy_kwargs = None
portfolio_file = (
- this_directory / metric_name / 'askl2_portfolios' / ('%s.json' % automl_policy)
+ self.this_directory
+ / metric_name
+ / "askl2_portfolios"
+ / ("%s.json" % automl_policy)
)
with open(portfolio_file) as fh:
portfolio_json = json.load(fh)
- portfolio = portfolio_json['portfolio']
+ portfolio = portfolio_json["portfolio"]
- if setting['fidelity'] == 'SH':
- smac_callback = SHObjectCallback('iterations', 4, 5.0, portfolio)
+ if setting["fidelity"] == "SH":
+ smac_callback = SHObjectCallback("iterations", 4, 5.0, portfolio)
else:
smac_callback = SmacObjectCallback(portfolio)
diff --git a/autosklearn/experimental/balanced_accuracy/askl2_portfolios/RF_None_10CV_iterative_es_if.json b/autosklearn/experimental/balanced_accuracy/askl2_portfolios/RF_None_10CV_iterative_es_if.json
index 752ffd8213..217c28e771 100644
--- a/autosklearn/experimental/balanced_accuracy/askl2_portfolios/RF_None_10CV_iterative_es_if.json
+++ b/autosklearn/experimental/balanced_accuracy/askl2_portfolios/RF_None_10CV_iterative_es_if.json
@@ -20,7 +20,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.11840768063684459,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.7668727846319956,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.28633349214759163,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"f1500cd": {
"balancing:strategy": "weighting",
@@ -42,7 +47,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.0013138596672908974,
"classifier:gradient_boosting:n_iter_no_change": 13,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"3e1a332": {
"balancing:strategy": "weighting",
@@ -62,7 +72,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.006544845548292656,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1062,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"22a29ef": {
"balancing:strategy": "weighting",
@@ -86,7 +101,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
"classifier:gradient_boosting:n_iter_no_change": 6,
"classifier:gradient_boosting:validation_fraction": 0.11549557161401015,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"5886349": {
"balancing:strategy": "weighting",
@@ -108,7 +128,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"classifier:gradient_boosting:n_iter_no_change": 1,
"classifier:gradient_boosting:validation_fraction": 0.16202792486532844,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"8f54538": {
"balancing:strategy": "none",
@@ -133,7 +158,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.004667222369408079,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"9b6c297": {
"balancing:strategy": "weighting",
@@ -155,7 +185,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.7972710156738574,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.2243563424345074,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"4dffb08": {
"balancing:strategy": "weighting",
@@ -175,7 +210,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.0002547219008659275,
"classifier:sgd:eta0": 2.0514805804594057e-06,
"classifier:sgd:l1_ratio": 7.163773086996508e-09,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"73611bf": {
"balancing:strategy": "weighting",
@@ -197,7 +237,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.047507672861535975,
"classifier:gradient_boosting:n_iter_no_change": 11,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"4d5e802": {
"balancing:strategy": "weighting",
@@ -219,7 +264,12 @@
"classifier:sgd:epsilon": 0.009309372110873482,
"classifier:sgd:eta0": 9.309855728417528e-05,
"classifier:sgd:l1_ratio": 0.6620356750678396,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"43a743f": {
"balancing:strategy": "none",
@@ -242,7 +292,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 82,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
"classifier:gradient_boosting:n_iter_no_change": 3,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"eac6b8e": {
"balancing:strategy": "weighting",
@@ -262,7 +317,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.001003862016677866,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1987,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"78a0240": {
"balancing:strategy": "weighting",
@@ -282,7 +342,12 @@
"classifier:extra_trees:min_samples_split": 5,
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.0036912148544896695,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"809eef4": {
"balancing:strategy": "weighting",
@@ -303,7 +368,12 @@
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.8563603866144365,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.041173114185507675,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"44e2072": {
"balancing:strategy": "weighting",
@@ -323,7 +393,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.0017922018692146033,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1528,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"f7d054b": {
"balancing:strategy": "weighting",
@@ -344,7 +419,12 @@
"classifier:gradient_boosting:scoring": "loss",
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.009713456035294782,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"87cff92": {
"balancing:strategy": "none",
@@ -366,7 +446,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.027146552434145918,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.7300216362356305,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.23992030699419034,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"8e0e269": {
"balancing:strategy": "weighting",
@@ -388,7 +473,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"classifier:gradient_boosting:n_iter_no_change": 3,
"classifier:gradient_boosting:validation_fraction": 0.14053446064492747,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"6b55ec8": {
"balancing:strategy": "weighting",
@@ -403,7 +493,12 @@
"classifier:passive_aggressive:fit_intercept": "True",
"classifier:passive_aggressive:loss": "hinge",
"classifier:passive_aggressive:tol": 6.434802101116561e-05,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"4e98e28": {
"balancing:strategy": "weighting",
@@ -425,7 +520,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"classifier:gradient_boosting:n_iter_no_change": 7,
"classifier:gradient_boosting:validation_fraction": 0.35540373091502847,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"3a5c250": {
"balancing:strategy": "weighting",
@@ -447,7 +547,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.009333921831143037,
"classifier:gradient_boosting:n_iter_no_change": 12,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"0dd7e47": {
"balancing:strategy": "weighting",
@@ -470,7 +575,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.0002677236491698818,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.702784013363627,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.06696502879142427,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"7ff3956": {
"balancing:strategy": "weighting",
@@ -487,7 +597,12 @@
"classifier:passive_aggressive:tol": 0.000403053149783872,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 907,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"77861ef": {
"balancing:strategy": "weighting",
@@ -509,7 +624,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.019347782064094907,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.8630731567690261,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.04372049630950089,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"79812c6": {
"balancing:strategy": "weighting",
@@ -531,7 +651,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.15443993532080638,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 999,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"85bfdfa": {
"balancing:strategy": "weighting",
@@ -551,7 +676,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.011947142763501097,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1194,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"ce012ed": {
"balancing:strategy": "weighting",
@@ -566,7 +696,12 @@
"classifier:passive_aggressive:fit_intercept": "True",
"classifier:passive_aggressive:loss": "squared_hinge",
"classifier:passive_aggressive:tol": 0.009455868366528827,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"e71ff84": {
"balancing:strategy": "none",
@@ -593,7 +728,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.7875382396139146,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.1913513845124594,
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"d4abc34": {
"balancing:strategy": "weighting",
@@ -619,7 +759,12 @@
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.748076585259295,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.23337355967199702,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"283cac3": {
"balancing:strategy": "weighting",
@@ -638,7 +783,12 @@
"classifier:extra_trees:min_samples_leaf": 2,
"classifier:extra_trees:min_samples_split": 15,
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"808d532": {
"balancing:strategy": "none",
@@ -659,7 +809,12 @@
"classifier:gradient_boosting:scoring": "loss",
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.05184374776387598,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"d2d2039": {
"balancing:strategy": "weighting",
@@ -681,7 +836,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"classifier:gradient_boosting:n_iter_no_change": 19,
"classifier:gradient_boosting:validation_fraction": 0.29656195578950684,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
}
},
"cutoffs": [
diff --git a/autosklearn/experimental/balanced_accuracy/askl2_portfolios/RF_None_3CV_iterative_es_if.json b/autosklearn/experimental/balanced_accuracy/askl2_portfolios/RF_None_3CV_iterative_es_if.json
index ea49298d0c..023a79e741 100644
--- a/autosklearn/experimental/balanced_accuracy/askl2_portfolios/RF_None_3CV_iterative_es_if.json
+++ b/autosklearn/experimental/balanced_accuracy/askl2_portfolios/RF_None_3CV_iterative_es_if.json
@@ -19,7 +19,12 @@
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1025,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"1cdd4cf": {
"balancing:strategy": "weighting",
@@ -35,7 +40,12 @@
"classifier:passive_aggressive:loss": "hinge",
"classifier:passive_aggressive:tol": 0.001538331512469445,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.0020876231314801,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"08d784a": {
"balancing:strategy": "weighting",
@@ -58,7 +68,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1719,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
"classifier:gradient_boosting:n_iter_no_change": 1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"9b1161c": {
"balancing:strategy": "weighting",
@@ -79,7 +94,12 @@
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.9371966597485295,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.2181009648699602,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"e6b6325": {
"balancing:strategy": "weighting",
@@ -101,7 +121,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.010402618316459437,
"classifier:gradient_boosting:n_iter_no_change": 17,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"c138595": {
"balancing:strategy": "weighting",
@@ -124,7 +149,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.03127012992564812,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.864877718024256,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.1412101268042483,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"53634b7": {
"balancing:strategy": "none",
@@ -148,7 +178,12 @@
"classifier:mlp:shuffle": "True",
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"d91f995": {
"balancing:strategy": "weighting",
@@ -168,7 +203,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.01665326684623352,
"classifier:sgd:eta0": 0.0008683421715267357,
"classifier:sgd:l1_ratio": 0.0037742853051687883,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"ba079f7": {
"balancing:strategy": "weighting",
@@ -191,7 +231,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.010000000000000004,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.7279628505501666,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.24610303193556213,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"5251c86": {
"balancing:strategy": "none",
@@ -212,7 +257,12 @@
"classifier:gradient_boosting:scoring": "loss",
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.28959335532988506,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"56a06e6": {
"balancing:strategy": "weighting",
@@ -231,7 +281,12 @@
"classifier:extra_trees:min_samples_leaf": 1,
"classifier:extra_trees:min_samples_split": 9,
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"762f55d": {
"balancing:strategy": "weighting",
@@ -253,7 +308,12 @@
"classifier:sgd:epsilon": 0.0060184515243734615,
"classifier:sgd:eta0": 8.74381130399994e-05,
"classifier:sgd:power_t": 0.040470778304526875,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"2a3926a": {
"balancing:strategy": "none",
@@ -274,7 +334,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.7323011212890899,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.2405664543729542,
"classifier:sgd:eta0": 0.031719126506159497,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"7fa43cd": {
"balancing:strategy": "weighting",
@@ -298,7 +363,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.1765511759516072,
"classifier:gradient_boosting:n_iter_no_change": 9,
"classifier:gradient_boosting:validation_fraction": 0.08017220174930173,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"9616495": {
"balancing:strategy": "none",
@@ -325,7 +395,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1675,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"1a2077f": {
"balancing:strategy": "weighting",
@@ -345,7 +420,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.00025247448315347646,
"classifier:sgd:eta0": 0.004594829994128768,
"classifier:sgd:power_t": 0.5032138663090429,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"6bdd0f7": {
"balancing:strategy": "weighting",
@@ -365,7 +445,12 @@
"classifier:sgd:epsilon": 0.0042774156243000246,
"classifier:sgd:eta0": 0.007202966496306807,
"classifier:sgd:power_t": 0.7972709828085118,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"fcd1bda": {
"balancing:strategy": "none",
@@ -392,7 +477,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.01186630902011325,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1281,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"3eada54": {
"balancing:strategy": "weighting",
@@ -407,7 +497,12 @@
"classifier:passive_aggressive:fit_intercept": "True",
"classifier:passive_aggressive:loss": "hinge",
"classifier:passive_aggressive:tol": 0.004257493012053348,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"4b06ddd": {
"balancing:strategy": "weighting",
@@ -426,7 +521,12 @@
"classifier:extra_trees:min_samples_leaf": 5,
"classifier:extra_trees:min_samples_split": 11,
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"8d8ad31": {
"balancing:strategy": "weighting",
@@ -447,7 +547,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1985,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
"classifier:sgd:l1_ratio": 0.0752197101694388,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"3bdcff2": {
"balancing:strategy": "weighting",
@@ -472,7 +577,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.005930785926441465,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"42ab0ad": {
"balancing:strategy": "weighting",
@@ -494,7 +604,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 839,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"f28f4ab": {
"balancing:strategy": "weighting",
@@ -517,7 +632,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.010000000000000004,
"classifier:gradient_boosting:n_iter_no_change": 14,
"classifier:gradient_boosting:validation_fraction": 0.23881663494669972,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"f8aba41": {
"balancing:strategy": "weighting",
@@ -532,7 +652,12 @@
"classifier:passive_aggressive:fit_intercept": "True",
"classifier:passive_aggressive:loss": "hinge",
"classifier:passive_aggressive:tol": 0.010384183550709794,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"1a6eed2": {
"balancing:strategy": "weighting",
@@ -553,7 +678,12 @@
"classifier:gradient_boosting:scoring": "loss",
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.04101667890187473,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"aba43bb": {
"balancing:strategy": "none",
@@ -579,7 +709,12 @@
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.011827601847252461,
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"15f5c2d": {
"balancing:strategy": "weighting",
@@ -604,7 +739,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.006279035684154215,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"d3402a5": {
"balancing:strategy": "weighting",
@@ -631,7 +771,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.7029655210876955,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.17384885170243727,
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"3e4ac17": {
"balancing:strategy": "weighting",
@@ -653,7 +798,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.02336976484926326,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 869,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"3e922f0": {
"balancing:strategy": "weighting",
@@ -676,7 +826,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.00550342373297507,
"classifier:gradient_boosting:n_iter_no_change": 4,
"classifier:gradient_boosting:validation_fraction": 0.20573288758098512,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"b2b7a37": {
"balancing:strategy": "weighting",
@@ -696,7 +851,12 @@
"classifier:random_forest:min_samples_split": 12,
"classifier:random_forest:min_weight_fraction_leaf": 0.0,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.0495878468366018,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
}
},
"cutoffs": [
diff --git a/autosklearn/experimental/balanced_accuracy/askl2_portfolios/RF_None_5CV_iterative_es_if.json b/autosklearn/experimental/balanced_accuracy/askl2_portfolios/RF_None_5CV_iterative_es_if.json
index ff0f42a423..a9a4f641d4 100644
--- a/autosklearn/experimental/balanced_accuracy/askl2_portfolios/RF_None_5CV_iterative_es_if.json
+++ b/autosklearn/experimental/balanced_accuracy/askl2_portfolios/RF_None_5CV_iterative_es_if.json
@@ -22,7 +22,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.026511611819517358,
"classifier:gradient_boosting:n_iter_no_change": 11,
"classifier:gradient_boosting:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"153cd48": {
"balancing:strategy": "weighting",
@@ -43,7 +48,12 @@
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.75,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.25870043409765325,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"517b840": {
"balancing:strategy": "weighting",
@@ -61,7 +71,12 @@
"classifier:sgd:penalty": "l1",
"classifier:sgd:tol": 6.822673172007927e-05,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.002283522288122212,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"251a3fa": {
"balancing:strategy": "none",
@@ -85,7 +100,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.75,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.265488755470474,
"classifier:gradient_boosting:n_iter_no_change": 10,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"65b7e1e": {
"balancing:strategy": "weighting",
@@ -104,7 +124,12 @@
"classifier:sgd:tol": 0.007271455729069826,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.0002459755534184245,
"classifier:sgd:eta0": 0.0050360748910538156,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"d814f50": {
"balancing:strategy": "none",
@@ -126,7 +151,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.04720821151303774,
"classifier:gradient_boosting:n_iter_no_change": 18,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"0c1ac11": {
"balancing:strategy": "weighting",
@@ -146,7 +176,12 @@
"classifier:extra_trees:min_samples_split": 8,
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.010386159652388809,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"17bd445": {
"balancing:strategy": "weighting",
@@ -169,7 +204,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.002843139741773352,
"classifier:gradient_boosting:n_iter_no_change": 6,
"classifier:gradient_boosting:validation_fraction": 0.25495619812848924,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"c2a619b": {
"balancing:strategy": "weighting",
@@ -191,7 +231,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1607,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"ddc3d8c": {
"balancing:strategy": "none",
@@ -218,7 +263,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.0029488266576420843,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.7247832818621729,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.025320522643377183,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"8c3de7c": {
"balancing:strategy": "weighting",
@@ -237,7 +287,12 @@
"classifier:extra_trees:min_samples_leaf": 9,
"classifier:extra_trees:min_samples_split": 9,
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"752352a": {
"balancing:strategy": "none",
@@ -257,7 +312,12 @@
"classifier:sgd:eta0": 0.015515899805519508,
"classifier:sgd:l1_ratio": 0.24549411803520174,
"classifier:sgd:power_t": 0.8667397248944955,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"b546b81": {
"balancing:strategy": "weighting",
@@ -280,7 +340,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.12608190257925678,
"classifier:gradient_boosting:n_iter_no_change": 15,
"classifier:gradient_boosting:validation_fraction": 0.061293139245041775,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"c3fe2a6": {
"balancing:strategy": "none",
@@ -300,7 +365,12 @@
"classifier:random_forest:min_samples_split": 6,
"classifier:random_forest:min_weight_fraction_leaf": 0.0,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.014623656915029211,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"8418723": {
"balancing:strategy": "weighting",
@@ -322,7 +392,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"classifier:gradient_boosting:n_iter_no_change": 1,
"classifier:gradient_boosting:validation_fraction": 0.07606225941532962,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"986a1dd": {
"balancing:strategy": "weighting",
@@ -341,7 +416,12 @@
"classifier:sgd:tol": 0.005424831840723169,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.3902406930524133,
"classifier:sgd:eta0": 0.0003678013182724422,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"ec57057": {
"balancing:strategy": "none",
@@ -368,7 +448,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.02479034287675088,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1962,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"54904c9": {
"balancing:strategy": "weighting",
@@ -388,7 +473,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1034,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
"classifier:sgd:l1_ratio": 0.5991767891921137,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"af700cf": {
"balancing:strategy": "weighting",
@@ -404,7 +494,12 @@
"classifier:passive_aggressive:loss": "squared_hinge",
"classifier:passive_aggressive:tol": 0.0038046183268173064,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.00016898186356742833,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"b05c113": {
"balancing:strategy": "weighting",
@@ -424,7 +519,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.017146688690423326,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1411,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"58aee7e": {
"balancing:strategy": "weighting",
@@ -445,7 +545,12 @@
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 920,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"fa44888": {
"balancing:strategy": "weighting",
@@ -465,7 +570,12 @@
"classifier:extra_trees:min_samples_split": 18,
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.1753262636941602,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"f0dcb8c": {
"balancing:strategy": "weighting",
@@ -484,7 +594,12 @@
"classifier:sgd:tol": 1.2296801183022839e-05,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.017661969506602388,
"classifier:sgd:eta0": 0.00021911747851387198,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"cc08355": {
"balancing:strategy": "weighting",
@@ -507,7 +622,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.8867957238309495,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.15743051287173188,
"classifier:gradient_boosting:n_iter_no_change": 19,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"9a87fb9": {
"balancing:strategy": "weighting",
@@ -530,7 +650,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.025494600018089123,
"classifier:gradient_boosting:n_iter_no_change": 1,
"classifier:gradient_boosting:validation_fraction": 0.22017588944163444,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"2c33a1b": {
"balancing:strategy": "weighting",
@@ -550,7 +675,12 @@
"classifier:extra_trees:min_samples_split": 4,
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.0016655553884381083,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"64ee1f9": {
"balancing:strategy": "none",
@@ -576,7 +706,12 @@
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.009442778909715678,
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"e9935d1": {
"balancing:strategy": "weighting",
@@ -596,7 +731,12 @@
"classifier:random_forest:min_samples_split": 3,
"classifier:random_forest:min_weight_fraction_leaf": 0.0,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.1510498275894446,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"1ccb4e6": {
"balancing:strategy": "none",
@@ -621,7 +761,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.05281695491485977,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"70fc8a2": {
"balancing:strategy": "weighting",
@@ -637,7 +782,12 @@
"classifier:passive_aggressive:loss": "hinge",
"classifier:passive_aggressive:tol": 0.0002974739873385178,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.13061993628467652,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"7c8dfce": {
"balancing:strategy": "weighting",
@@ -660,7 +810,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.018048175604161183,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.8687568474210459,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.14957346857371134,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"9aa60c4": {
"balancing:strategy": "weighting",
@@ -681,7 +836,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1287,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
"classifier:sgd:eta0": 1.4478521162990933e-05,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
}
},
"cutoffs": [
diff --git a/autosklearn/experimental/balanced_accuracy/askl2_portfolios/RF_None_holdout_iterative_es_if.json b/autosklearn/experimental/balanced_accuracy/askl2_portfolios/RF_None_holdout_iterative_es_if.json
index ece0700362..2f5a0b4a78 100644
--- a/autosklearn/experimental/balanced_accuracy/askl2_portfolios/RF_None_holdout_iterative_es_if.json
+++ b/autosklearn/experimental/balanced_accuracy/askl2_portfolios/RF_None_holdout_iterative_es_if.json
@@ -19,7 +19,12 @@
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.7586354047062515,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.2909702584195659,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"d38f297": {
"balancing:strategy": "weighting",
@@ -38,7 +43,12 @@
"classifier:sgd:tol": 0.002985016693669331,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.017077961357419263,
"classifier:sgd:l1_ratio": 0.1498449443375854,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"2d4a38e": {
"balancing:strategy": "weighting",
@@ -59,7 +69,12 @@
"classifier:gradient_boosting:scoring": "loss",
"classifier:gradient_boosting:tol": 1e-07,
"classifier:gradient_boosting:n_iter_no_change": 10,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"77764f9": {
"balancing:strategy": "weighting",
@@ -80,7 +95,12 @@
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1555,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"099d657": {
"balancing:strategy": "none",
@@ -104,7 +124,12 @@
"classifier:mlp:shuffle": "True",
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"1684610": {
"balancing:strategy": "weighting",
@@ -127,7 +152,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.0014640920658176558,
"classifier:gradient_boosting:n_iter_no_change": 1,
"classifier:gradient_boosting:validation_fraction": 0.2845052418735526,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"58e7fea": {
"balancing:strategy": "weighting",
@@ -150,7 +180,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.0002326124527892098,
"classifier:gradient_boosting:n_iter_no_change": 3,
"classifier:gradient_boosting:validation_fraction": 0.1810774322857213,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"05690ed": {
"balancing:strategy": "none",
@@ -172,7 +207,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"classifier:gradient_boosting:n_iter_no_change": 1,
"classifier:gradient_boosting:validation_fraction": 0.3886555767510837,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"35e418d": {
"balancing:strategy": "none",
@@ -190,7 +230,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.003855451379285886,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.8617966077274979,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.22153286208031922,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"074da43": {
"balancing:strategy": "weighting",
@@ -210,7 +255,12 @@
"classifier:extra_trees:min_samples_split": 18,
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.005314900629177162,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"7cab112": {
"balancing:strategy": "weighting",
@@ -234,7 +284,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 880,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
"classifier:gradient_boosting:n_iter_no_change": 15,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"c2b918a": {
"balancing:strategy": "weighting",
@@ -258,7 +313,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.7838632555847499,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.25508233724220736,
"classifier:gradient_boosting:n_iter_no_change": 18,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"811cba1": {
"balancing:strategy": "weighting",
@@ -279,7 +339,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1840,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
"classifier:sgd:l1_ratio": 0.38121990909845516,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"05327ff": {
"balancing:strategy": "weighting",
@@ -304,7 +369,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
"classifier:gradient_boosting:n_iter_no_change": 7,
"classifier:gradient_boosting:validation_fraction": 0.07231375899575396,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"32974d1": {
"balancing:strategy": "none",
@@ -323,7 +393,12 @@
"classifier:sgd:tol": 1.5416139747247184e-05,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1249,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"a3ba205": {
"balancing:strategy": "weighting",
@@ -341,7 +416,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.00019794194843467043,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 498,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"864161c": {
"balancing:strategy": "weighting",
@@ -362,7 +442,12 @@
"classifier:random_forest:min_weight_fraction_leaf": 0.0,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.7101136721559372,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.2189336903315928,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"d033742": {
"balancing:strategy": "weighting",
@@ -387,7 +472,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.05231190906849011,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"eb6f7f1": {
"balancing:strategy": "weighting",
@@ -410,7 +500,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.015284619454999112,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.744931703493537,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.25686097824357923,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"c29f8bf": {
"balancing:strategy": "weighting",
@@ -432,7 +527,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"classifier:gradient_boosting:n_iter_no_change": 16,
"classifier:gradient_boosting:validation_fraction": 0.2596215430064656,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"c832fa9": {
"balancing:strategy": "weighting",
@@ -459,7 +559,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.045160438360813265,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.7646233641739945,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.27760286286692604,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"5d877bc": {
"balancing:strategy": "none",
@@ -483,7 +588,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.20810753982499872,
"classifier:gradient_boosting:n_iter_no_change": 4,
"classifier:gradient_boosting:validation_fraction": 0.24990038505909143,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"c7158d0": {
"balancing:strategy": "weighting",
@@ -508,7 +618,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.058558914473045606,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"d618703": {
"balancing:strategy": "weighting",
@@ -534,7 +649,12 @@
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.007482244990304125,
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"82a2502": {
"balancing:strategy": "weighting",
@@ -556,7 +676,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.011016084673362935,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.7724545671283145,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.2792304829849116,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"8e1bfe6": {
"balancing:strategy": "none",
@@ -578,7 +703,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.02643262139119834,
"classifier:gradient_boosting:n_iter_no_change": 12,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"4ad9cee": {
"balancing:strategy": "weighting",
@@ -600,7 +730,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.04411723614126245,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1679,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"2f6c8d6": {
"balancing:strategy": "none",
@@ -619,7 +754,12 @@
"classifier:sgd:tol": 0.0029940704007657176,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.009786276322275262,
"classifier:sgd:epsilon": 0.0008711023589705295,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"6be5932": {
"balancing:strategy": "none",
@@ -641,7 +781,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.00017275410211396532,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.7322313773989851,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.299087564302405,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"c5628ce": {
"balancing:strategy": "weighting",
@@ -668,7 +813,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.013193967050327956,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1291,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"564103f": {
"balancing:strategy": "weighting",
@@ -693,7 +843,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
"classifier:gradient_boosting:n_iter_no_change": 4,
"classifier:gradient_boosting:validation_fraction": 0.2551421881660974,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"9ce49a5": {
"balancing:strategy": "none",
@@ -714,7 +869,12 @@
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.8700451777827737,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.24811629410747338,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
}
},
"cutoffs": [
diff --git a/autosklearn/experimental/balanced_accuracy/askl2_portfolios/RF_SH-eta4-i_10CV_iterative_es_if.json b/autosklearn/experimental/balanced_accuracy/askl2_portfolios/RF_SH-eta4-i_10CV_iterative_es_if.json
index 652659bda0..59c8b2a50b 100644
--- a/autosklearn/experimental/balanced_accuracy/askl2_portfolios/RF_SH-eta4-i_10CV_iterative_es_if.json
+++ b/autosklearn/experimental/balanced_accuracy/askl2_portfolios/RF_SH-eta4-i_10CV_iterative_es_if.json
@@ -20,7 +20,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.11840768063684459,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.7668727846319956,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.28633349214759163,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"f1500cd": {
"balancing:strategy": "weighting",
@@ -42,7 +47,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.0013138596672908974,
"classifier:gradient_boosting:n_iter_no_change": 13,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"3e1a332": {
"balancing:strategy": "weighting",
@@ -62,7 +72,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.006544845548292656,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1062,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"7500838": {
"balancing:strategy": "weighting",
@@ -85,7 +100,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 607,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
"classifier:gradient_boosting:n_iter_no_change": 11,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"dd3a0e8": {
"balancing:strategy": "weighting",
@@ -106,7 +126,12 @@
"classifier:gradient_boosting:scoring": "loss",
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.002589355496626201,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"13d94b3": {
"balancing:strategy": "weighting",
@@ -132,7 +157,12 @@
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.05217337165303994,
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"78a0240": {
"balancing:strategy": "weighting",
@@ -152,7 +182,12 @@
"classifier:extra_trees:min_samples_split": 5,
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.0036912148544896695,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"4dffb08": {
"balancing:strategy": "weighting",
@@ -172,7 +207,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.0002547219008659275,
"classifier:sgd:eta0": 2.0514805804594057e-06,
"classifier:sgd:l1_ratio": 7.163773086996508e-09,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"43a743f": {
"balancing:strategy": "none",
@@ -195,7 +235,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 82,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
"classifier:gradient_boosting:n_iter_no_change": 3,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"4d5e802": {
"balancing:strategy": "weighting",
@@ -217,7 +262,12 @@
"classifier:sgd:epsilon": 0.009309372110873482,
"classifier:sgd:eta0": 9.309855728417528e-05,
"classifier:sgd:l1_ratio": 0.6620356750678396,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"9b6c297": {
"balancing:strategy": "weighting",
@@ -239,7 +289,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.7972710156738574,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.2243563424345074,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"5886349": {
"balancing:strategy": "weighting",
@@ -261,7 +316,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"classifier:gradient_boosting:n_iter_no_change": 1,
"classifier:gradient_boosting:validation_fraction": 0.16202792486532844,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"5c6b775": {
"balancing:strategy": "weighting",
@@ -285,7 +345,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.2395023772591513,
"classifier:gradient_boosting:n_iter_no_change": 5,
"classifier:gradient_boosting:validation_fraction": 0.10823188437286345,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"87cff92": {
"balancing:strategy": "none",
@@ -307,7 +372,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.027146552434145918,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.7300216362356305,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.23992030699419034,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"05cb256": {
"balancing:strategy": "none",
@@ -335,7 +405,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1822,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"4e50c99": {
"balancing:strategy": "weighting",
@@ -357,7 +432,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.0020131193326745865,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.7478194148240691,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.09486350397967545,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"73611bf": {
"balancing:strategy": "weighting",
@@ -379,7 +459,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.047507672861535975,
"classifier:gradient_boosting:n_iter_no_change": 11,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"a885804": {
"balancing:strategy": "weighting",
@@ -399,7 +484,12 @@
"classifier:extra_trees:min_samples_split": 13,
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.0016446890106470516,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"44e2072": {
"balancing:strategy": "weighting",
@@ -419,7 +509,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.0017922018692146033,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1528,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"4b79861": {
"balancing:strategy": "none",
@@ -441,7 +536,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.0001253687001786426,
"classifier:gradient_boosting:n_iter_no_change": 4,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"79812c6": {
"balancing:strategy": "weighting",
@@ -463,7 +563,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.15443993532080638,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 999,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"8e0e269": {
"balancing:strategy": "weighting",
@@ -485,7 +590,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"classifier:gradient_boosting:n_iter_no_change": 3,
"classifier:gradient_boosting:validation_fraction": 0.14053446064492747,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"66c3cf8": {
"balancing:strategy": "weighting",
@@ -505,7 +615,12 @@
"classifier:extra_trees:min_samples_split": 13,
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.00017390212311709855,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"bff26f9": {
"balancing:strategy": "weighting",
@@ -523,7 +638,12 @@
"classifier:sgd:penalty": "l2",
"classifier:sgd:tol": 0.0005655131678222187,
"classifier:sgd:eta0": 0.00022254320651933049,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"ce012ed": {
"balancing:strategy": "weighting",
@@ -538,7 +658,12 @@
"classifier:passive_aggressive:fit_intercept": "True",
"classifier:passive_aggressive:loss": "squared_hinge",
"classifier:passive_aggressive:tol": 0.009455868366528827,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"d2d2039": {
"balancing:strategy": "weighting",
@@ -560,7 +685,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"classifier:gradient_boosting:n_iter_no_change": 19,
"classifier:gradient_boosting:validation_fraction": 0.29656195578950684,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"eac6b8e": {
"balancing:strategy": "weighting",
@@ -580,7 +710,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.001003862016677866,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1987,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"a31129f": {
"balancing:strategy": "none",
@@ -605,7 +740,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.014794967781387955,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"da90184": {
"balancing:strategy": "weighting",
@@ -626,7 +766,12 @@
"classifier:gradient_boosting:scoring": "loss",
"classifier:gradient_boosting:tol": 1e-07,
"classifier:gradient_boosting:n_iter_no_change": 11,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"85bfdfa": {
"balancing:strategy": "weighting",
@@ -646,7 +791,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.011947142763501097,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1194,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"d4abc34": {
"balancing:strategy": "weighting",
@@ -672,7 +822,12 @@
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.748076585259295,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.23337355967199702,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"acb3733": {
"balancing:strategy": "weighting",
@@ -695,7 +850,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.0017046848576384626,
"classifier:gradient_boosting:n_iter_no_change": 1,
"classifier:gradient_boosting:validation_fraction": 0.2938721928434415,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
}
},
"cutoffs": [
diff --git a/autosklearn/experimental/balanced_accuracy/askl2_portfolios/RF_SH-eta4-i_3CV_iterative_es_if.json b/autosklearn/experimental/balanced_accuracy/askl2_portfolios/RF_SH-eta4-i_3CV_iterative_es_if.json
index 08a5a03a9b..8a7e0e8263 100644
--- a/autosklearn/experimental/balanced_accuracy/askl2_portfolios/RF_SH-eta4-i_3CV_iterative_es_if.json
+++ b/autosklearn/experimental/balanced_accuracy/askl2_portfolios/RF_SH-eta4-i_3CV_iterative_es_if.json
@@ -19,7 +19,12 @@
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1025,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"08d784a": {
"balancing:strategy": "weighting",
@@ -42,7 +47,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1719,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
"classifier:gradient_boosting:n_iter_no_change": 1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"d91f995": {
"balancing:strategy": "weighting",
@@ -62,7 +72,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.01665326684623352,
"classifier:sgd:eta0": 0.0008683421715267357,
"classifier:sgd:l1_ratio": 0.0037742853051687883,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"9b1161c": {
"balancing:strategy": "weighting",
@@ -83,7 +98,12 @@
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.9371966597485295,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.2181009648699602,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"e6b6325": {
"balancing:strategy": "weighting",
@@ -105,7 +125,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.010402618316459437,
"classifier:gradient_boosting:n_iter_no_change": 17,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"53634b7": {
"balancing:strategy": "none",
@@ -129,7 +154,12 @@
"classifier:mlp:shuffle": "True",
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"ba079f7": {
"balancing:strategy": "weighting",
@@ -152,7 +182,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.010000000000000004,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.7279628505501666,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.24610303193556213,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"eb4397b": {
"balancing:strategy": "weighting",
@@ -173,7 +208,12 @@
"classifier:gradient_boosting:scoring": "loss",
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.22276171190430583,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"8d8ad31": {
"balancing:strategy": "weighting",
@@ -194,7 +234,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1985,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
"classifier:sgd:l1_ratio": 0.0752197101694388,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"56a06e6": {
"balancing:strategy": "weighting",
@@ -213,7 +258,12 @@
"classifier:extra_trees:min_samples_leaf": 1,
"classifier:extra_trees:min_samples_split": 9,
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"1a2077f": {
"balancing:strategy": "weighting",
@@ -233,7 +283,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.00025247448315347646,
"classifier:sgd:eta0": 0.004594829994128768,
"classifier:sgd:power_t": 0.5032138663090429,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"762f55d": {
"balancing:strategy": "weighting",
@@ -255,7 +310,12 @@
"classifier:sgd:epsilon": 0.0060184515243734615,
"classifier:sgd:eta0": 8.74381130399994e-05,
"classifier:sgd:power_t": 0.040470778304526875,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"3eada54": {
"balancing:strategy": "weighting",
@@ -270,7 +330,12 @@
"classifier:passive_aggressive:fit_intercept": "True",
"classifier:passive_aggressive:loss": "hinge",
"classifier:passive_aggressive:tol": 0.004257493012053348,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"42ab0ad": {
"balancing:strategy": "weighting",
@@ -292,7 +357,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 839,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"6bdd0f7": {
"balancing:strategy": "weighting",
@@ -312,7 +382,12 @@
"classifier:sgd:epsilon": 0.0042774156243000246,
"classifier:sgd:eta0": 0.007202966496306807,
"classifier:sgd:power_t": 0.7972709828085118,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"aba43bb": {
"balancing:strategy": "none",
@@ -338,7 +413,12 @@
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.011827601847252461,
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"fcd1bda": {
"balancing:strategy": "none",
@@ -365,7 +445,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.01186630902011325,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1281,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"418cb3d": {
"balancing:strategy": "weighting",
@@ -387,7 +472,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.11023361689757208,
"classifier:gradient_boosting:n_iter_no_change": 1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"6f053cf": {
"balancing:strategy": "none",
@@ -412,7 +502,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.012107086584047114,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"c138595": {
"balancing:strategy": "weighting",
@@ -435,7 +530,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.03127012992564812,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.864877718024256,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.1412101268042483,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"237eb9d": {
"balancing:strategy": "weighting",
@@ -457,7 +557,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.06393243974085384,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.9024015752738876,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.014225351110017392,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"1a6eed2": {
"balancing:strategy": "weighting",
@@ -478,7 +583,12 @@
"classifier:gradient_boosting:scoring": "loss",
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.04101667890187473,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"fda1986": {
"balancing:strategy": "none",
@@ -501,7 +611,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.06676574893796923,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.7580247800718314,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.24565571245859932,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"2036c61": {
"balancing:strategy": "weighting",
@@ -526,7 +641,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
"classifier:gradient_boosting:n_iter_no_change": 10,
"classifier:gradient_boosting:validation_fraction": 0.2850324540065884,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"c8cc243": {
"balancing:strategy": "weighting",
@@ -546,7 +666,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.02758033528296938,
"classifier:sgd:epsilon": 0.0001239933011012703,
"classifier:sgd:eta0": 4.9354139509949256e-05,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"61b0988": {
"balancing:strategy": "none",
@@ -563,7 +688,12 @@
"classifier:passive_aggressive:tol": 2.8118963978316873e-05,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1742,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"2190cfc": {
"balancing:strategy": "weighting",
@@ -583,7 +713,12 @@
"classifier:extra_trees:min_samples_split": 7,
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.0015359113788476297,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"15debe5": {
"balancing:strategy": "weighting",
@@ -608,7 +743,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"45b9541": {
"balancing:strategy": "none",
@@ -634,7 +774,12 @@
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.006209804535789879,
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"48cd11d": {
"balancing:strategy": "none",
@@ -658,7 +803,12 @@
"classifier:mlp:shuffle": "True",
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"95e43e6": {
"balancing:strategy": "none",
@@ -685,7 +835,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.018226002574401903,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1093,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"f8aba41": {
"balancing:strategy": "weighting",
@@ -700,7 +855,12 @@
"classifier:passive_aggressive:fit_intercept": "True",
"classifier:passive_aggressive:loss": "hinge",
"classifier:passive_aggressive:tol": 0.010384183550709794,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
}
},
"cutoffs": [
diff --git a/autosklearn/experimental/balanced_accuracy/askl2_portfolios/RF_SH-eta4-i_5CV_iterative_es_if.json b/autosklearn/experimental/balanced_accuracy/askl2_portfolios/RF_SH-eta4-i_5CV_iterative_es_if.json
index 556c888df0..3f45fe26e9 100644
--- a/autosklearn/experimental/balanced_accuracy/askl2_portfolios/RF_SH-eta4-i_5CV_iterative_es_if.json
+++ b/autosklearn/experimental/balanced_accuracy/askl2_portfolios/RF_SH-eta4-i_5CV_iterative_es_if.json
@@ -22,7 +22,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.026511611819517358,
"classifier:gradient_boosting:n_iter_no_change": 11,
"classifier:gradient_boosting:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"153cd48": {
"balancing:strategy": "weighting",
@@ -43,7 +48,12 @@
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.75,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.25870043409765325,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"65b7e1e": {
"balancing:strategy": "weighting",
@@ -62,7 +72,12 @@
"classifier:sgd:tol": 0.007271455729069826,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.0002459755534184245,
"classifier:sgd:eta0": 0.0050360748910538156,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"7c8dfce": {
"balancing:strategy": "weighting",
@@ -85,7 +100,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.018048175604161183,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.8687568474210459,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.14957346857371134,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"64ee1f9": {
"balancing:strategy": "none",
@@ -111,7 +131,12 @@
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.009442778909715678,
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"8ceec1c": {
"balancing:strategy": "weighting",
@@ -132,7 +157,12 @@
"classifier:gradient_boosting:scoring": "loss",
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.03734669642625855,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"97dbfc8": {
"balancing:strategy": "none",
@@ -157,7 +187,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.011240853770638584,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"aa37c31": {
"balancing:strategy": "weighting",
@@ -178,7 +213,12 @@
"classifier:gradient_boosting:scoring": "loss",
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.000128792623655875,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"0c1ac11": {
"balancing:strategy": "weighting",
@@ -198,7 +238,12 @@
"classifier:extra_trees:min_samples_split": 8,
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.010386159652388809,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"8418723": {
"balancing:strategy": "weighting",
@@ -220,7 +265,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"classifier:gradient_boosting:n_iter_no_change": 1,
"classifier:gradient_boosting:validation_fraction": 0.07606225941532962,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"752352a": {
"balancing:strategy": "none",
@@ -240,7 +290,12 @@
"classifier:sgd:eta0": 0.015515899805519508,
"classifier:sgd:l1_ratio": 0.24549411803520174,
"classifier:sgd:power_t": 0.8667397248944955,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"b546b81": {
"balancing:strategy": "weighting",
@@ -263,7 +318,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.12608190257925678,
"classifier:gradient_boosting:n_iter_no_change": 15,
"classifier:gradient_boosting:validation_fraction": 0.061293139245041775,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"17bd445": {
"balancing:strategy": "weighting",
@@ -286,7 +346,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.002843139741773352,
"classifier:gradient_boosting:n_iter_no_change": 6,
"classifier:gradient_boosting:validation_fraction": 0.25495619812848924,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"c3fe2a6": {
"balancing:strategy": "none",
@@ -306,7 +371,12 @@
"classifier:random_forest:min_samples_split": 6,
"classifier:random_forest:min_weight_fraction_leaf": 0.0,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.014623656915029211,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"58aee7e": {
"balancing:strategy": "weighting",
@@ -327,7 +397,12 @@
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 920,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"4d1f885": {
"balancing:strategy": "weighting",
@@ -348,7 +423,12 @@
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.7945158247012963,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.03711649896366117,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"d814f50": {
"balancing:strategy": "none",
@@ -370,7 +450,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.04720821151303774,
"classifier:gradient_boosting:n_iter_no_change": 18,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"b05c113": {
"balancing:strategy": "weighting",
@@ -390,7 +475,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.017146688690423326,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1411,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"f0dcb8c": {
"balancing:strategy": "weighting",
@@ -409,7 +499,12 @@
"classifier:sgd:tol": 1.2296801183022839e-05,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.017661969506602388,
"classifier:sgd:eta0": 0.00021911747851387198,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"9b48748": {
"balancing:strategy": "none",
@@ -427,7 +522,12 @@
"classifier:sgd:penalty": "elasticnet",
"classifier:sgd:tol": 2.4009915484750738e-05,
"classifier:sgd:l1_ratio": 4.444732732155884e-08,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"986a1dd": {
"balancing:strategy": "weighting",
@@ -446,7 +546,12 @@
"classifier:sgd:tol": 0.005424831840723169,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.3902406930524133,
"classifier:sgd:eta0": 0.0003678013182724422,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"9a87fb9": {
"balancing:strategy": "weighting",
@@ -469,7 +574,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.025494600018089123,
"classifier:gradient_boosting:n_iter_no_change": 1,
"classifier:gradient_boosting:validation_fraction": 0.22017588944163444,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"fa44888": {
"balancing:strategy": "weighting",
@@ -489,7 +599,12 @@
"classifier:extra_trees:min_samples_split": 18,
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.1753262636941602,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"7ed7866": {
"balancing:strategy": "weighting",
@@ -512,7 +627,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.02263817900937648,
"classifier:gradient_boosting:n_iter_no_change": 1,
"classifier:gradient_boosting:validation_fraction": 0.16830986246481522,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"517b840": {
"balancing:strategy": "weighting",
@@ -530,7 +650,12 @@
"classifier:sgd:penalty": "l1",
"classifier:sgd:tol": 6.822673172007927e-05,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.002283522288122212,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"8ee7265": {
"balancing:strategy": "weighting",
@@ -550,7 +675,12 @@
"classifier:extra_trees:min_samples_split": 8,
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.0031964981754476628,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"67da9ba": {
"balancing:strategy": "weighting",
@@ -570,7 +700,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.3222595116142074,
"classifier:sgd:eta0": 0.0014873014786985059,
"classifier:sgd:l1_ratio": 0.4083537005226397,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"e9935d1": {
"balancing:strategy": "weighting",
@@ -590,7 +725,12 @@
"classifier:random_forest:min_samples_split": 3,
"classifier:random_forest:min_weight_fraction_leaf": 0.0,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.1510498275894446,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"70fc8a2": {
"balancing:strategy": "weighting",
@@ -606,7 +746,12 @@
"classifier:passive_aggressive:loss": "hinge",
"classifier:passive_aggressive:tol": 0.0002974739873385178,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.13061993628467652,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"6bc51e2": {
"balancing:strategy": "weighting",
@@ -633,7 +778,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.0017163643283156385,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1636,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"ddebe02": {
"balancing:strategy": "weighting",
@@ -654,7 +804,12 @@
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 781,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"af700cf": {
"balancing:strategy": "weighting",
@@ -670,7 +825,12 @@
"classifier:passive_aggressive:loss": "squared_hinge",
"classifier:passive_aggressive:tol": 0.0038046183268173064,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.00016898186356742833,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
}
},
"cutoffs": [
diff --git a/autosklearn/experimental/balanced_accuracy/askl2_portfolios/RF_SH-eta4-i_holdout_iterative_es_if.json b/autosklearn/experimental/balanced_accuracy/askl2_portfolios/RF_SH-eta4-i_holdout_iterative_es_if.json
index 5d528fc589..93ca58b6b5 100644
--- a/autosklearn/experimental/balanced_accuracy/askl2_portfolios/RF_SH-eta4-i_holdout_iterative_es_if.json
+++ b/autosklearn/experimental/balanced_accuracy/askl2_portfolios/RF_SH-eta4-i_holdout_iterative_es_if.json
@@ -19,7 +19,12 @@
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.7586354047062515,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.2909702584195659,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"eb6f7f1": {
"balancing:strategy": "weighting",
@@ -42,7 +47,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.015284619454999112,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.744931703493537,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.25686097824357923,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"d033742": {
"balancing:strategy": "weighting",
@@ -67,7 +77,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.05231190906849011,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"d38f297": {
"balancing:strategy": "weighting",
@@ -86,7 +101,12 @@
"classifier:sgd:tol": 0.002985016693669331,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.017077961357419263,
"classifier:sgd:l1_ratio": 0.1498449443375854,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"e249036": {
"balancing:strategy": "weighting",
@@ -109,7 +129,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.9922123945957716,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.05865685179492095,
"classifier:gradient_boosting:n_iter_no_change": 13,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"978b964": {
"balancing:strategy": "weighting",
@@ -128,7 +153,12 @@
"classifier:extra_trees:min_samples_leaf": 3,
"classifier:extra_trees:min_samples_split": 16,
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"07bebab": {
"balancing:strategy": "weighting",
@@ -151,7 +181,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.02508567646126689,
"classifier:gradient_boosting:n_iter_no_change": 13,
"classifier:gradient_boosting:validation_fraction": 0.08774268414289245,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"05690ed": {
"balancing:strategy": "none",
@@ -173,7 +208,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"classifier:gradient_boosting:n_iter_no_change": 1,
"classifier:gradient_boosting:validation_fraction": 0.3886555767510837,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"8cd6304": {
"balancing:strategy": "none",
@@ -190,7 +230,12 @@
"classifier:passive_aggressive:tol": 1.260088833717425e-05,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.8506611891906092,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.03562927538205138,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"73fc45b": {
"balancing:strategy": "weighting",
@@ -211,7 +256,12 @@
"classifier:gradient_boosting:scoring": "loss",
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.012073097426292019,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"074da43": {
"balancing:strategy": "weighting",
@@ -231,7 +281,12 @@
"classifier:extra_trees:min_samples_split": 18,
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.005314900629177162,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"c832fa9": {
"balancing:strategy": "weighting",
@@ -258,7 +313,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.045160438360813265,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.7646233641739945,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.27760286286692604,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"a3ba205": {
"balancing:strategy": "weighting",
@@ -276,7 +336,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.00019794194843467043,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 498,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"68e6606": {
"balancing:strategy": "weighting",
@@ -295,7 +360,12 @@
"classifier:extra_trees:min_samples_leaf": 11,
"classifier:extra_trees:min_samples_split": 11,
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"f187687": {
"balancing:strategy": "weighting",
@@ -316,7 +386,12 @@
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.7136760024277188,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.2452308844548974,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"a479a3d": {
"balancing:strategy": "none",
@@ -339,7 +414,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.20607286607457198,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1739,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"c2b918a": {
"balancing:strategy": "weighting",
@@ -363,7 +443,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.7838632555847499,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.25508233724220736,
"classifier:gradient_boosting:n_iter_no_change": 18,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"8c75f1c": {
"balancing:strategy": "weighting",
@@ -378,7 +463,12 @@
"classifier:passive_aggressive:fit_intercept": "True",
"classifier:passive_aggressive:loss": "hinge",
"classifier:passive_aggressive:tol": 0.028504951505057607,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"77764f9": {
"balancing:strategy": "weighting",
@@ -399,7 +489,12 @@
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1555,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"05327ff": {
"balancing:strategy": "weighting",
@@ -424,7 +519,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
"classifier:gradient_boosting:n_iter_no_change": 7,
"classifier:gradient_boosting:validation_fraction": 0.07231375899575396,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"2d4a38e": {
"balancing:strategy": "weighting",
@@ -445,7 +545,12 @@
"classifier:gradient_boosting:scoring": "loss",
"classifier:gradient_boosting:tol": 1e-07,
"classifier:gradient_boosting:n_iter_no_change": 10,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"d618703": {
"balancing:strategy": "weighting",
@@ -471,7 +576,12 @@
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.007482244990304125,
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"c437bea": {
"balancing:strategy": "none",
@@ -492,7 +602,12 @@
"classifier:random_forest:min_weight_fraction_leaf": 0.0,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.7450275104929482,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.1376588753599898,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"2f6c8d6": {
"balancing:strategy": "none",
@@ -511,7 +626,12 @@
"classifier:sgd:tol": 0.0029940704007657176,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.009786276322275262,
"classifier:sgd:epsilon": 0.0008711023589705295,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"811cba1": {
"balancing:strategy": "weighting",
@@ -532,7 +652,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1840,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
"classifier:sgd:l1_ratio": 0.38121990909845516,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"c29f8bf": {
"balancing:strategy": "weighting",
@@ -554,7 +679,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"classifier:gradient_boosting:n_iter_no_change": 16,
"classifier:gradient_boosting:validation_fraction": 0.2596215430064656,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"d54bc21": {
"balancing:strategy": "weighting",
@@ -581,7 +711,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.010000000000000004,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.7334232077430226,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.242115842769868,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"b8184c0": {
"balancing:strategy": "weighting",
@@ -601,7 +736,12 @@
"classifier:extra_trees:min_samples_split": 13,
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.0021020326930237176,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"7cab112": {
"balancing:strategy": "weighting",
@@ -625,7 +765,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 880,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
"classifier:gradient_boosting:n_iter_no_change": 15,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"b86e0b2": {
"balancing:strategy": "weighting",
@@ -641,7 +786,12 @@
"classifier:passive_aggressive:loss": "hinge",
"classifier:passive_aggressive:tol": 0.0007201456859918386,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.00011374720541793248,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"a2734c9": {
"balancing:strategy": "weighting",
@@ -665,7 +815,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
"classifier:gradient_boosting:n_iter_no_change": 15,
"classifier:gradient_boosting:validation_fraction": 0.373905213665471,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"564103f": {
"balancing:strategy": "weighting",
@@ -690,7 +845,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
"classifier:gradient_boosting:n_iter_no_change": 4,
"classifier:gradient_boosting:validation_fraction": 0.2551421881660974,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
}
},
"cutoffs": [
diff --git a/autosklearn/experimental/log_loss/askl2_portfolios/RF_None_10CV_iterative_es_if.json b/autosklearn/experimental/log_loss/askl2_portfolios/RF_None_10CV_iterative_es_if.json
index e99742213f..769ec64959 100644
--- a/autosklearn/experimental/log_loss/askl2_portfolios/RF_None_10CV_iterative_es_if.json
+++ b/autosklearn/experimental/log_loss/askl2_portfolios/RF_None_10CV_iterative_es_if.json
@@ -21,7 +21,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.006280403042161129,
"classifier:gradient_boosting:n_iter_no_change": 17,
"classifier:gradient_boosting:validation_fraction": 0.12820455722720236,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"69a5aed": {
"balancing:strategy": "weighting",
@@ -46,7 +51,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.020507323983341024,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"278a3b7": {
"balancing:strategy": "weighting",
@@ -64,7 +74,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.043140728717176134,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.7565211582936167,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.2427714099781957,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"2b21643": {
"balancing:strategy": "weighting",
@@ -90,7 +105,12 @@
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.010000000000000004,
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"186f4ea": {
"balancing:strategy": "none",
@@ -113,7 +133,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.03190939434652121,
"classifier:gradient_boosting:n_iter_no_change": 19,
"classifier:gradient_boosting:validation_fraction": 0.08286587058642826,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"bad65bc": {
"balancing:strategy": "none",
@@ -135,7 +160,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.003937421024884423,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.72347247098442,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.08078070293132113,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"f4ccd76": {
"balancing:strategy": "weighting",
@@ -157,7 +187,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"classifier:gradient_boosting:n_iter_no_change": 15,
"classifier:gradient_boosting:validation_fraction": 0.12630183388364327,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"eaafcc8": {
"balancing:strategy": "weighting",
@@ -183,7 +218,12 @@
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.29539064866700554,
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"5d883d6": {
"balancing:strategy": "none",
@@ -209,7 +249,12 @@
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 935,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"bacc15e": {
"balancing:strategy": "weighting",
@@ -236,7 +281,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.016645612062423754,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.9727250293787837,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.250997962904354,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"6c9a546": {
"balancing:strategy": "none",
@@ -262,7 +312,12 @@
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.013890722403949634,
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"fe1cf61": {
"balancing:strategy": "weighting",
@@ -285,7 +340,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.00011785529033562006,
"classifier:gradient_boosting:n_iter_no_change": 13,
"classifier:gradient_boosting:validation_fraction": 0.09015136537905166,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"231dfac": {
"balancing:strategy": "none",
@@ -308,7 +368,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 853,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
"classifier:gradient_boosting:n_iter_no_change": 2,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"2d37001": {
"balancing:strategy": "weighting",
@@ -333,7 +398,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"c11fce2": {
"balancing:strategy": "weighting",
@@ -352,7 +422,12 @@
"classifier:sgd:tol": 0.0012132021248816036,
"classifier:sgd:eta0": 0.00030412915402748054,
"classifier:sgd:power_t": 0.3300398436994113,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"040bd83": {
"balancing:strategy": "none",
@@ -379,7 +454,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.0011172717285354973,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 848,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"0ed94fb": {
"balancing:strategy": "weighting",
@@ -405,7 +485,12 @@
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.19791986161487668,
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"72903c0": {
"balancing:strategy": "none",
@@ -425,7 +510,12 @@
"classifier:gradient_boosting:min_samples_leaf": 154,
"classifier:gradient_boosting:scoring": "loss",
"classifier:gradient_boosting:tol": 1e-07,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"a82be94": {
"balancing:strategy": "none",
@@ -446,7 +536,12 @@
"classifier:gradient_boosting:scoring": "loss",
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.006309920762276479,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"395da55": {
"balancing:strategy": "weighting",
@@ -473,7 +568,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.7076847567891946,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.005025836888703367,
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"121a4ef": {
"balancing:strategy": "none",
@@ -494,7 +594,12 @@
"classifier:gradient_boosting:scoring": "loss",
"classifier:gradient_boosting:tol": 1e-07,
"classifier:gradient_boosting:n_iter_no_change": 16,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"e403bce": {
"balancing:strategy": "weighting",
@@ -516,7 +621,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"classifier:gradient_boosting:n_iter_no_change": 15,
"classifier:gradient_boosting:validation_fraction": 0.038609281162211954,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"e1d88ff": {
"balancing:strategy": "weighting",
@@ -537,7 +647,12 @@
"classifier:gradient_boosting:scoring": "loss",
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.0017745595112219766,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"3b1767a": {
"balancing:strategy": "none",
@@ -559,7 +674,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.00022212882394186624,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 877,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"7e1cc2b": {
"balancing:strategy": "none",
@@ -579,7 +699,12 @@
"classifier:gradient_boosting:min_samples_leaf": 151,
"classifier:gradient_boosting:scoring": "loss",
"classifier:gradient_boosting:tol": 1e-07,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"5ac8b4c": {
"balancing:strategy": "weighting",
@@ -604,7 +729,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.009080589517127771,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"f3b27fb": {
"balancing:strategy": "none",
@@ -629,7 +759,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
"classifier:gradient_boosting:n_iter_no_change": 17,
"classifier:gradient_boosting:validation_fraction": 0.05025878941909691,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"d70e605": {
"balancing:strategy": "none",
@@ -654,7 +789,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.14157483812028196,
"classifier:gradient_boosting:n_iter_no_change": 7,
"classifier:gradient_boosting:validation_fraction": 0.044378278898149826,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"db042ae": {
"balancing:strategy": "none",
@@ -679,7 +819,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.003132312313645055,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"46148ef": {
"balancing:strategy": "none",
@@ -703,7 +848,12 @@
"classifier:mlp:shuffle": "True",
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"b96cfb3": {
"balancing:strategy": "none",
@@ -728,7 +878,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.0032925248120968527,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"a69fc2a": {
"balancing:strategy": "none",
@@ -748,7 +903,12 @@
"classifier:gradient_boosting:min_samples_leaf": 2,
"classifier:gradient_boosting:scoring": "loss",
"classifier:gradient_boosting:tol": 1e-07,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
}
},
"cutoffs": [
diff --git a/autosklearn/experimental/log_loss/askl2_portfolios/RF_None_3CV_iterative_es_if.json b/autosklearn/experimental/log_loss/askl2_portfolios/RF_None_3CV_iterative_es_if.json
index 1bf572ad28..0564fd890a 100644
--- a/autosklearn/experimental/log_loss/askl2_portfolios/RF_None_3CV_iterative_es_if.json
+++ b/autosklearn/experimental/log_loss/askl2_portfolios/RF_None_3CV_iterative_es_if.json
@@ -21,7 +21,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.0004981969019895562,
"classifier:gradient_boosting:n_iter_no_change": 4,
"classifier:gradient_boosting:validation_fraction": 0.23374473517320618,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"a4cc318": {
"balancing:strategy": "weighting",
@@ -45,7 +50,12 @@
"classifier:mlp:shuffle": "True",
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"5d80c27": {
"balancing:strategy": "weighting",
@@ -61,7 +71,12 @@
"classifier:passive_aggressive:loss": "squared_hinge",
"classifier:passive_aggressive:tol": 0.0005787736999064962,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.012192559003185645,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"772b3a8": {
"balancing:strategy": "none",
@@ -87,7 +102,12 @@
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1000,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"aa1d8b0": {
"balancing:strategy": "none",
@@ -109,7 +129,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.00016137881434812206,
"classifier:gradient_boosting:n_iter_no_change": 14,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"9650698": {
"balancing:strategy": "weighting",
@@ -130,7 +155,12 @@
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.8072961431686058,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.06513185741134703,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"f203d83": {
"balancing:strategy": "weighting",
@@ -150,7 +180,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.011532350506496297,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.7645599467770883,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.0011091488435678166,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"128ee5e": {
"balancing:strategy": "weighting",
@@ -172,7 +207,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.1696500193573621,
"classifier:gradient_boosting:n_iter_no_change": 20,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"6859bdd": {
"balancing:strategy": "none",
@@ -192,7 +232,12 @@
"classifier:gradient_boosting:min_samples_leaf": 1,
"classifier:gradient_boosting:scoring": "loss",
"classifier:gradient_boosting:tol": 1e-07,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"7cd518a": {
"balancing:strategy": "none",
@@ -212,7 +257,12 @@
"classifier:gradient_boosting:min_samples_leaf": 50,
"classifier:gradient_boosting:scoring": "loss",
"classifier:gradient_boosting:tol": 1e-07,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"94dbc0f": {
"balancing:strategy": "weighting",
@@ -237,7 +287,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.03612642945550235,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"e623b07": {
"balancing:strategy": "none",
@@ -261,7 +316,12 @@
"classifier:mlp:shuffle": "True",
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"05b30ee": {
"balancing:strategy": "weighting",
@@ -286,7 +346,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.25794352734499837,
"classifier:gradient_boosting:n_iter_no_change": 20,
"classifier:gradient_boosting:validation_fraction": 0.06937115181143882,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"ab05de1": {
"balancing:strategy": "weighting",
@@ -307,7 +372,12 @@
"classifier:gradient_boosting:scoring": "loss",
"classifier:gradient_boosting:tol": 1e-07,
"classifier:gradient_boosting:n_iter_no_change": 7,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"fb3d3cb": {
"balancing:strategy": "none",
@@ -331,7 +401,12 @@
"classifier:mlp:shuffle": "True",
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"b351885": {
"balancing:strategy": "weighting",
@@ -352,7 +427,12 @@
"classifier:gradient_boosting:scoring": "loss",
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.000146347913583504,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"e5e92a5": {
"balancing:strategy": "none",
@@ -373,7 +453,12 @@
"classifier:gradient_boosting:scoring": "loss",
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.08571475959664131,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"80918df": {
"balancing:strategy": "none",
@@ -399,7 +484,12 @@
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.0001412449266478571,
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"789a57a": {
"balancing:strategy": "none",
@@ -414,7 +504,12 @@
"classifier:passive_aggressive:fit_intercept": "True",
"classifier:passive_aggressive:loss": "squared_hinge",
"classifier:passive_aggressive:tol": 1.0363812529580006e-05,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"09b2b34": {
"balancing:strategy": "none",
@@ -435,7 +530,12 @@
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1224,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"7d4aa2b": {
"balancing:strategy": "none",
@@ -457,7 +557,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.008716562758519649,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.8571845899461281,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.19501334084607963,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"9fc9b85": {
"balancing:strategy": "none",
@@ -484,7 +589,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.009053655288403547,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.8467720098829357,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.047087352576090154,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"0a12676": {
"balancing:strategy": "weighting",
@@ -509,7 +619,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.0922294472020022,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"7cbf2e6": {
"balancing:strategy": "none",
@@ -531,7 +646,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.9759927921712355,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.20606466507293525,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"daaac58": {
"balancing:strategy": "weighting",
@@ -552,7 +672,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 744,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
"classifier:sgd:eta0": 3.9926577858082636e-05,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"9605116": {
"balancing:strategy": "weighting",
@@ -573,7 +698,12 @@
"classifier:gradient_boosting:scoring": "loss",
"classifier:gradient_boosting:tol": 1e-07,
"classifier:gradient_boosting:n_iter_no_change": 2,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"6c51fe7": {
"balancing:strategy": "none",
@@ -598,7 +728,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"255fc3f": {
"balancing:strategy": "weighting",
@@ -623,7 +758,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"5c39af4": {
"balancing:strategy": "none",
@@ -647,7 +787,12 @@
"classifier:mlp:shuffle": "True",
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"3b71f29": {
"balancing:strategy": "none",
@@ -667,7 +812,12 @@
"classifier:gradient_boosting:min_samples_leaf": 87,
"classifier:gradient_boosting:scoring": "loss",
"classifier:gradient_boosting:tol": 1e-07,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"1726d64": {
"balancing:strategy": "weighting",
@@ -688,7 +838,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1124,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
"classifier:sgd:l1_ratio": 0.02078042178977767,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"ddf8743": {
"balancing:strategy": "none",
@@ -710,7 +865,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.05399187079844309,
"classifier:gradient_boosting:n_iter_no_change": 7,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
}
},
"cutoffs": [
diff --git a/autosklearn/experimental/log_loss/askl2_portfolios/RF_None_5CV_iterative_es_if.json b/autosklearn/experimental/log_loss/askl2_portfolios/RF_None_5CV_iterative_es_if.json
index 451749d104..f4cf97ce11 100644
--- a/autosklearn/experimental/log_loss/askl2_portfolios/RF_None_5CV_iterative_es_if.json
+++ b/autosklearn/experimental/log_loss/askl2_portfolios/RF_None_5CV_iterative_es_if.json
@@ -20,7 +20,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"classifier:gradient_boosting:n_iter_no_change": 20,
"classifier:gradient_boosting:validation_fraction": 0.021337414482578282,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"04ec8ba": {
"balancing:strategy": "weighting",
@@ -45,7 +50,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.00010526559119801658,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"e26c1a0": {
"balancing:strategy": "weighting",
@@ -62,7 +72,12 @@
"classifier:passive_aggressive:tol": 0.00025296373413613473,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.7108230489029451,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.06748030619538817,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"d0b9bfd": {
"balancing:strategy": "none",
@@ -87,7 +102,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.23517300107634806,
"classifier:gradient_boosting:n_iter_no_change": 12,
"classifier:gradient_boosting:validation_fraction": 0.24423608491087115,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"f697bc3": {
"balancing:strategy": "weighting",
@@ -102,7 +122,12 @@
"classifier:passive_aggressive:fit_intercept": "True",
"classifier:passive_aggressive:loss": "squared_hinge",
"classifier:passive_aggressive:tol": 0.051629399473293695,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"91ff063": {
"balancing:strategy": "weighting",
@@ -127,7 +152,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.005591829098246694,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"5b541c8": {
"balancing:strategy": "weighting",
@@ -147,7 +177,12 @@
"classifier:gradient_boosting:min_samples_leaf": 1,
"classifier:gradient_boosting:scoring": "loss",
"classifier:gradient_boosting:tol": 1e-07,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"6e089e5": {
"balancing:strategy": "weighting",
@@ -169,7 +204,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.007472986296445738,
"classifier:gradient_boosting:n_iter_no_change": 13,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"6ab4d4e": {
"balancing:strategy": "none",
@@ -190,7 +230,12 @@
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.9922490300053561,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.1829734948438291,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"0ba913b": {
"balancing:strategy": "weighting",
@@ -214,7 +259,12 @@
"classifier:mlp:shuffle": "True",
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"cf5bab1": {
"balancing:strategy": "weighting",
@@ -236,7 +286,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.0665604479729369,
"classifier:gradient_boosting:n_iter_no_change": 7,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"297e609": {
"balancing:strategy": "none",
@@ -263,7 +318,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.1952627887023227,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1185,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"ace7e83": {
"balancing:strategy": "none",
@@ -288,7 +348,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.010000000000000004,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"1e5b9a4": {
"balancing:strategy": "none",
@@ -311,7 +376,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.048162732860304876,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.8792240750956144,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.034132091107883594,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"f937be6": {
"balancing:strategy": "none",
@@ -336,7 +406,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"2ff9871": {
"balancing:strategy": "weighting",
@@ -362,7 +437,12 @@
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.0042422461274778025,
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"4fe636f": {
"balancing:strategy": "weighting",
@@ -384,7 +464,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1928,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"35682b2": {
"balancing:strategy": "weighting",
@@ -406,7 +491,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"classifier:gradient_boosting:n_iter_no_change": 6,
"classifier:gradient_boosting:validation_fraction": 0.05433003250418479,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"eb711e8": {
"balancing:strategy": "none",
@@ -427,7 +517,12 @@
"classifier:gradient_boosting:scoring": "loss",
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.00016634440875057025,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"7e9bd96": {
"balancing:strategy": "none",
@@ -455,7 +550,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 115,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"27587a2": {
"balancing:strategy": "none",
@@ -474,7 +574,12 @@
"classifier:extra_trees:min_samples_leaf": 2,
"classifier:extra_trees:min_samples_split": 8,
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"10f054f": {
"balancing:strategy": "none",
@@ -500,7 +605,12 @@
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.35589413017822663,
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"dc9d49c": {
"balancing:strategy": "none",
@@ -523,7 +633,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 124,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
"classifier:gradient_boosting:n_iter_no_change": 12,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"ca6d42d": {
"balancing:strategy": "none",
@@ -538,7 +653,12 @@
"classifier:passive_aggressive:fit_intercept": "True",
"classifier:passive_aggressive:loss": "hinge",
"classifier:passive_aggressive:tol": 1.874391988200189e-05,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"b7fb7d2": {
"balancing:strategy": "none",
@@ -563,7 +683,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.00932259508492168,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"4f414ab": {
"balancing:strategy": "weighting",
@@ -591,7 +716,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.75,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.25,
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"f1c1dc4": {
"balancing:strategy": "none",
@@ -613,7 +743,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.010000000000000004,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.9154437384186069,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.008654073085106122,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"83d9459": {
"balancing:strategy": "weighting",
@@ -635,7 +770,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.0004231632120121218,
"classifier:gradient_boosting:n_iter_no_change": 10,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"0824a55": {
"balancing:strategy": "none",
@@ -662,7 +802,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.005344635230824308,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 778,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"82d949b": {
"balancing:strategy": "weighting",
@@ -684,7 +829,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.4281393406025153,
"classifier:gradient_boosting:n_iter_no_change": 10,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"1892815": {
"balancing:strategy": "weighting",
@@ -708,7 +858,12 @@
"classifier:mlp:shuffle": "True",
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"26b1b25": {
"balancing:strategy": "none",
@@ -730,7 +885,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
"classifier:sgd:eta0": 0.07771727761177989,
"classifier:sgd:l1_ratio": 0.1318744496830014,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
}
},
"cutoffs": [
diff --git a/autosklearn/experimental/log_loss/askl2_portfolios/RF_None_holdout_iterative_es_if.json b/autosklearn/experimental/log_loss/askl2_portfolios/RF_None_holdout_iterative_es_if.json
index 0c67a990c6..403d673e07 100644
--- a/autosklearn/experimental/log_loss/askl2_portfolios/RF_None_holdout_iterative_es_if.json
+++ b/autosklearn/experimental/log_loss/askl2_portfolios/RF_None_holdout_iterative_es_if.json
@@ -22,7 +22,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
"classifier:gradient_boosting:n_iter_no_change": 19,
"classifier:gradient_boosting:validation_fraction": 0.15148736566968157,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"fac7a35": {
"balancing:strategy": "weighting",
@@ -38,7 +43,12 @@
"classifier:passive_aggressive:loss": "squared_hinge",
"classifier:passive_aggressive:tol": 3.994632621445496e-05,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.01158501039849096,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"73f216b": {
"balancing:strategy": "none",
@@ -62,7 +72,12 @@
"classifier:mlp:shuffle": "True",
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"b73823e": {
"balancing:strategy": "weighting",
@@ -83,7 +98,12 @@
"classifier:gradient_boosting:scoring": "loss",
"classifier:gradient_boosting:tol": 1e-07,
"classifier:gradient_boosting:n_iter_no_change": 10,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"bdbd472": {
"balancing:strategy": "none",
@@ -105,7 +125,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.01261738379662994,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1909,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"e447ff7": {
"balancing:strategy": "none",
@@ -132,7 +157,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.003422849226960965,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1243,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"2224037": {
"balancing:strategy": "weighting",
@@ -157,7 +187,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
"classifier:gradient_boosting:n_iter_no_change": 19,
"classifier:gradient_boosting:validation_fraction": 0.10022515354287,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"c7e9f6c": {
"balancing:strategy": "weighting",
@@ -176,7 +211,12 @@
"classifier:sgd:tol": 0.03697396801428135,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.7964677178110968,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.06005060035154881,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"6931b43": {
"balancing:strategy": "weighting",
@@ -200,7 +240,12 @@
"classifier:mlp:shuffle": "True",
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"281561b": {
"balancing:strategy": "none",
@@ -222,7 +267,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1089,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"a1fef2a": {
"balancing:strategy": "none",
@@ -243,7 +293,12 @@
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.7455371733579619,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.24115086053573728,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"999a070": {
"balancing:strategy": "weighting",
@@ -268,7 +323,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.0033457517477028763,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"440cf58": {
"balancing:strategy": "weighting",
@@ -295,7 +355,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.4664181155003834,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1323,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"2656bee": {
"balancing:strategy": "none",
@@ -318,7 +383,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.3087815620469713,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.9632661624804552,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.22254575317019368,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"aeb22c7": {
"balancing:strategy": "weighting",
@@ -340,7 +410,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.9604530553396105,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.06974330376119121,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"8e0506b": {
"balancing:strategy": "none",
@@ -366,7 +441,12 @@
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.9664568034228809,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.05941132539521904,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"857fdf6": {
"balancing:strategy": "none",
@@ -387,7 +467,12 @@
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.8441739960764526,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.23854700600096612,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"8f4c1a4": {
"balancing:strategy": "none",
@@ -408,7 +493,12 @@
"classifier:gradient_boosting:scoring": "loss",
"classifier:gradient_boosting:tol": 1e-07,
"classifier:gradient_boosting:n_iter_no_change": 6,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"d2d26b1": {
"balancing:strategy": "none",
@@ -433,7 +523,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"4332765": {
"balancing:strategy": "weighting",
@@ -456,7 +551,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.0002722722541363936,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1000,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"92c8a99": {
"balancing:strategy": "none",
@@ -480,7 +580,12 @@
"classifier:mlp:shuffle": "True",
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"84e34e8": {
"balancing:strategy": "weighting",
@@ -505,7 +610,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.3203579669193948,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"ca1afa4": {
"balancing:strategy": "weighting",
@@ -530,7 +640,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
"classifier:gradient_boosting:n_iter_no_change": 12,
"classifier:gradient_boosting:validation_fraction": 0.118292989361417,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"ca3d668": {
"balancing:strategy": "weighting",
@@ -555,7 +670,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.24247020764480076,
"classifier:gradient_boosting:n_iter_no_change": 19,
"classifier:gradient_boosting:validation_fraction": 0.08951357569062098,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"619947f": {
"balancing:strategy": "none",
@@ -577,7 +697,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
"classifier:sgd:eta0": 0.03270619114268647,
"classifier:sgd:l1_ratio": 0.6366758350267246,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"947b3e7": {
"balancing:strategy": "none",
@@ -597,7 +722,12 @@
"classifier:sgd:epsilon": 3.6337948240776306e-05,
"classifier:sgd:eta0": 0.09955932870902368,
"classifier:sgd:power_t": 0.0367818634524381,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"d366b63": {
"balancing:strategy": "weighting",
@@ -619,7 +749,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.002079412293512545,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.7046267798963054,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.0017897043469726544,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"e2da229": {
"balancing:strategy": "none",
@@ -646,7 +781,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.02039634466784319,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 440,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"47c2f5a": {
"balancing:strategy": "weighting",
@@ -663,7 +803,12 @@
"classifier:passive_aggressive:tol": 6.097797160523113e-05,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1167,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"2dca877": {
"balancing:strategy": "weighting",
@@ -690,7 +835,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.019932607318403454,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.7540212502267323,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.2188034291959503,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"95b36c5": {
"balancing:strategy": "weighting",
@@ -711,7 +861,12 @@
"classifier:gradient_boosting:scoring": "loss",
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.025627338912525964,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"981e948": {
"balancing:strategy": "none",
@@ -731,7 +886,12 @@
"classifier:extra_trees:min_samples_split": 4,
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.026426948053594874,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
}
},
"cutoffs": [
diff --git a/autosklearn/experimental/log_loss/askl2_portfolios/RF_SH-eta4-i_10CV_iterative_es_if.json b/autosklearn/experimental/log_loss/askl2_portfolios/RF_SH-eta4-i_10CV_iterative_es_if.json
index 9da086e17f..2f4886e3ca 100644
--- a/autosklearn/experimental/log_loss/askl2_portfolios/RF_SH-eta4-i_10CV_iterative_es_if.json
+++ b/autosklearn/experimental/log_loss/askl2_portfolios/RF_SH-eta4-i_10CV_iterative_es_if.json
@@ -21,7 +21,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.006280403042161129,
"classifier:gradient_boosting:n_iter_no_change": 17,
"classifier:gradient_boosting:validation_fraction": 0.12820455722720236,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"69a5aed": {
"balancing:strategy": "weighting",
@@ -46,7 +51,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.020507323983341024,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"278a3b7": {
"balancing:strategy": "weighting",
@@ -64,7 +74,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.043140728717176134,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.7565211582936167,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.2427714099781957,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"c708530": {
"balancing:strategy": "weighting",
@@ -87,7 +102,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.002118717924807201,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.8009018127136468,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.2568493863080247,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"eaafcc8": {
"balancing:strategy": "weighting",
@@ -113,7 +133,12 @@
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.29539064866700554,
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"11b4871": {
"balancing:strategy": "weighting",
@@ -134,7 +159,12 @@
"classifier:gradient_boosting:scoring": "loss",
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.0003108062131055452,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"cccabeb": {
"balancing:strategy": "weighting",
@@ -158,7 +188,12 @@
"classifier:mlp:shuffle": "True",
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"121a4ef": {
"balancing:strategy": "none",
@@ -179,7 +214,12 @@
"classifier:gradient_boosting:scoring": "loss",
"classifier:gradient_boosting:tol": 1e-07,
"classifier:gradient_boosting:n_iter_no_change": 16,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"bad65bc": {
"balancing:strategy": "none",
@@ -201,7 +241,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.003937421024884423,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.72347247098442,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.08078070293132113,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"6da0e1a": {
"balancing:strategy": "weighting",
@@ -228,7 +273,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.002154531744214574,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.9579499947602039,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.027650749608723685,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"2d37001": {
"balancing:strategy": "weighting",
@@ -253,7 +303,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"8285215": {
"balancing:strategy": "weighting",
@@ -275,7 +330,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.010000000000000004,
"classifier:gradient_boosting:n_iter_no_change": 3,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"0ed94fb": {
"balancing:strategy": "weighting",
@@ -301,7 +361,12 @@
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.19791986161487668,
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"c11fce2": {
"balancing:strategy": "weighting",
@@ -320,7 +385,12 @@
"classifier:sgd:tol": 0.0012132021248816036,
"classifier:sgd:eta0": 0.00030412915402748054,
"classifier:sgd:power_t": 0.3300398436994113,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"0c324d9": {
"balancing:strategy": "weighting",
@@ -341,7 +411,12 @@
"classifier:gradient_boosting:scoring": "loss",
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.0036104288679581994,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"423e2f5": {
"balancing:strategy": "none",
@@ -364,7 +439,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.024016514710042437,
"classifier:gradient_boosting:n_iter_no_change": 11,
"classifier:gradient_boosting:validation_fraction": 0.17382179382094004,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"a82be94": {
"balancing:strategy": "none",
@@ -385,7 +465,12 @@
"classifier:gradient_boosting:scoring": "loss",
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.006309920762276479,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"040bd83": {
"balancing:strategy": "none",
@@ -412,7 +497,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.0011172717285354973,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 848,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"72903c0": {
"balancing:strategy": "none",
@@ -432,7 +522,12 @@
"classifier:gradient_boosting:min_samples_leaf": 154,
"classifier:gradient_boosting:scoring": "loss",
"classifier:gradient_boosting:tol": 1e-07,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"5ac8b4c": {
"balancing:strategy": "weighting",
@@ -457,7 +552,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.009080589517127771,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"dcc5375": {
"balancing:strategy": "weighting",
@@ -476,7 +576,12 @@
"classifier:extra_trees:min_samples_leaf": 1,
"classifier:extra_trees:min_samples_split": 2,
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"6c58cb5": {
"balancing:strategy": "weighting",
@@ -497,7 +602,12 @@
"classifier:gradient_boosting:scoring": "loss",
"classifier:gradient_boosting:tol": 1e-07,
"classifier:gradient_boosting:n_iter_no_change": 2,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"494f574": {
"balancing:strategy": "none",
@@ -522,7 +632,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.029737145358412768,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"395da55": {
"balancing:strategy": "weighting",
@@ -549,7 +664,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.7076847567891946,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.005025836888703367,
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"bf2837d": {
"balancing:strategy": "none",
@@ -569,7 +689,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1545,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
"classifier:sgd:l1_ratio": 0.0016338977047375043,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"4cf2fa5": {
"balancing:strategy": "none",
@@ -591,7 +716,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
"classifier:sgd:epsilon": 0.0012221439818245812,
"classifier:sgd:eta0": 0.008943497619680189,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"5fc1b32": {
"balancing:strategy": "none",
@@ -614,7 +744,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.010000000000000004,
"classifier:gradient_boosting:n_iter_no_change": 14,
"classifier:gradient_boosting:validation_fraction": 0.06548786110794219,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"d87827f": {
"balancing:strategy": "none",
@@ -636,7 +771,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"classifier:gradient_boosting:n_iter_no_change": 19,
"classifier:gradient_boosting:validation_fraction": 0.07410624737565304,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"596ffcf": {
"balancing:strategy": "none",
@@ -657,7 +797,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1244,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
"classifier:sgd:l1_ratio": 0.030723049722014444,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"7e1cc2b": {
"balancing:strategy": "none",
@@ -677,7 +822,12 @@
"classifier:gradient_boosting:min_samples_leaf": 151,
"classifier:gradient_boosting:scoring": "loss",
"classifier:gradient_boosting:tol": 1e-07,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"e5c9543": {
"balancing:strategy": "weighting",
@@ -694,7 +844,12 @@
"classifier:sgd:loss": "squared_hinge",
"classifier:sgd:penalty": "l1",
"classifier:sgd:tol": 0.0004953514665899356,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"231dfac": {
"balancing:strategy": "none",
@@ -717,7 +872,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 853,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
"classifier:gradient_boosting:n_iter_no_change": 2,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
}
},
"cutoffs": [
diff --git a/autosklearn/experimental/log_loss/askl2_portfolios/RF_SH-eta4-i_3CV_iterative_es_if.json b/autosklearn/experimental/log_loss/askl2_portfolios/RF_SH-eta4-i_3CV_iterative_es_if.json
index 5d5a8153ec..26070cd64f 100644
--- a/autosklearn/experimental/log_loss/askl2_portfolios/RF_SH-eta4-i_3CV_iterative_es_if.json
+++ b/autosklearn/experimental/log_loss/askl2_portfolios/RF_SH-eta4-i_3CV_iterative_es_if.json
@@ -21,7 +21,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.0004981969019895562,
"classifier:gradient_boosting:n_iter_no_change": 4,
"classifier:gradient_boosting:validation_fraction": 0.23374473517320618,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"772b3a8": {
"balancing:strategy": "none",
@@ -47,7 +52,12 @@
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1000,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"5d80c27": {
"balancing:strategy": "weighting",
@@ -63,7 +73,12 @@
"classifier:passive_aggressive:loss": "squared_hinge",
"classifier:passive_aggressive:tol": 0.0005787736999064962,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.012192559003185645,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"6ddc39b": {
"balancing:strategy": "weighting",
@@ -87,7 +102,12 @@
"classifier:mlp:shuffle": "True",
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"e5c81ce": {
"balancing:strategy": "weighting",
@@ -110,7 +130,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1259,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
"classifier:gradient_boosting:n_iter_no_change": 8,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"e623b07": {
"balancing:strategy": "none",
@@ -134,7 +159,12 @@
"classifier:mlp:shuffle": "True",
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"05b30ee": {
"balancing:strategy": "weighting",
@@ -159,7 +189,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.25794352734499837,
"classifier:gradient_boosting:n_iter_no_change": 20,
"classifier:gradient_boosting:validation_fraction": 0.06937115181143882,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"f203d83": {
"balancing:strategy": "weighting",
@@ -179,7 +214,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.011532350506496297,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.7645599467770883,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.0011091488435678166,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"6554e90": {
"balancing:strategy": "none",
@@ -200,7 +240,12 @@
"classifier:gradient_boosting:scoring": "loss",
"classifier:gradient_boosting:tol": 1e-07,
"classifier:gradient_boosting:n_iter_no_change": 11,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"0e85550": {
"balancing:strategy": "weighting",
@@ -220,7 +265,12 @@
"classifier:extra_trees:min_samples_split": 2,
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.0007621274530118648,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"50aa7ad": {
"balancing:strategy": "weighting",
@@ -245,7 +295,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.006044015369290746,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"7d4aa2b": {
"balancing:strategy": "none",
@@ -267,7 +322,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.008716562758519649,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.8571845899461281,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.19501334084607963,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"1d66cc6": {
"balancing:strategy": "none",
@@ -291,7 +351,12 @@
"classifier:mlp:shuffle": "True",
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"ef2bd55": {
"balancing:strategy": "weighting",
@@ -315,7 +380,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.714072222464285,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.22744859003175397,
"classifier:gradient_boosting:n_iter_no_change": 19,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"6973c5f": {
"balancing:strategy": "none",
@@ -336,7 +406,12 @@
"classifier:gradient_boosting:scoring": "loss",
"classifier:gradient_boosting:tol": 1e-07,
"classifier:gradient_boosting:n_iter_no_change": 3,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"789a57a": {
"balancing:strategy": "none",
@@ -351,7 +426,12 @@
"classifier:passive_aggressive:fit_intercept": "True",
"classifier:passive_aggressive:loss": "squared_hinge",
"classifier:passive_aggressive:tol": 1.0363812529580006e-05,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"919a721": {
"balancing:strategy": "none",
@@ -373,7 +453,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.009771641527306337,
"classifier:gradient_boosting:n_iter_no_change": 9,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"3b71f29": {
"balancing:strategy": "none",
@@ -393,7 +478,12 @@
"classifier:gradient_boosting:min_samples_leaf": 87,
"classifier:gradient_boosting:scoring": "loss",
"classifier:gradient_boosting:tol": 1e-07,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"a43e999": {
"balancing:strategy": "weighting",
@@ -418,7 +508,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.0006140067400027361,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"6859bdd": {
"balancing:strategy": "none",
@@ -438,7 +533,12 @@
"classifier:gradient_boosting:min_samples_leaf": 1,
"classifier:gradient_boosting:scoring": "loss",
"classifier:gradient_boosting:tol": 1e-07,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"a03ce20": {
"balancing:strategy": "none",
@@ -465,7 +565,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.01240791256186952,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.7665670083302197,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.25,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"172a64f": {
"balancing:strategy": "none",
@@ -487,7 +592,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.0014143330565857936,
"classifier:gradient_boosting:n_iter_no_change": 16,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"80918df": {
"balancing:strategy": "none",
@@ -513,7 +623,12 @@
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.0001412449266478571,
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"1726d64": {
"balancing:strategy": "weighting",
@@ -534,7 +649,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1124,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
"classifier:sgd:l1_ratio": 0.02078042178977767,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"9605116": {
"balancing:strategy": "weighting",
@@ -555,7 +675,12 @@
"classifier:gradient_boosting:scoring": "loss",
"classifier:gradient_boosting:tol": 1e-07,
"classifier:gradient_boosting:n_iter_no_change": 2,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"ef6b017": {
"balancing:strategy": "weighting",
@@ -580,7 +705,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.038744762665218604,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"ce18bb7": {
"balancing:strategy": "weighting",
@@ -604,7 +734,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.7521302105202969,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.2619140818965272,
"classifier:gradient_boosting:n_iter_no_change": 20,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"255fc3f": {
"balancing:strategy": "weighting",
@@ -629,7 +764,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"e5e92a5": {
"balancing:strategy": "none",
@@ -650,7 +790,12 @@
"classifier:gradient_boosting:scoring": "loss",
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.08571475959664131,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"6d8d3a8": {
"balancing:strategy": "weighting",
@@ -677,7 +822,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.014867338455589912,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1762,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"84f6f17": {
"balancing:strategy": "weighting",
@@ -698,7 +848,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
"classifier:sgd:eta0": 3.091511250116596e-05,
"classifier:sgd:l1_ratio": 0.3302336331591988,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"6c51fe7": {
"balancing:strategy": "none",
@@ -723,7 +878,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
}
},
"cutoffs": [
diff --git a/autosklearn/experimental/log_loss/askl2_portfolios/RF_SH-eta4-i_5CV_iterative_es_if.json b/autosklearn/experimental/log_loss/askl2_portfolios/RF_SH-eta4-i_5CV_iterative_es_if.json
index 52f4d84c26..c3ea7b6241 100644
--- a/autosklearn/experimental/log_loss/askl2_portfolios/RF_SH-eta4-i_5CV_iterative_es_if.json
+++ b/autosklearn/experimental/log_loss/askl2_portfolios/RF_SH-eta4-i_5CV_iterative_es_if.json
@@ -20,7 +20,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"classifier:gradient_boosting:n_iter_no_change": 20,
"classifier:gradient_boosting:validation_fraction": 0.021337414482578282,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"0ba913b": {
"balancing:strategy": "weighting",
@@ -44,7 +49,12 @@
"classifier:mlp:shuffle": "True",
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"e26c1a0": {
"balancing:strategy": "weighting",
@@ -61,7 +71,12 @@
"classifier:passive_aggressive:tol": 0.00025296373413613473,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.7108230489029451,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.06748030619538817,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"1bbdcf7": {
"balancing:strategy": "weighting",
@@ -83,7 +98,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1994,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"f937be6": {
"balancing:strategy": "none",
@@ -108,7 +128,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"b174099": {
"balancing:strategy": "weighting",
@@ -123,7 +148,12 @@
"classifier:passive_aggressive:fit_intercept": "True",
"classifier:passive_aggressive:loss": "squared_hinge",
"classifier:passive_aggressive:tol": 0.00016439976946376104,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"04ec8ba": {
"balancing:strategy": "weighting",
@@ -148,7 +178,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.00010526559119801658,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"403e74e": {
"balancing:strategy": "weighting",
@@ -166,7 +201,12 @@
"classifier:sgd:penalty": "l1",
"classifier:sgd:tol": 2.7281817942839784e-05,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.00930364133754329,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"83d9459": {
"balancing:strategy": "weighting",
@@ -188,7 +228,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.0004231632120121218,
"classifier:gradient_boosting:n_iter_no_change": 10,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"6ab4d4e": {
"balancing:strategy": "none",
@@ -209,7 +254,12 @@
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.9922490300053561,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.1829734948438291,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"6e089e5": {
"balancing:strategy": "weighting",
@@ -231,7 +281,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.007472986296445738,
"classifier:gradient_boosting:n_iter_no_change": 13,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"2ff9871": {
"balancing:strategy": "weighting",
@@ -257,7 +312,12 @@
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.0042422461274778025,
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"8933744": {
"balancing:strategy": "weighting",
@@ -277,7 +337,12 @@
"classifier:gradient_boosting:min_samples_leaf": 95,
"classifier:gradient_boosting:scoring": "loss",
"classifier:gradient_boosting:tol": 1e-07,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"271d452": {
"balancing:strategy": "weighting",
@@ -297,7 +362,12 @@
"classifier:gradient_boosting:min_samples_leaf": 24,
"classifier:gradient_boosting:scoring": "loss",
"classifier:gradient_boosting:tol": 1e-07,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"8c9971b": {
"balancing:strategy": "none",
@@ -324,7 +394,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.02245420281185916,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1103,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"9eaeaab": {
"balancing:strategy": "weighting",
@@ -350,7 +425,12 @@
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.8926817285376502,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.08139907756830209,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"ace7e83": {
"balancing:strategy": "none",
@@ -375,7 +455,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.010000000000000004,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"dc9d49c": {
"balancing:strategy": "none",
@@ -398,7 +483,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 124,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
"classifier:gradient_boosting:n_iter_no_change": 12,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"7e9bd96": {
"balancing:strategy": "none",
@@ -426,7 +516,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 115,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"0c11923": {
"balancing:strategy": "none",
@@ -447,7 +542,12 @@
"classifier:gradient_boosting:scoring": "loss",
"classifier:gradient_boosting:tol": 1e-07,
"classifier:gradient_boosting:n_iter_no_change": 8,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"ca6d42d": {
"balancing:strategy": "none",
@@ -462,7 +562,12 @@
"classifier:passive_aggressive:fit_intercept": "True",
"classifier:passive_aggressive:loss": "hinge",
"classifier:passive_aggressive:tol": 1.874391988200189e-05,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"27587a2": {
"balancing:strategy": "none",
@@ -481,7 +586,12 @@
"classifier:extra_trees:min_samples_leaf": 2,
"classifier:extra_trees:min_samples_split": 8,
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"5367b82": {
"balancing:strategy": "none",
@@ -503,7 +613,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"classifier:gradient_boosting:n_iter_no_change": 6,
"classifier:gradient_boosting:validation_fraction": 0.11360221557156701,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"050fe7a": {
"balancing:strategy": "none",
@@ -524,7 +639,12 @@
"classifier:gradient_boosting:scoring": "loss",
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.008032087955934802,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"4f50608": {
"balancing:strategy": "weighting",
@@ -550,7 +670,12 @@
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.724059454604005,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.25,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"4f414ab": {
"balancing:strategy": "weighting",
@@ -578,7 +703,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.75,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.25,
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"26b1b25": {
"balancing:strategy": "none",
@@ -600,7 +730,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
"classifier:sgd:eta0": 0.07771727761177989,
"classifier:sgd:l1_ratio": 0.1318744496830014,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"8f7d99e": {
"balancing:strategy": "weighting",
@@ -625,7 +760,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.0030354192387363837,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"cca9a17": {
"balancing:strategy": "weighting",
@@ -644,7 +784,12 @@
"classifier:extra_trees:min_samples_leaf": 1,
"classifier:extra_trees:min_samples_split": 2,
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"28afd70": {
"balancing:strategy": "weighting",
@@ -669,7 +814,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.010271514866804008,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"2373433": {
"balancing:strategy": "weighting",
@@ -694,7 +844,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.004960032952598242,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"0a2cd4a": {
"balancing:strategy": "none",
@@ -717,7 +872,12 @@
"classifier:sgd:epsilon": 0.001977206007382024,
"classifier:sgd:eta0": 0.008049468921629188,
"classifier:sgd:l1_ratio": 0.5783321148432391,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
}
},
"cutoffs": [
diff --git a/autosklearn/experimental/log_loss/askl2_portfolios/RF_SH-eta4-i_holdout_iterative_es_if.json b/autosklearn/experimental/log_loss/askl2_portfolios/RF_SH-eta4-i_holdout_iterative_es_if.json
index e3aead06b8..2cfd4df384 100644
--- a/autosklearn/experimental/log_loss/askl2_portfolios/RF_SH-eta4-i_holdout_iterative_es_if.json
+++ b/autosklearn/experimental/log_loss/askl2_portfolios/RF_SH-eta4-i_holdout_iterative_es_if.json
@@ -22,7 +22,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
"classifier:gradient_boosting:n_iter_no_change": 19,
"classifier:gradient_boosting:validation_fraction": 0.15148736566968157,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"fac7a35": {
"balancing:strategy": "weighting",
@@ -38,7 +43,12 @@
"classifier:passive_aggressive:loss": "squared_hinge",
"classifier:passive_aggressive:tol": 3.994632621445496e-05,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.01158501039849096,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"73f216b": {
"balancing:strategy": "none",
@@ -62,7 +72,12 @@
"classifier:mlp:shuffle": "True",
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"e447ff7": {
"balancing:strategy": "none",
@@ -89,7 +104,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.003422849226960965,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1243,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"b73823e": {
"balancing:strategy": "weighting",
@@ -110,7 +130,12 @@
"classifier:gradient_boosting:scoring": "loss",
"classifier:gradient_boosting:tol": 1e-07,
"classifier:gradient_boosting:n_iter_no_change": 10,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"6931b43": {
"balancing:strategy": "weighting",
@@ -134,7 +159,12 @@
"classifier:mlp:shuffle": "True",
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"c7e9f6c": {
"balancing:strategy": "weighting",
@@ -153,7 +183,12 @@
"classifier:sgd:tol": 0.03697396801428135,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.7964677178110968,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.06005060035154881,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"667d5f4": {
"balancing:strategy": "weighting",
@@ -176,7 +211,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.36511540247766866,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1178,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"bdbd472": {
"balancing:strategy": "none",
@@ -198,7 +238,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.01261738379662994,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1909,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"eff997e": {
"balancing:strategy": "none",
@@ -223,7 +268,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.33902719556466043,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"aeb22c7": {
"balancing:strategy": "weighting",
@@ -245,7 +295,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.9604530553396105,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.06974330376119121,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"a1fef2a": {
"balancing:strategy": "none",
@@ -266,7 +321,12 @@
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.7455371733579619,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.24115086053573728,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"2224037": {
"balancing:strategy": "weighting",
@@ -291,7 +351,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
"classifier:gradient_boosting:n_iter_no_change": 19,
"classifier:gradient_boosting:validation_fraction": 0.10022515354287,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"a56e902": {
"balancing:strategy": "none",
@@ -316,7 +381,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.00023767330801058195,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"fa2f9f6": {
"balancing:strategy": "none",
@@ -339,7 +409,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.014822445699720594,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.7960305699594659,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.14211609181777107,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"47c2f5a": {
"balancing:strategy": "weighting",
@@ -356,7 +431,12 @@
"classifier:passive_aggressive:tol": 6.097797160523113e-05,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1167,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"d2d26b1": {
"balancing:strategy": "none",
@@ -381,7 +461,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"d366b63": {
"balancing:strategy": "weighting",
@@ -403,7 +488,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.002079412293512545,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.7046267798963054,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.0017897043469726544,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"4be27b9": {
"balancing:strategy": "none",
@@ -426,7 +516,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.03523864695144828,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.7180329793931687,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.1033159225302246,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"8e0506b": {
"balancing:strategy": "none",
@@ -452,7 +547,12 @@
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.9664568034228809,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.05941132539521904,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"7fb1f2e": {
"balancing:strategy": "none",
@@ -474,7 +574,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.024951678116134714,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.9967193703099239,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.2759853763113785,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"4332765": {
"balancing:strategy": "weighting",
@@ -497,7 +602,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.0002722722541363936,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1000,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"95b36c5": {
"balancing:strategy": "weighting",
@@ -518,7 +628,12 @@
"classifier:gradient_boosting:scoring": "loss",
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.025627338912525964,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"947b3e7": {
"balancing:strategy": "none",
@@ -538,7 +653,12 @@
"classifier:sgd:epsilon": 3.6337948240776306e-05,
"classifier:sgd:eta0": 0.09955932870902368,
"classifier:sgd:power_t": 0.0367818634524381,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"aaa7355": {
"balancing:strategy": "none",
@@ -563,7 +683,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.0248276234342269,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"619947f": {
"balancing:strategy": "none",
@@ -585,7 +710,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
"classifier:sgd:eta0": 0.03270619114268647,
"classifier:sgd:l1_ratio": 0.6366758350267246,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"84e34e8": {
"balancing:strategy": "weighting",
@@ -610,7 +740,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.3203579669193948,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"440cf58": {
"balancing:strategy": "weighting",
@@ -637,7 +772,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.4664181155003834,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1323,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"92a2b3e": {
"balancing:strategy": "none",
@@ -662,7 +802,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.043137864752964175,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"d4b1080": {
"balancing:strategy": "weighting",
@@ -684,7 +829,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
"classifier:sgd:eta0": 2.7253528919863838e-05,
"classifier:sgd:l1_ratio": 0.09043420706418341,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"6933af5": {
"balancing:strategy": "weighting",
@@ -711,7 +861,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.15923339809849205,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.9921300116071107,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.006717144431080089,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"c2bfd40": {
"balancing:strategy": "none",
@@ -727,7 +882,12 @@
"classifier:passive_aggressive:loss": "squared_hinge",
"classifier:passive_aggressive:tol": 1.197123139199107e-05,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.40377267259792243,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
}
},
"cutoffs": [
diff --git a/autosklearn/experimental/roc_auc/askl2_portfolios/RF_None_10CV_iterative_es_if.json b/autosklearn/experimental/roc_auc/askl2_portfolios/RF_None_10CV_iterative_es_if.json
index 6be8f31b8c..71eaa43a07 100644
--- a/autosklearn/experimental/roc_auc/askl2_portfolios/RF_None_10CV_iterative_es_if.json
+++ b/autosklearn/experimental/roc_auc/askl2_portfolios/RF_None_10CV_iterative_es_if.json
@@ -20,7 +20,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.8406370831514932,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.20389889461388738,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"2e26e3e": {
"balancing:strategy": "weighting",
@@ -47,7 +52,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 275,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"e457d9c": {
"balancing:strategy": "weighting",
@@ -69,7 +79,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.10560266211666122,
"classifier:gradient_boosting:n_iter_no_change": 1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"ef55f97": {
"balancing:strategy": "weighting",
@@ -91,7 +106,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"classifier:gradient_boosting:n_iter_no_change": 1,
"classifier:gradient_boosting:validation_fraction": 0.17760188229708512,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"1e0cf43": {
"balancing:strategy": "none",
@@ -115,7 +135,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.76063568788292,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.189634987062873,
"classifier:gradient_boosting:n_iter_no_change": 17,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"40266af": {
"balancing:strategy": "weighting",
@@ -136,7 +161,12 @@
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.7569526888248843,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.2412613619390905,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"7b353d7": {
"balancing:strategy": "weighting",
@@ -154,7 +184,12 @@
"classifier:sgd:penalty": "l2",
"classifier:sgd:tol": 0.0005568060975043113,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.01053168704546915,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"5aab6e4": {
"balancing:strategy": "weighting",
@@ -180,7 +215,12 @@
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.021809618585192588,
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"cb9ea50": {
"balancing:strategy": "none",
@@ -202,7 +242,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"classifier:gradient_boosting:n_iter_no_change": 11,
"classifier:gradient_boosting:validation_fraction": 0.2767824989716728,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"4a934ee": {
"balancing:strategy": "weighting",
@@ -222,7 +267,12 @@
"classifier:gradient_boosting:min_samples_leaf": 1,
"classifier:gradient_boosting:scoring": "loss",
"classifier:gradient_boosting:tol": 1e-07,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"c0e6259": {
"balancing:strategy": "weighting",
@@ -249,7 +299,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.8207344726818622,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.050481917951293094,
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"14994aa": {
"balancing:strategy": "weighting",
@@ -264,7 +319,12 @@
"classifier:passive_aggressive:fit_intercept": "True",
"classifier:passive_aggressive:loss": "hinge",
"classifier:passive_aggressive:tol": 0.0002511675542745637,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"d2c45b7": {
"balancing:strategy": "none",
@@ -286,7 +346,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.0002152450797464187,
"classifier:gradient_boosting:n_iter_no_change": 11,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"604e370": {
"balancing:strategy": "none",
@@ -305,7 +370,12 @@
"classifier:extra_trees:min_samples_leaf": 2,
"classifier:extra_trees:min_samples_split": 6,
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"13bae60": {
"balancing:strategy": "weighting",
@@ -325,7 +395,12 @@
"classifier:extra_trees:min_samples_split": 2,
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.29531825974563736,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"bdb1c5c": {
"balancing:strategy": "weighting",
@@ -352,7 +427,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1087,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"6dec707": {
"balancing:strategy": "weighting",
@@ -378,7 +458,12 @@
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 66,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"5e8eb5f": {
"balancing:strategy": "none",
@@ -400,7 +485,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"classifier:gradient_boosting:n_iter_no_change": 1,
"classifier:gradient_boosting:validation_fraction": 0.10464338969122808,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"3373e00": {
"balancing:strategy": "weighting",
@@ -423,7 +513,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.49938868874150427,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1000,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"f37dd59": {
"balancing:strategy": "none",
@@ -448,7 +543,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"7c0aca6": {
"balancing:strategy": "weighting",
@@ -472,7 +572,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
"classifier:gradient_boosting:n_iter_no_change": 20,
"classifier:gradient_boosting:validation_fraction": 0.010012508536536974,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"8fda287": {
"balancing:strategy": "weighting",
@@ -490,7 +595,12 @@
"classifier:sgd:penalty": "l1",
"classifier:sgd:tol": 0.0003330846127844381,
"classifier:sgd:epsilon": 0.05751688066646589,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"3cc8774": {
"balancing:strategy": "none",
@@ -510,7 +620,12 @@
"classifier:extra_trees:min_samples_split": 12,
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.00232269597189769,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"b62b471": {
"balancing:strategy": "weighting",
@@ -533,7 +648,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.002371604450008676,
"classifier:gradient_boosting:n_iter_no_change": 12,
"classifier:gradient_boosting:validation_fraction": 0.3170800989882104,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"188a54d": {
"balancing:strategy": "none",
@@ -560,7 +680,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.010279798317958203,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.7529347700207953,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.2586577204324494,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"929f30b": {
"balancing:strategy": "weighting",
@@ -582,7 +707,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"classifier:gradient_boosting:n_iter_no_change": 9,
"classifier:gradient_boosting:validation_fraction": 0.21013865145204977,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"b03f092": {
"balancing:strategy": "none",
@@ -604,7 +734,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
"classifier:sgd:eta0": 4.881378632007281e-07,
"classifier:sgd:l1_ratio": 0.20434166925375527,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"27509eb": {
"balancing:strategy": "none",
@@ -626,7 +761,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"classifier:gradient_boosting:n_iter_no_change": 2,
"classifier:gradient_boosting:validation_fraction": 0.24884856514753817,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"f208775": {
"balancing:strategy": "none",
@@ -647,7 +787,12 @@
"classifier:gradient_boosting:scoring": "loss",
"classifier:gradient_boosting:tol": 1e-07,
"classifier:gradient_boosting:n_iter_no_change": 18,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"e3b07fe": {
"balancing:strategy": "none",
@@ -673,7 +818,12 @@
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.04337401371806109,
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"32ab8ad": {
"balancing:strategy": "none",
@@ -695,7 +845,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.4884138170289074,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1000,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"952e617": {
"balancing:strategy": "none",
@@ -714,7 +869,12 @@
"classifier:random_forest:min_samples_leaf": 15,
"classifier:random_forest:min_samples_split": 19,
"classifier:random_forest:min_weight_fraction_leaf": 0.0,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
}
},
"cutoffs": [
diff --git a/autosklearn/experimental/roc_auc/askl2_portfolios/RF_None_3CV_iterative_es_if.json b/autosklearn/experimental/roc_auc/askl2_portfolios/RF_None_3CV_iterative_es_if.json
index 1928878da9..d8b44f3c4c 100644
--- a/autosklearn/experimental/roc_auc/askl2_portfolios/RF_None_3CV_iterative_es_if.json
+++ b/autosklearn/experimental/roc_auc/askl2_portfolios/RF_None_3CV_iterative_es_if.json
@@ -19,7 +19,12 @@
"classifier:gradient_boosting:scoring": "loss",
"classifier:gradient_boosting:tol": 1e-07,
"classifier:gradient_boosting:n_iter_no_change": 1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"057a961": {
"balancing:strategy": "none",
@@ -44,7 +49,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.019274771652010168,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"18b9a26": {
"balancing:strategy": "none",
@@ -66,7 +76,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.01722698747911545,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.7057003246095424,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.19450556868583982,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"32b95ed": {
"balancing:strategy": "weighting",
@@ -88,7 +103,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.3101738987949935,
"classifier:gradient_boosting:n_iter_no_change": 9,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"3b9d012": {
"balancing:strategy": "weighting",
@@ -109,7 +129,12 @@
"classifier:gradient_boosting:scoring": "loss",
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.02647964603290406,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"e94ffc2": {
"balancing:strategy": "none",
@@ -132,7 +157,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.012839593604617927,
"classifier:gradient_boosting:n_iter_no_change": 3,
"classifier:gradient_boosting:validation_fraction": 0.3750029060371645,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"611fcd2": {
"balancing:strategy": "none",
@@ -153,7 +183,12 @@
"classifier:gradient_boosting:scoring": "loss",
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.04973487764532246,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"2cc7fa6": {
"balancing:strategy": "none",
@@ -177,7 +212,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
"classifier:gradient_boosting:n_iter_no_change": 5,
"classifier:gradient_boosting:validation_fraction": 0.11007668926581227,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"a507a1e": {
"balancing:strategy": "weighting",
@@ -200,7 +240,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.0012964352618779716,
"classifier:gradient_boosting:n_iter_no_change": 1,
"classifier:gradient_boosting:validation_fraction": 0.3636102017236529,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"6270995": {
"balancing:strategy": "weighting",
@@ -226,7 +271,12 @@
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.006335781408792834,
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"09c23ea": {
"balancing:strategy": "none",
@@ -248,7 +298,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"classifier:gradient_boosting:n_iter_no_change": 3,
"classifier:gradient_boosting:validation_fraction": 0.3906906166499882,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"8436422": {
"balancing:strategy": "none",
@@ -270,7 +325,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.000542132677012084,
"classifier:gradient_boosting:n_iter_no_change": 7,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"4fcd4ce": {
"balancing:strategy": "weighting",
@@ -286,7 +346,12 @@
"classifier:passive_aggressive:loss": "hinge",
"classifier:passive_aggressive:tol": 0.00013387314937898115,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.13494044036226888,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"fbd4f3d": {
"balancing:strategy": "none",
@@ -311,7 +376,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.00467217089753154,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"81f13b6": {
"balancing:strategy": "none",
@@ -339,7 +409,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 530,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"937efac": {
"balancing:strategy": "none",
@@ -366,7 +441,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.01037687839425122,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1474,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"670fc33": {
"balancing:strategy": "weighting",
@@ -389,7 +469,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.10770240781492481,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.7080966958283763,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.19286444496554173,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"ccb5d37": {
"balancing:strategy": "none",
@@ -411,7 +496,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"classifier:gradient_boosting:n_iter_no_change": 1,
"classifier:gradient_boosting:validation_fraction": 0.03367077228049312,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"0d34b7c": {
"balancing:strategy": "weighting",
@@ -438,7 +528,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.0009439166308847433,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 629,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"ea94610": {
"balancing:strategy": "none",
@@ -458,7 +553,12 @@
"classifier:random_forest:min_samples_split": 4,
"classifier:random_forest:min_weight_fraction_leaf": 0.0,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.2641107295539412,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"fb21bf8": {
"balancing:strategy": "weighting",
@@ -480,7 +580,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.010763165432149112,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 10,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"83389d2": {
"balancing:strategy": "weighting",
@@ -501,7 +606,12 @@
"classifier:gradient_boosting:scoring": "loss",
"classifier:gradient_boosting:tol": 1e-07,
"classifier:gradient_boosting:n_iter_no_change": 14,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"70d7673": {
"balancing:strategy": "weighting",
@@ -528,7 +638,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.03562417892318361,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 124,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"b4eb798": {
"balancing:strategy": "none",
@@ -556,7 +671,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.9780439652611674,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.17229915174281954,
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"0a5c673": {
"balancing:strategy": "weighting",
@@ -582,7 +702,12 @@
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.008616859943826969,
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"3c13b23": {
"balancing:strategy": "none",
@@ -607,7 +732,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"7a1979d": {
"balancing:strategy": "none",
@@ -627,7 +757,12 @@
"classifier:gradient_boosting:min_samples_leaf": 174,
"classifier:gradient_boosting:scoring": "loss",
"classifier:gradient_boosting:tol": 1e-07,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"db47f52": {
"balancing:strategy": "none",
@@ -649,7 +784,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.07919351726301219,
"classifier:gradient_boosting:n_iter_no_change": 18,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"57c1d3d": {
"balancing:strategy": "weighting",
@@ -668,7 +808,12 @@
"classifier:extra_trees:min_samples_leaf": 4,
"classifier:extra_trees:min_samples_split": 7,
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"330533a": {
"balancing:strategy": "weighting",
@@ -689,7 +834,12 @@
"classifier:gradient_boosting:scoring": "loss",
"classifier:gradient_boosting:tol": 1e-07,
"classifier:gradient_boosting:n_iter_no_change": 9,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"31f9d24": {
"balancing:strategy": "weighting",
@@ -709,7 +859,12 @@
"classifier:gradient_boosting:min_samples_leaf": 43,
"classifier:gradient_boosting:scoring": "loss",
"classifier:gradient_boosting:tol": 1e-07,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"b7a81a4": {
"balancing:strategy": "none",
@@ -732,7 +887,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.16522606245500177,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.8360632014850663,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.2880796632700646,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
}
},
"cutoffs": [
diff --git a/autosklearn/experimental/roc_auc/askl2_portfolios/RF_None_5CV_iterative_es_if.json b/autosklearn/experimental/roc_auc/askl2_portfolios/RF_None_5CV_iterative_es_if.json
index 4752cc9ecc..2f41ca5ffb 100644
--- a/autosklearn/experimental/roc_auc/askl2_portfolios/RF_None_5CV_iterative_es_if.json
+++ b/autosklearn/experimental/roc_auc/askl2_portfolios/RF_None_5CV_iterative_es_if.json
@@ -23,7 +23,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
"classifier:gradient_boosting:n_iter_no_change": 14,
"classifier:gradient_boosting:validation_fraction": 0.12446922535601551,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"48e7d72": {
"balancing:strategy": "none",
@@ -51,7 +56,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1633,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"3b5725c": {
"balancing:strategy": "none",
@@ -70,7 +80,12 @@
"classifier:random_forest:min_samples_leaf": 7,
"classifier:random_forest:min_samples_split": 6,
"classifier:random_forest:min_weight_fraction_leaf": 0.0,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"e8567fc": {
"balancing:strategy": "weighting",
@@ -88,7 +103,12 @@
"classifier:sgd:penalty": "l1",
"classifier:sgd:tol": 0.0007858702630144828,
"classifier:sgd:eta0": 1.3077397138182272e-06,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"8aac02c": {
"balancing:strategy": "weighting",
@@ -107,7 +127,12 @@
"classifier:extra_trees:min_samples_leaf": 1,
"classifier:extra_trees:min_samples_split": 2,
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"0b0f56b": {
"balancing:strategy": "none",
@@ -132,7 +157,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
"classifier:gradient_boosting:n_iter_no_change": 2,
"classifier:gradient_boosting:validation_fraction": 0.015980615501050732,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"766431d": {
"balancing:strategy": "none",
@@ -153,7 +183,12 @@
"classifier:sgd:eta0": 0.06268029581129594,
"classifier:sgd:l1_ratio": 1.2408983312626301e-05,
"classifier:sgd:power_t": 0.15473248271262038,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"9454294": {
"balancing:strategy": "weighting",
@@ -180,7 +215,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.8593841208287082,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.27443311984781915,
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"b85628c": {
"balancing:strategy": "none",
@@ -205,7 +245,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.014783697119079647,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"a8f7d08": {
"balancing:strategy": "weighting",
@@ -227,7 +272,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.03763157344543798,
"classifier:gradient_boosting:n_iter_no_change": 11,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"c5fe8c1": {
"balancing:strategy": "weighting",
@@ -250,7 +300,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.7772622710360984,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.25,
"classifier:gradient_boosting:n_iter_no_change": 14,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"7eb153b": {
"balancing:strategy": "weighting",
@@ -274,7 +329,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.1023091397829033,
"classifier:gradient_boosting:n_iter_no_change": 3,
"classifier:gradient_boosting:validation_fraction": 0.20375603224425706,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"ed48b1e": {
"balancing:strategy": "weighting",
@@ -300,7 +360,12 @@
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.19593845967764903,
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"fd4841b": {
"balancing:strategy": "weighting",
@@ -324,7 +389,12 @@
"classifier:mlp:shuffle": "True",
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"c3c5a39": {
"balancing:strategy": "none",
@@ -346,7 +416,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"classifier:gradient_boosting:n_iter_no_change": 13,
"classifier:gradient_boosting:validation_fraction": 0.09536739955550967,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"aeea657": {
"balancing:strategy": "weighting",
@@ -374,7 +449,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 968,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"d41cebb": {
"balancing:strategy": "none",
@@ -401,7 +481,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.014993536738242566,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1006,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"546748f": {
"balancing:strategy": "none",
@@ -428,7 +513,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.012282943233395478,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.7431608138789764,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.22854594976143439,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"ca57fd6": {
"balancing:strategy": "none",
@@ -449,7 +539,12 @@
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1805,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"d2a3f8c": {
"balancing:strategy": "none",
@@ -476,7 +571,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.010000000000000004,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1074,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"36cd396": {
"balancing:strategy": "weighting",
@@ -501,7 +601,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"09555d8": {
"balancing:strategy": "none",
@@ -523,7 +628,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.0016129570857335443,
"classifier:gradient_boosting:n_iter_no_change": 12,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"8ffaacd": {
"balancing:strategy": "none",
@@ -543,7 +653,12 @@
"classifier:random_forest:min_samples_split": 2,
"classifier:random_forest:min_weight_fraction_leaf": 0.0,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.0013280091773466545,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"4396ea3": {
"balancing:strategy": "weighting",
@@ -570,7 +685,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.04792826808552058,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 499,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"1d4a4c2": {
"balancing:strategy": "weighting",
@@ -591,7 +711,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 775,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
"classifier:sgd:eta0": 0.08292033673052085,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"038e453": {
"balancing:strategy": "weighting",
@@ -613,7 +738,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.010000000000000004,
"classifier:gradient_boosting:n_iter_no_change": 9,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"d2b0a20": {
"balancing:strategy": "none",
@@ -641,7 +771,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 989,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"7308c48": {
"balancing:strategy": "none",
@@ -666,7 +801,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
"classifier:gradient_boosting:n_iter_no_change": 17,
"classifier:gradient_boosting:validation_fraction": 0.19824476185559053,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"4e08230": {
"balancing:strategy": "weighting",
@@ -688,7 +828,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"classifier:gradient_boosting:n_iter_no_change": 11,
"classifier:gradient_boosting:validation_fraction": 0.010048821300335692,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"a9c54cc": {
"balancing:strategy": "none",
@@ -710,7 +855,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.0025534629798769164,
"classifier:gradient_boosting:n_iter_no_change": 19,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"c9ed268": {
"balancing:strategy": "none",
@@ -730,7 +880,12 @@
"classifier:extra_trees:min_samples_split": 2,
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.28925175872648934,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"3995b86": {
"balancing:strategy": "none",
@@ -752,7 +907,12 @@
"classifier:sgd:eta0": 0.0004991855218627508,
"classifier:sgd:l1_ratio": 0.5188568617869495,
"classifier:sgd:power_t": 0.040071332883664504,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
}
},
"cutoffs": [
diff --git a/autosklearn/experimental/roc_auc/askl2_portfolios/RF_None_holdout_iterative_es_if.json b/autosklearn/experimental/roc_auc/askl2_portfolios/RF_None_holdout_iterative_es_if.json
index 6d05ceeb47..29e15fc327 100644
--- a/autosklearn/experimental/roc_auc/askl2_portfolios/RF_None_holdout_iterative_es_if.json
+++ b/autosklearn/experimental/roc_auc/askl2_portfolios/RF_None_holdout_iterative_es_if.json
@@ -20,7 +20,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 467,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"c85132e": {
"balancing:strategy": "weighting",
@@ -45,7 +50,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.006864495761122583,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"3ffac3d": {
"balancing:strategy": "weighting",
@@ -65,7 +75,12 @@
"classifier:extra_trees:min_samples_split": 12,
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.011646685244486026,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"3861869": {
"balancing:strategy": "weighting",
@@ -89,7 +104,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.7621841297339562,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.2689588813833072,
"classifier:gradient_boosting:n_iter_no_change": 19,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"1d8ad41": {
"balancing:strategy": "none",
@@ -112,7 +132,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.010000000000000004,
"classifier:gradient_boosting:n_iter_no_change": 2,
"classifier:gradient_boosting:validation_fraction": 0.35512720315660784,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"93a94fc": {
"balancing:strategy": "none",
@@ -137,7 +162,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"88839ea": {
"balancing:strategy": "weighting",
@@ -159,7 +189,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.014692720865382352,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.8369065529221758,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.2738199611008032,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"98dcda2": {
"balancing:strategy": "weighting",
@@ -184,7 +219,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.019065400719425012,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"f3b89ca": {
"balancing:strategy": "none",
@@ -204,7 +244,12 @@
"classifier:random_forest:min_samples_split": 3,
"classifier:random_forest:min_weight_fraction_leaf": 0.0,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.00894575704976732,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"5da6d7e": {
"balancing:strategy": "none",
@@ -227,7 +272,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.02106255140796113,
"classifier:gradient_boosting:n_iter_no_change": 14,
"classifier:gradient_boosting:validation_fraction": 0.1034846810050602,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"4522c5c": {
"balancing:strategy": "none",
@@ -253,7 +303,12 @@
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.0031947270899377388,
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"3e86168": {
"balancing:strategy": "none",
@@ -277,7 +332,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.25,
"classifier:gradient_boosting:n_iter_no_change": 14,
"classifier:gradient_boosting:validation_fraction": 0.3430240407684142,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"e33f5d1": {
"balancing:strategy": "weighting",
@@ -295,7 +355,12 @@
"classifier:sgd:penalty": "elasticnet",
"classifier:sgd:tol": 0.00820385437008071,
"classifier:sgd:l1_ratio": 3.8082985862639424e-06,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"6e537da": {
"balancing:strategy": "none",
@@ -320,7 +385,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"e72a0d0": {
"balancing:strategy": "none",
@@ -343,7 +413,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.007883404564380745,
"classifier:gradient_boosting:n_iter_no_change": 1,
"classifier:gradient_boosting:validation_fraction": 0.058283078146253935,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"59fbdef": {
"balancing:strategy": "none",
@@ -361,7 +436,12 @@
"classifier:sgd:penalty": "l1",
"classifier:sgd:tol": 8.27706111772171e-05,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.007965405034296472,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"cbcd40f": {
"balancing:strategy": "none",
@@ -385,7 +465,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1229,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
"classifier:gradient_boosting:n_iter_no_change": 15,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"04b6e1c": {
"balancing:strategy": "weighting",
@@ -400,7 +485,12 @@
"classifier:passive_aggressive:fit_intercept": "True",
"classifier:passive_aggressive:loss": "squared_hinge",
"classifier:passive_aggressive:tol": 0.01877747201567429,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"4a96439": {
"balancing:strategy": "none",
@@ -419,7 +509,12 @@
"classifier:extra_trees:min_samples_leaf": 1,
"classifier:extra_trees:min_samples_split": 5,
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"80bc363": {
"balancing:strategy": "weighting",
@@ -441,7 +536,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"classifier:gradient_boosting:n_iter_no_change": 5,
"classifier:gradient_boosting:validation_fraction": 0.023676944177253988,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"4464ce4": {
"balancing:strategy": "none",
@@ -460,7 +560,12 @@
"classifier:extra_trees:min_samples_leaf": 5,
"classifier:extra_trees:min_samples_split": 10,
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"3a1cf69": {
"balancing:strategy": "weighting",
@@ -485,7 +590,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"3230065": {
"balancing:strategy": "none",
@@ -510,7 +620,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"e5d8521": {
"balancing:strategy": "weighting",
@@ -537,7 +652,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 445,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"ee1d407": {
"balancing:strategy": "weighting",
@@ -558,7 +678,12 @@
"classifier:gradient_boosting:scoring": "loss",
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.0030968404387141115,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"c9d403d": {
"balancing:strategy": "none",
@@ -579,7 +704,12 @@
"classifier:gradient_boosting:scoring": "loss",
"classifier:gradient_boosting:tol": 1e-07,
"classifier:gradient_boosting:n_iter_no_change": 18,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"5ec319d": {
"balancing:strategy": "none",
@@ -601,7 +731,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.0059309332615295795,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.7122245303069001,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.26765391722955484,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"0c4f40a": {
"balancing:strategy": "none",
@@ -625,7 +760,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 255,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
"classifier:gradient_boosting:n_iter_no_change": 11,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"87bd8ad": {
"balancing:strategy": "weighting",
@@ -652,7 +792,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.000219180545401084,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 88,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"33ce15c": {
"balancing:strategy": "none",
@@ -674,7 +819,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.2974843930412059,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 817,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"7166357": {
"balancing:strategy": "weighting",
@@ -693,7 +843,12 @@
"classifier:extra_trees:min_samples_leaf": 1,
"classifier:extra_trees:min_samples_split": 3,
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"fcfb9a5": {
"balancing:strategy": "none",
@@ -717,7 +872,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.1962257728731655,
"classifier:gradient_boosting:n_iter_no_change": 5,
"classifier:gradient_boosting:validation_fraction": 0.03541393128244466,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
}
},
"cutoffs": [
diff --git a/autosklearn/experimental/roc_auc/askl2_portfolios/RF_SH-eta4-i_10CV_iterative_es_if.json b/autosklearn/experimental/roc_auc/askl2_portfolios/RF_SH-eta4-i_10CV_iterative_es_if.json
index 8048cc14d4..41cbdbe805 100644
--- a/autosklearn/experimental/roc_auc/askl2_portfolios/RF_SH-eta4-i_10CV_iterative_es_if.json
+++ b/autosklearn/experimental/roc_auc/askl2_portfolios/RF_SH-eta4-i_10CV_iterative_es_if.json
@@ -20,7 +20,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.8406370831514932,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.20389889461388738,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"078caa9": {
"balancing:strategy": "none",
@@ -44,7 +49,12 @@
"classifier:mlp:shuffle": "True",
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"e457d9c": {
"balancing:strategy": "weighting",
@@ -66,7 +76,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.10560266211666122,
"classifier:gradient_boosting:n_iter_no_change": 1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"ef55f97": {
"balancing:strategy": "weighting",
@@ -88,7 +103,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"classifier:gradient_boosting:n_iter_no_change": 1,
"classifier:gradient_boosting:validation_fraction": 0.17760188229708512,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"3aa9e49": {
"balancing:strategy": "weighting",
@@ -109,7 +129,12 @@
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.7260837314772796,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.2525954332120069,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"2e26e3e": {
"balancing:strategy": "weighting",
@@ -136,7 +161,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 275,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"9fbdad0": {
"balancing:strategy": "weighting",
@@ -156,7 +186,12 @@
"classifier:extra_trees:min_samples_split": 11,
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.030049905394547737,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"f3d6f6e": {
"balancing:strategy": "weighting",
@@ -182,7 +217,12 @@
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.20784580770691768,
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"344e35a": {
"balancing:strategy": "none",
@@ -203,7 +243,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1023,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
"classifier:sgd:eta0": 0.09184400512698869,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"fd44352": {
"balancing:strategy": "weighting",
@@ -223,7 +268,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.002492245059153722,
"classifier:sgd:eta0": 0.011565276785900419,
"classifier:sgd:l1_ratio": 0.3056450833447781,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"3c53750": {
"balancing:strategy": "weighting",
@@ -250,7 +300,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.21581356247444286,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.75,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.25174946101953555,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"5e8eb5f": {
"balancing:strategy": "none",
@@ -272,7 +327,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"classifier:gradient_boosting:n_iter_no_change": 1,
"classifier:gradient_boosting:validation_fraction": 0.10464338969122808,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"48c6703": {
"balancing:strategy": "none",
@@ -298,7 +358,12 @@
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.0010897445029567562,
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"09b43c4": {
"balancing:strategy": "weighting",
@@ -322,7 +387,12 @@
"classifier:mlp:shuffle": "True",
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"fc6f1b6": {
"balancing:strategy": "weighting",
@@ -339,7 +409,12 @@
"classifier:sgd:loss": "perceptron",
"classifier:sgd:penalty": "l1",
"classifier:sgd:tol": 0.000390333931153359,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"11c8920": {
"balancing:strategy": "none",
@@ -364,7 +439,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.2686670281481464,
"classifier:gradient_boosting:n_iter_no_change": 18,
"classifier:gradient_boosting:validation_fraction": 0.1143156303642018,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"7b353d7": {
"balancing:strategy": "weighting",
@@ -382,7 +462,12 @@
"classifier:sgd:penalty": "l2",
"classifier:sgd:tol": 0.0005568060975043113,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.01053168704546915,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"1e0cf43": {
"balancing:strategy": "none",
@@ -406,7 +491,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.76063568788292,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.189634987062873,
"classifier:gradient_boosting:n_iter_no_change": 17,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"cb9ea50": {
"balancing:strategy": "none",
@@ -428,7 +518,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"classifier:gradient_boosting:n_iter_no_change": 11,
"classifier:gradient_boosting:validation_fraction": 0.2767824989716728,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"13bae60": {
"balancing:strategy": "weighting",
@@ -448,7 +543,12 @@
"classifier:extra_trees:min_samples_split": 2,
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.29531825974563736,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"14994aa": {
"balancing:strategy": "weighting",
@@ -463,7 +563,12 @@
"classifier:passive_aggressive:fit_intercept": "True",
"classifier:passive_aggressive:loss": "hinge",
"classifier:passive_aggressive:tol": 0.0002511675542745637,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"3cc8774": {
"balancing:strategy": "none",
@@ -483,7 +588,12 @@
"classifier:extra_trees:min_samples_split": 12,
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.00232269597189769,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"9f0a242": {
"balancing:strategy": "weighting",
@@ -508,7 +618,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"9395997": {
"balancing:strategy": "weighting",
@@ -534,7 +649,12 @@
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.7339627907968111,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.25,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"8bb9a95": {
"balancing:strategy": "weighting",
@@ -560,7 +680,12 @@
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.06335085775892373,
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"0c52278": {
"balancing:strategy": "none",
@@ -580,7 +705,12 @@
"classifier:extra_trees:min_samples_split": 2,
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.02066959974142489,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"9c231f7": {
"balancing:strategy": "none",
@@ -597,7 +727,12 @@
"classifier:passive_aggressive:tol": 0.092080536430842,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1282,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"fd2205c": {
"balancing:strategy": "weighting",
@@ -621,7 +756,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
"classifier:gradient_boosting:n_iter_no_change": 18,
"classifier:gradient_boosting:validation_fraction": 0.3960774294744941,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"9d1c409": {
"balancing:strategy": "weighting",
@@ -647,7 +787,12 @@
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.002464603685466766,
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"bdb1c5c": {
"balancing:strategy": "weighting",
@@ -674,7 +819,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1087,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"4a934ee": {
"balancing:strategy": "weighting",
@@ -694,7 +844,12 @@
"classifier:gradient_boosting:min_samples_leaf": 1,
"classifier:gradient_boosting:scoring": "loss",
"classifier:gradient_boosting:tol": 1e-07,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"32ab8ad": {
"balancing:strategy": "none",
@@ -716,7 +871,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.4884138170289074,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1000,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
}
},
"cutoffs": [
diff --git a/autosklearn/experimental/roc_auc/askl2_portfolios/RF_SH-eta4-i_3CV_iterative_es_if.json b/autosklearn/experimental/roc_auc/askl2_portfolios/RF_SH-eta4-i_3CV_iterative_es_if.json
index a74fd998d7..c88e7485a4 100644
--- a/autosklearn/experimental/roc_auc/askl2_portfolios/RF_SH-eta4-i_3CV_iterative_es_if.json
+++ b/autosklearn/experimental/roc_auc/askl2_portfolios/RF_SH-eta4-i_3CV_iterative_es_if.json
@@ -19,7 +19,12 @@
"classifier:gradient_boosting:scoring": "loss",
"classifier:gradient_boosting:tol": 1e-07,
"classifier:gradient_boosting:n_iter_no_change": 1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"4e2d8ae": {
"balancing:strategy": "none",
@@ -41,7 +46,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.00683041403778669,
"classifier:gradient_boosting:n_iter_no_change": 6,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"a507a1e": {
"balancing:strategy": "weighting",
@@ -64,7 +74,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.0012964352618779716,
"classifier:gradient_boosting:n_iter_no_change": 1,
"classifier:gradient_boosting:validation_fraction": 0.3636102017236529,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"3b9d012": {
"balancing:strategy": "weighting",
@@ -85,7 +100,12 @@
"classifier:gradient_boosting:scoring": "loss",
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.02647964603290406,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"37ed3d0": {
"balancing:strategy": "none",
@@ -110,7 +130,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"09c23ea": {
"balancing:strategy": "none",
@@ -132,7 +157,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"classifier:gradient_boosting:n_iter_no_change": 3,
"classifier:gradient_boosting:validation_fraction": 0.3906906166499882,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"70fe9d8": {
"balancing:strategy": "weighting",
@@ -153,7 +183,12 @@
"classifier:gradient_boosting:scoring": "loss",
"classifier:gradient_boosting:tol": 1e-07,
"classifier:gradient_boosting:n_iter_no_change": 1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"da2661d": {
"balancing:strategy": "none",
@@ -180,7 +215,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.028836545069486425,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 453,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"4134365": {
"balancing:strategy": "none",
@@ -206,7 +246,12 @@
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.22345877503930808,
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"0823d35": {
"balancing:strategy": "weighting",
@@ -228,7 +273,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
"classifier:sgd:eta0": 0.0010980866884579832,
"classifier:sgd:l1_ratio": 0.016536672161617715,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"b4eb798": {
"balancing:strategy": "none",
@@ -256,7 +306,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.9780439652611674,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.17229915174281954,
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"499a242": {
"balancing:strategy": "weighting",
@@ -275,7 +330,12 @@
"classifier:extra_trees:min_samples_leaf": 1,
"classifier:extra_trees:min_samples_split": 3,
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"86e9ac7": {
"balancing:strategy": "none",
@@ -295,7 +355,12 @@
"classifier:gradient_boosting:min_samples_leaf": 2,
"classifier:gradient_boosting:scoring": "loss",
"classifier:gradient_boosting:tol": 1e-07,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"4fcd4ce": {
"balancing:strategy": "weighting",
@@ -311,7 +376,12 @@
"classifier:passive_aggressive:loss": "hinge",
"classifier:passive_aggressive:tol": 0.00013387314937898115,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.13494044036226888,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"81f13b6": {
"balancing:strategy": "none",
@@ -339,7 +409,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 530,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"1981b8a": {
"balancing:strategy": "none",
@@ -359,7 +434,12 @@
"classifier:extra_trees:min_samples_split": 11,
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.006673931442484339,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"32b95ed": {
"balancing:strategy": "weighting",
@@ -381,7 +461,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.3101738987949935,
"classifier:gradient_boosting:n_iter_no_change": 9,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"611fcd2": {
"balancing:strategy": "none",
@@ -402,7 +487,12 @@
"classifier:gradient_boosting:scoring": "loss",
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.04973487764532246,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"0a5c673": {
"balancing:strategy": "weighting",
@@ -428,7 +518,12 @@
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.008616859943826969,
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"7d32be9": {
"balancing:strategy": "none",
@@ -451,7 +546,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.010741517843298895,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.8961119811910175,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.22345657028939003,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"1f3235b": {
"balancing:strategy": "none",
@@ -476,7 +576,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.02556245191219976,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"c1bec76": {
"balancing:strategy": "none",
@@ -499,7 +604,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.9477141363567148,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.04448426739885174,
"classifier:gradient_boosting:n_iter_no_change": 6,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"9bdce9a": {
"balancing:strategy": "none",
@@ -519,7 +629,12 @@
"classifier:gradient_boosting:min_samples_leaf": 7,
"classifier:gradient_boosting:scoring": "loss",
"classifier:gradient_boosting:tol": 1e-07,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"18708a2": {
"balancing:strategy": "weighting",
@@ -540,7 +655,12 @@
"classifier:gradient_boosting:scoring": "loss",
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.03356994088246032,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"b54d05f": {
"balancing:strategy": "none",
@@ -568,7 +688,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.9455241437884573,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.059719120916944465,
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"3f02da0": {
"balancing:strategy": "weighting",
@@ -590,7 +715,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
"classifier:sgd:epsilon": 3.109676948490368e-05,
"classifier:sgd:eta0": 2.5983206022574477e-05,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"21ff716": {
"balancing:strategy": "weighting",
@@ -616,7 +746,12 @@
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.02075801210840713,
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"f2e837c": {
"balancing:strategy": "weighting",
@@ -641,7 +776,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"fb21bf8": {
"balancing:strategy": "weighting",
@@ -663,7 +803,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.010763165432149112,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 10,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"d13d06f": {
"balancing:strategy": "none",
@@ -688,7 +833,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.01896761231951133,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"d95f74b": {
"balancing:strategy": "none",
@@ -709,7 +859,12 @@
"classifier:random_forest:min_weight_fraction_leaf": 0.0,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1456,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"7167f0c": {
"balancing:strategy": "none",
@@ -734,7 +889,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.05789857129625832,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
}
},
"cutoffs": [
diff --git a/autosklearn/experimental/roc_auc/askl2_portfolios/RF_SH-eta4-i_5CV_iterative_es_if.json b/autosklearn/experimental/roc_auc/askl2_portfolios/RF_SH-eta4-i_5CV_iterative_es_if.json
index 6580310dcb..0136a3bd2b 100644
--- a/autosklearn/experimental/roc_auc/askl2_portfolios/RF_SH-eta4-i_5CV_iterative_es_if.json
+++ b/autosklearn/experimental/roc_auc/askl2_portfolios/RF_SH-eta4-i_5CV_iterative_es_if.json
@@ -23,7 +23,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
"classifier:gradient_boosting:n_iter_no_change": 14,
"classifier:gradient_boosting:validation_fraction": 0.12446922535601551,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"48e7d72": {
"balancing:strategy": "none",
@@ -51,7 +56,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1633,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"3b5725c": {
"balancing:strategy": "none",
@@ -70,7 +80,12 @@
"classifier:random_forest:min_samples_leaf": 7,
"classifier:random_forest:min_samples_split": 6,
"classifier:random_forest:min_weight_fraction_leaf": 0.0,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"e8567fc": {
"balancing:strategy": "weighting",
@@ -88,7 +103,12 @@
"classifier:sgd:penalty": "l1",
"classifier:sgd:tol": 0.0007858702630144828,
"classifier:sgd:eta0": 1.3077397138182272e-06,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"d3b1c98": {
"balancing:strategy": "none",
@@ -110,7 +130,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.0023870347307066147,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.8079255643542401,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.2778427452756853,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"49b696b": {
"balancing:strategy": "none",
@@ -132,7 +157,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.16529056183915525,
"classifier:gradient_boosting:n_iter_no_change": 3,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"766431d": {
"balancing:strategy": "none",
@@ -153,7 +183,12 @@
"classifier:sgd:eta0": 0.06268029581129594,
"classifier:sgd:l1_ratio": 1.2408983312626301e-05,
"classifier:sgd:power_t": 0.15473248271262038,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"901862d": {
"balancing:strategy": "weighting",
@@ -178,7 +213,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.00018198102665192584,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"9454294": {
"balancing:strategy": "weighting",
@@ -205,7 +245,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.8593841208287082,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.27443311984781915,
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"ed48b1e": {
"balancing:strategy": "weighting",
@@ -231,7 +276,12 @@
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.19593845967764903,
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"7eb153b": {
"balancing:strategy": "weighting",
@@ -255,7 +305,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.1023091397829033,
"classifier:gradient_boosting:n_iter_no_change": 3,
"classifier:gradient_boosting:validation_fraction": 0.20375603224425706,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"d41cebb": {
"balancing:strategy": "none",
@@ -282,7 +337,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.014993536738242566,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1006,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"c3c5a39": {
"balancing:strategy": "none",
@@ -304,7 +364,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"classifier:gradient_boosting:n_iter_no_change": 13,
"classifier:gradient_boosting:validation_fraction": 0.09536739955550967,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"09555d8": {
"balancing:strategy": "none",
@@ -326,7 +391,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.0016129570857335443,
"classifier:gradient_boosting:n_iter_no_change": 12,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"c9ed268": {
"balancing:strategy": "none",
@@ -346,7 +416,12 @@
"classifier:extra_trees:min_samples_split": 2,
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.28925175872648934,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"07b7a65": {
"balancing:strategy": "weighting",
@@ -367,7 +442,12 @@
"classifier:gradient_boosting:scoring": "loss",
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.09025942637402559,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"fd4841b": {
"balancing:strategy": "weighting",
@@ -391,7 +471,12 @@
"classifier:mlp:shuffle": "True",
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"a9c54cc": {
"balancing:strategy": "none",
@@ -413,7 +498,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.0025534629798769164,
"classifier:gradient_boosting:n_iter_no_change": 19,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"d2a3f8c": {
"balancing:strategy": "none",
@@ -440,7 +530,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.010000000000000004,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1074,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"36cd396": {
"balancing:strategy": "weighting",
@@ -465,7 +560,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"ed2b3cf": {
"balancing:strategy": "weighting",
@@ -486,7 +586,12 @@
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1979,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"aeea657": {
"balancing:strategy": "weighting",
@@ -514,7 +619,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 968,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"1d4a4c2": {
"balancing:strategy": "weighting",
@@ -535,7 +645,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 775,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "uniform",
"classifier:sgd:eta0": 0.08292033673052085,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"5dbf9b3": {
"balancing:strategy": "weighting",
@@ -561,7 +676,12 @@
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.8390771744382532,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.16778149289904426,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"225eba5": {
"balancing:strategy": "none",
@@ -588,7 +708,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.11290792173552132,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1206,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"51f697f": {
"balancing:strategy": "weighting",
@@ -607,7 +732,12 @@
"classifier:extra_trees:min_samples_leaf": 2,
"classifier:extra_trees:min_samples_split": 4,
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"b84962e": {
"balancing:strategy": "weighting",
@@ -629,7 +759,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.006793803631342036,
"classifier:gradient_boosting:n_iter_no_change": 10,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"05ee5be": {
"balancing:strategy": "none",
@@ -654,7 +789,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.0018907050183624963,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"0ca3806": {
"balancing:strategy": "weighting",
@@ -682,7 +822,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.8684897085540992,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.16274635771642468,
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"015ec3b": {
"balancing:strategy": "none",
@@ -704,7 +849,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.1090812196521665,
"classifier:gradient_boosting:n_iter_no_change": 8,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"53af782": {
"balancing:strategy": "none",
@@ -730,7 +880,12 @@
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.7158581239252029,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.23727106534304226,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"04ee8cd": {
"balancing:strategy": "weighting",
@@ -752,7 +907,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.010000000000000004,
"classifier:gradient_boosting:n_iter_no_change": 14,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
}
},
"cutoffs": [
diff --git a/autosklearn/experimental/roc_auc/askl2_portfolios/RF_SH-eta4-i_holdout_iterative_es_if.json b/autosklearn/experimental/roc_auc/askl2_portfolios/RF_SH-eta4-i_holdout_iterative_es_if.json
index 3c444db186..727752ba57 100644
--- a/autosklearn/experimental/roc_auc/askl2_portfolios/RF_SH-eta4-i_holdout_iterative_es_if.json
+++ b/autosklearn/experimental/roc_auc/askl2_portfolios/RF_SH-eta4-i_holdout_iterative_es_if.json
@@ -20,7 +20,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 467,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"c85132e": {
"balancing:strategy": "weighting",
@@ -45,7 +50,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.006864495761122583,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"3ffac3d": {
"balancing:strategy": "weighting",
@@ -65,7 +75,12 @@
"classifier:extra_trees:min_samples_split": 12,
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.011646685244486026,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"3861869": {
"balancing:strategy": "weighting",
@@ -89,7 +104,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.7621841297339562,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.2689588813833072,
"classifier:gradient_boosting:n_iter_no_change": 19,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"1d8ad41": {
"balancing:strategy": "none",
@@ -112,7 +132,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.010000000000000004,
"classifier:gradient_boosting:n_iter_no_change": 2,
"classifier:gradient_boosting:validation_fraction": 0.35512720315660784,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"e8c14b9": {
"balancing:strategy": "weighting",
@@ -134,7 +159,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.012653276795134658,
"classifier:gradient_boosting:n_iter_no_change": 7,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"93a94fc": {
"balancing:strategy": "none",
@@ -159,7 +189,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"ee1d407": {
"balancing:strategy": "weighting",
@@ -180,7 +215,12 @@
"classifier:gradient_boosting:scoring": "loss",
"classifier:gradient_boosting:tol": 1e-07,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.0030968404387141115,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"4464ce4": {
"balancing:strategy": "none",
@@ -199,7 +239,12 @@
"classifier:extra_trees:min_samples_leaf": 5,
"classifier:extra_trees:min_samples_split": 10,
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"5da6d7e": {
"balancing:strategy": "none",
@@ -222,7 +267,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.02106255140796113,
"classifier:gradient_boosting:n_iter_no_change": 14,
"classifier:gradient_boosting:validation_fraction": 0.1034846810050602,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"4522c5c": {
"balancing:strategy": "none",
@@ -248,7 +298,12 @@
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.0031947270899377388,
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"cbcd40f": {
"balancing:strategy": "none",
@@ -272,7 +327,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1229,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
"classifier:gradient_boosting:n_iter_no_change": 15,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"7166357": {
"balancing:strategy": "weighting",
@@ -291,7 +351,12 @@
"classifier:extra_trees:min_samples_leaf": 1,
"classifier:extra_trees:min_samples_split": 3,
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"3a1cf69": {
"balancing:strategy": "weighting",
@@ -316,7 +381,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"e99d884": {
"balancing:strategy": "none",
@@ -341,7 +411,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.14403104186842322,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"1312886": {
"balancing:strategy": "weighting",
@@ -365,7 +440,12 @@
"classifier:mlp:shuffle": "True",
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"59fbdef": {
"balancing:strategy": "none",
@@ -383,7 +463,12 @@
"classifier:sgd:penalty": "l1",
"classifier:sgd:tol": 8.27706111772171e-05,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.007965405034296472,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"33ce15c": {
"balancing:strategy": "none",
@@ -405,7 +490,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.2974843930412059,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 817,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"e72a0d0": {
"balancing:strategy": "none",
@@ -428,7 +518,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.007883404564380745,
"classifier:gradient_boosting:n_iter_no_change": 1,
"classifier:gradient_boosting:validation_fraction": 0.058283078146253935,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"5ec319d": {
"balancing:strategy": "none",
@@ -450,7 +545,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.0059309332615295795,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_max": 0.7122245303069001,
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.26765391722955484,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"3230065": {
"balancing:strategy": "none",
@@ -475,7 +575,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"4a96439": {
"balancing:strategy": "none",
@@ -494,7 +599,12 @@
"classifier:extra_trees:min_samples_leaf": 1,
"classifier:extra_trees:min_samples_split": 5,
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"970f3cb": {
"balancing:strategy": "none",
@@ -520,7 +630,12 @@
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 70,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"3e86168": {
"balancing:strategy": "none",
@@ -544,7 +659,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.25,
"classifier:gradient_boosting:n_iter_no_change": 14,
"classifier:gradient_boosting:validation_fraction": 0.3430240407684142,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"6e537da": {
"balancing:strategy": "none",
@@ -569,7 +689,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"e5d8521": {
"balancing:strategy": "weighting",
@@ -596,7 +721,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 445,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
"classifier:mlp:validation_fraction": 0.1,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"ac7f6c4": {
"balancing:strategy": "weighting",
@@ -619,7 +749,12 @@
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.010000000000000004,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:n_quantiles": 1000,
"data_preprocessor:feature_type:numerical_transformer:rescaling:quantile_transformer:output_distribution": "normal",
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"f1d04fb": {
"balancing:strategy": "none",
@@ -644,7 +779,12 @@
"classifier:mlp:solver": "adam",
"classifier:mlp:tol": 0.0001,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.008845358970001308,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"191b6ed": {
"balancing:strategy": "none",
@@ -666,7 +806,12 @@
"classifier:gradient_boosting:tol": 1e-07,
"classifier:gradient_boosting:n_iter_no_change": 10,
"classifier:gradient_boosting:validation_fraction": 0.17784647507533755,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"9c8ea9e": {
"balancing:strategy": "weighting",
@@ -685,7 +830,12 @@
"classifier:extra_trees:min_samples_leaf": 8,
"classifier:extra_trees:min_samples_split": 16,
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"842e6d9": {
"balancing:strategy": "none",
@@ -709,7 +859,12 @@
"data_preprocessor:feature_type:numerical_transformer:rescaling:robust_scaler:q_min": 0.2221036462435928,
"classifier:gradient_boosting:n_iter_no_change": 16,
"classifier:gradient_boosting:validation_fraction": 0.0796022383640014,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
},
"c867baf": {
"balancing:strategy": "weighting",
@@ -729,7 +884,12 @@
"classifier:extra_trees:min_samples_split": 20,
"classifier:extra_trees:min_weight_fraction_leaf": 0.0,
"data_preprocessor:feature_type:categorical_transformer:category_coalescence:minority_coalescer:minimum_fraction": 0.01973840362555204,
- "data_preprocessor:__choice__": "feature_type"
+ "data_preprocessor:__choice__": "feature_type",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:__choice__": "bag_of_word_encoding",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_absolute": 1,
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:min_df_choice": "min_df_absolute",
+ "data_preprocessor:feature_type:text_transformer:text_encoding:bag_of_word_encoding:ngram_upper_bound": 1,
+ "data_preprocessor:feature_type:text_transformer:text_feature_reduction:n_components": 100
}
},
"cutoffs": [
diff --git a/autosklearn/experimental/selector.py b/autosklearn/experimental/selector.py
index 3ef681bef8..b854c7b440 100644
--- a/autosklearn/experimental/selector.py
+++ b/autosklearn/experimental/selector.py
@@ -1,6 +1,7 @@
+import typing
+
import copy
import itertools
-import typing
import numpy as np
import pandas as pd
@@ -9,7 +10,6 @@
class AbstractSelector:
-
def fit(
self,
X: pd.DataFrame,
@@ -19,18 +19,22 @@ def fit(
) -> None:
raise NotImplementedError()
- def predict(self, X: pd.DataFrame, y: typing.Optional[pd.DataFrame] = None) -> pd.DataFrame:
+ def predict(
+ self, X: pd.DataFrame, y: typing.Optional[pd.DataFrame] = None
+ ) -> pd.DataFrame:
prediction = self._predict(X, y)
for col, series in prediction.iteritems():
assert series.dtype == float, (col, series)
np.testing.assert_array_almost_equal(
- prediction.sum(axis='columns').to_numpy(),
+ prediction.sum(axis="columns").to_numpy(),
np.ones(X.shape[0]),
err_msg=prediction.to_csv(),
)
return prediction
- def _predict(self, X: pd.DataFrame, y: typing.Optional[pd.DataFrame]) -> pd.DataFrame:
+ def _predict(
+ self, X: pd.DataFrame, y: typing.Optional[pd.DataFrame]
+ ) -> pd.DataFrame:
raise NotImplementedError()
@@ -68,43 +72,60 @@ def fit(
weights[i] = dict()
for j in range(i + 1, len(target_indices)):
- if self.configuration['normalization'] in ('all', 'binary', 'y', 'all1',
- 'binary1'):
+ if self.configuration["normalization"] in (
+ "all",
+ "binary",
+ "y",
+ "all1",
+ "binary1",
+ ):
minimum2 = np.ones(len(X)) * np.inf
maximum2 = np.zeros(len(X))
- if self.configuration['normalization'] in ('all', 'all1'):
+ if self.configuration["normalization"] in ("all", "all1"):
for idx, task_id in enumerate(X.index):
for method_id in range(len(target_indices)):
- minimum2[idx] = np.nanmin((
- minimum2[idx],
- minima[task_id][self.strategies_[method_id]]
- ))
- maximum2[idx] = np.nanmax((
- maximum2[idx],
- maxima[task_id][self.strategies_[method_id]]
- ))
- if self.configuration['normalization'] == 'all1':
+ minimum2[idx] = np.nanmin(
+ (
+ minimum2[idx],
+ minima[task_id][self.strategies_[method_id]],
+ )
+ )
+ maximum2[idx] = np.nanmax(
+ (
+ maximum2[idx],
+ maxima[task_id][self.strategies_[method_id]],
+ )
+ )
+ if self.configuration["normalization"] == "all1":
maximum2 = np.ones_like(maximum2)
- elif self.configuration['normalization'] in ('binary', 'binary1'):
+ elif self.configuration["normalization"] in ("binary", "binary1"):
for idx, task_id in enumerate(X.index):
for method_id in (i, j):
- minimum2[idx] = np.nanmin((
- minimum2[idx],
- minima[task_id][self.strategies_[method_id]]
- ))
- maximum2[idx] = np.nanmax((
- maximum2[idx],
- maxima[task_id][self.strategies_[method_id]]
- ))
- if self.configuration['normalization'] == 'binary1':
+ minimum2[idx] = np.nanmin(
+ (
+ minimum2[idx],
+ minima[task_id][self.strategies_[method_id]],
+ )
+ )
+ maximum2[idx] = np.nanmax(
+ (
+ maximum2[idx],
+ maxima[task_id][self.strategies_[method_id]],
+ )
+ )
+ if self.configuration["normalization"] == "binary1":
maximum2 = np.ones_like(maximum2)
- elif self.configuration['normalization'] == 'y':
+ elif self.configuration["normalization"] == "y":
for idx, task_id in enumerate(X.index):
- minimum2[idx] = np.nanmin((minimum2[idx], y_pd.loc[task_id].min()))
- maximum2[idx] = np.nanmax((maximum2[idx], y_pd.loc[task_id].max()))
+ minimum2[idx] = np.nanmin(
+ (minimum2[idx], y_pd.loc[task_id].min())
+ )
+ maximum2[idx] = np.nanmax(
+ (maximum2[idx], y_pd.loc[task_id].max())
+ )
else:
- raise ValueError(self.configuration['normalization'])
+ raise ValueError(self.configuration["normalization"])
y_i_j = y[:, i] < y[:, j]
mask = np.isfinite(y[:, i]) & np.isfinite(y[:, j])
@@ -121,7 +142,7 @@ def fit(
weights_i_j = np.abs(normalized_y_i - normalized_y_j)
- elif self.configuration['normalization'] == 'rank':
+ elif self.configuration["normalization"] == "rank":
y_i_j = y[:, i] < y[:, j]
mask = np.isfinite(y[:, i]) & np.isfinite(y[:, j])
X_ = X.to_numpy()[mask]
@@ -129,7 +150,7 @@ def fit(
ranks = scipy.stats.rankdata(y[mask], axis=1)
weights_i_j = np.abs(ranks[:, i] - ranks[:, j])
- elif self.configuration['normalization'] == 'None':
+ elif self.configuration["normalization"] == "None":
y_i_j = y[:, i] < y[:, j]
mask = np.isfinite(y[:, i]) & np.isfinite(y[:, j])
X_ = X.to_numpy()[mask]
@@ -137,7 +158,7 @@ def fit(
weights_i_j = np.ones_like(y_i_j).astype(int)
else:
- raise ValueError(self.configuration['normalization'])
+ raise ValueError(self.configuration["normalization"])
if len(y_i_j) == 0:
models[i][j] = None
@@ -148,21 +169,25 @@ def fit(
n_zeros = int(np.ceil(len(y_i_j) / 2))
n_ones = int(np.floor(len(y_i_j) / 2))
import sklearn.dummy
- base_model = sklearn.dummy.DummyClassifier(strategy='constant',
- constant=y_i_j[0])
+
+ base_model = sklearn.dummy.DummyClassifier(
+ strategy="constant", constant=y_i_j[0]
+ )
base_model.fit(
X_,
np.array(([[0]] * n_zeros) + ([[1]] * n_ones)).flatten(),
sample_weight=weights_i_j,
)
else:
- if self.configuration.get('max_depth') == 0:
+ if self.configuration.get("max_depth") == 0:
import sklearn.dummy
+
loss_i = np.sum((y_i_j == 0) * weights_i_j)
loss_j = np.sum((y_i_j == 1) * weights_i_j)
base_model = sklearn.dummy.DummyClassifier(
- strategy='constant', constant=1 if loss_i < loss_j else 0,
+ strategy="constant",
+ constant=1 if loss_i < loss_j else 0,
)
base_model.fit(
X_,
@@ -171,7 +196,11 @@ def fit(
)
else:
base_model = self.fit_pairwise_model(
- X_, y_i_j, weights_i_j, self.rng, self.configuration,
+ X_,
+ y_i_j,
+ weights_i_j,
+ self.rng,
+ self.configuration,
)
models[i][j] = base_model
weights[i][j] = weights_i_j
@@ -179,7 +208,9 @@ def fit(
self.weights_ = weights
self.target_indices = target_indices
- def _predict(self, X: pd.DataFrame, y: typing.Optional[pd.DataFrame]) -> pd.DataFrame:
+ def _predict(
+ self, X: pd.DataFrame, y: typing.Optional[pd.DataFrame]
+ ) -> pd.DataFrame:
if y is not None:
raise ValueError("y must not be provided")
@@ -193,7 +224,9 @@ def _predict(self, X: pd.DataFrame, y: typing.Optional[pd.DataFrame]) -> pd.Data
raw_probas[(i, j)] = self.models[i][j].predict_proba(X)
if len(raw_predictions) == 0:
- predictions = pd.DataFrame(0, index=X.index, columns=self.strategies_).astype(float)
+ predictions = pd.DataFrame(
+ 0, index=X.index, columns=self.strategies_
+ ).astype(float)
predictions.iloc[:, self.single_strategy_idx] = 1.0
return predictions
@@ -203,21 +236,21 @@ def _predict(self, X: pd.DataFrame, y: typing.Optional[pd.DataFrame]) -> pd.Data
for i in range(len(self.target_indices)):
for j in range(i + 1, len(self.target_indices)):
if (i, j) in raw_predictions:
- if self.configuration['prediction'] == 'soft':
+ if self.configuration["prediction"] == "soft":
if raw_probas[(i, j)].shape[1] == 1:
proba = raw_probas[(i, j)][x_idx][0]
else:
proba = raw_probas[(i, j)][x_idx][1]
wins[i] += proba
wins[j] += 1 - proba
- elif self.configuration['prediction'] == 'hard':
+ elif self.configuration["prediction"] == "hard":
prediction = raw_predictions[(i, j)][x_idx]
if prediction == 1:
wins[i] += 1
else:
wins[j] += 1
else:
- raise ValueError(self.configuration['prediction'])
+ raise ValueError(self.configuration["prediction"])
n_prev = np.inf
# Tie breaking
@@ -236,7 +269,9 @@ def _predict(self, X: pd.DataFrame, y: typing.Optional[pd.DataFrame]) -> pd.Data
hit = True
break
if not hit:
- wins[int(self.rng.choice(np.argwhere(most_wins_mask).flatten()))] += 1
+ wins[
+ int(self.rng.choice(np.argwhere(most_wins_mask).flatten()))
+ ] += 1
elif np.sum(most_wins_mask) > 1:
n_prev = np.sum(most_wins_mask)
where = np.argwhere(most_wins_mask).flatten()
@@ -250,10 +285,9 @@ def _predict(self, X: pd.DataFrame, y: typing.Optional[pd.DataFrame]) -> pd.Data
else:
method_i = self.strategies_[i]
method_j = self.strategies_[j]
- if (
- self.tie_break_order.index(method_i)
- < self.tie_break_order.index(method_j)
- ):
+ if self.tie_break_order.index(
+ method_i
+ ) <