demo/wine_classification.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Intro"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<img width=600 src=\"images/wine.jpg\">\n",
    "<p>This notebook demonstrates solving a classification problem on the example of the <a href=\"https://archive.ics.uci.edu/ml/datasets/Wine\">Wine Data Set</a>. The dataset contains chemical analysis of wines from three different cultivars. We will train models that will be able to predict to which cultivar does a certain wine belong.</p>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Import dependencies"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "\n",
    "import sys\n",
    "sys.path.insert(0, \"..\")\n",
    "import mymllib\n",
    "\n",
    "%matplotlib inline\n",
    "sns.set(font_scale=1.2)\n",
    "\n",
    "PALETTE = \"deep\""
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Exploratory data analysis"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Download the dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Class</th>\n",
       "      <th>Alcohol</th>\n",
       "      <th>Malic acid</th>\n",
       "      <th>Ash</th>\n",
       "      <th>Alcalinity of ash</th>\n",
       "      <th>Magnesium</th>\n",
       "      <th>Total phenols</th>\n",
       "      <th>Flavanoids</th>\n",
       "      <th>Nonflavanoid phenols</th>\n",
       "      <th>Proanthocyanins</th>\n",
       "      <th>Color intensity</th>\n",
       "      <th>Hue</th>\n",
       "      <th>OD280/OD315 of diluted wines</th>\n",
       "      <th>Proline</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>14.23</td>\n",
       "      <td>1.71</td>\n",
       "      <td>2.43</td>\n",
       "      <td>15.6</td>\n",
       "      <td>127</td>\n",
       "      <td>2.80</td>\n",
       "      <td>3.06</td>\n",
       "      <td>0.28</td>\n",
       "      <td>2.29</td>\n",
       "      <td>5.64</td>\n",
       "      <td>1.04</td>\n",
       "      <td>3.92</td>\n",
       "      <td>1065</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>13.20</td>\n",
       "      <td>1.78</td>\n",
       "      <td>2.14</td>\n",
       "      <td>11.2</td>\n",
       "      <td>100</td>\n",
       "      <td>2.65</td>\n",
       "      <td>2.76</td>\n",
       "      <td>0.26</td>\n",
       "      <td>1.28</td>\n",
       "      <td>4.38</td>\n",
       "      <td>1.05</td>\n",
       "      <td>3.40</td>\n",
       "      <td>1050</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>13.16</td>\n",
       "      <td>2.36</td>\n",
       "      <td>2.67</td>\n",
       "      <td>18.6</td>\n",
       "      <td>101</td>\n",
       "      <td>2.80</td>\n",
       "      <td>3.24</td>\n",
       "      <td>0.30</td>\n",
       "      <td>2.81</td>\n",
       "      <td>5.68</td>\n",
       "      <td>1.03</td>\n",
       "      <td>3.17</td>\n",
       "      <td>1185</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1</td>\n",
       "      <td>14.37</td>\n",
       "      <td>1.95</td>\n",
       "      <td>2.50</td>\n",
       "      <td>16.8</td>\n",
       "      <td>113</td>\n",
       "      <td>3.85</td>\n",
       "      <td>3.49</td>\n",
       "      <td>0.24</td>\n",
       "      <td>2.18</td>\n",
       "      <td>7.80</td>\n",
       "      <td>0.86</td>\n",
       "      <td>3.45</td>\n",
       "      <td>1480</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1</td>\n",
       "      <td>13.24</td>\n",
       "      <td>2.59</td>\n",
       "      <td>2.87</td>\n",
       "      <td>21.0</td>\n",
       "      <td>118</td>\n",
       "      <td>2.80</td>\n",
       "      <td>2.69</td>\n",
       "      <td>0.39</td>\n",
       "      <td>1.82</td>\n",
       "      <td>4.32</td>\n",
       "      <td>1.04</td>\n",
       "      <td>2.93</td>\n",
       "      <td>735</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Class  Alcohol  Malic acid   Ash  Alcalinity of ash  Magnesium  \\\n",
       "0      1    14.23        1.71  2.43               15.6        127   \n",
       "1      1    13.20        1.78  2.14               11.2        100   \n",
       "2      1    13.16        2.36  2.67               18.6        101   \n",
       "3      1    14.37        1.95  2.50               16.8        113   \n",
       "4      1    13.24        2.59  2.87               21.0        118   \n",
       "\n",
       "   Total phenols  Flavanoids  Nonflavanoid phenols  Proanthocyanins  \\\n",
       "0           2.80        3.06                  0.28             2.29   \n",
       "1           2.65        2.76                  0.26             1.28   \n",
       "2           2.80        3.24                  0.30             2.81   \n",
       "3           3.85        3.49                  0.24             2.18   \n",
       "4           2.80        2.69                  0.39             1.82   \n",
       "\n",
       "   Color intensity   Hue  OD280/OD315 of diluted wines  Proline  \n",
       "0             5.64  1.04                          3.92     1065  \n",
       "1             4.38  1.05                          3.40     1050  \n",
       "2             5.68  1.03                          3.17     1185  \n",
       "3             7.80  0.86                          3.45     1480  \n",
       "4             4.32  1.04                          2.93      735  "
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "columns = [\"Class\", \"Alcohol\", \"Malic acid\", \"Ash\", \"Alcalinity of ash\", \"Magnesium\",\n",
    "           \"Total phenols\", \"Flavanoids\", \"Nonflavanoid phenols\", \"Proanthocyanins\",\n",
    "           \"Color intensity\", \"Hue\", \"OD280/OD315 of diluted wines\", \"Proline\"]\n",
    "dataset = pd.read_table(\n",
    "    \"https://archive.ics.uci.edu/ml/machine-learning-databases/wine/wine.data\",\n",
    "    names=columns, sep=\",\", na_values = \"?\")\n",
    "dataset.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Check the dataset for missing data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As we can see, there is no missing values in the dataset:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Class                           0\n",
       "Alcohol                         0\n",
       "Malic acid                      0\n",
       "Ash                             0\n",
       "Alcalinity of ash               0\n",
       "Magnesium                       0\n",
       "Total phenols                   0\n",
       "Flavanoids                      0\n",
       "Nonflavanoid phenols            0\n",
       "Proanthocyanins                 0\n",
       "Color intensity                 0\n",
       "Hue                             0\n",
       "OD280/OD315 of diluted wines    0\n",
       "Proline                         0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dataset.isna().sum()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Select features and target"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We select the Class column as the target value and all other columns as features:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Target: Class\n",
      "Features: ['Alcohol', 'Malic acid', 'Ash', 'Alcalinity of ash', 'Magnesium', 'Total phenols', 'Flavanoids', 'Nonflavanoid phenols', 'Proanthocyanins', 'Color intensity', 'Hue', 'OD280/OD315 of diluted wines', 'Proline']\n"
     ]
    }
   ],
   "source": [
    "target_column = columns[0]\n",
    "feature_columns = columns[1:]\n",
    "\n",
    "print(\"Target:\", target_column)\n",
    "print(\"Features:\", feature_columns)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Train/Test split"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Here we use a simple 80%/20% train/test split:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train dataset size: 142\n",
      "Test dataset size: 36\n"
     ]
    }
   ],
   "source": [
    "train_dataset = dataset.sample(frac=0.8, random_state=123)\n",
    "test_dataset = dataset.drop(train_dataset.index)\n",
    "\n",
    "print(\"Train dataset size:\", train_dataset.shape[0])\n",
    "print(\"Test dataset size:\", test_dataset.shape[0])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Classes examples count"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's compare number of examples of each class:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "class_counts = train_dataset[target_column].value_counts()\n",
    "plt.title(\"Number of exmaples by classes\")\n",
    "sns.barplot(class_counts.index, class_counts, palette=PALETTE)\n",
    "_ = plt.ylabel(\"Count\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As we can see from the plot, classes have different number of examples and thus the dataset is not stricktly balanced. However difference between the classes is not too big and shouldn't cause any significant problems."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Visualization with PCA"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "To visualize the data, its dimensionality should be reduced to 2 dimension, which can be achieved with PCA. For PCA to work properly the data should be normalized:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "data_scaler = mymllib.preprocessing.DataScaler().fit(train_dataset[feature_columns])\n",
    "normalized_features = data_scaler.scale(train_dataset[feature_columns], by_std=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's run PCA. As we can see, the retained variance of two principal components is equal to 99.98% which makes them a great represantation of the data:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Retained variance: 0.9998390273646086\n"
     ]
    }
   ],
   "source": [
    "pca = mymllib.dim_reduction.PCA(2).fit(normalized_features)\n",
    "print(\"Retained variance:\", sum(pca.retained_variance))\n",
    "features_2D = pca.transform(normalized_features)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now when we have a 2D representation of our data we can plot it and check how examples of different classes are located with respect to each other:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1008x504 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(14, 7))\n",
    "plt.title(\"Classes scatter plot\")\n",
    "sns.scatterplot(x=features_2D[:, 0], y=features_2D[:, 1],\n",
    "                hue=train_dataset[target_column], legend=\"full\", palette=PALETTE,\n",
    "                s=100)\n",
    "plt.xlabel(\"PC1\")\n",
    "_ = plt.ylabel(\"PC2\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The scatter plot shows that examples from different classes are not linearly separable, which means that a model would need to learn a nonlinear function to perform better on this dataset. We will try out two models that are capable of doing this: a logistic regression with polynomial features and a neural network with a hidden layer."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Models"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Features/target split"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We split both train and test datasets into X (features) and y (target) to make them suitable for training and testing models:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train = train_dataset[feature_columns]\n",
    "y_train = train_dataset[target_column]\n",
    "\n",
    "X_test = test_dataset[feature_columns]\n",
    "y_test = test_dataset[target_column]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Logistic regression"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As were shown above, examples of different classes are not linearly separable and thus for logistic regression to better classify them them we need to create polynomial features:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train_poly = mymllib.preprocessing.add_polynomial(X_train, 2)\n",
    "X_test_poly = mymllib.preprocessing.add_polynomial(X_test, 2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We also normalize the data so that all features would have a mean of 0 and a variance of 1:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "data_scaler = mymllib.preprocessing.DataScaler().fit(X_train_poly)\n",
    "X_train_norm = data_scaler.scale(X_train_poly)\n",
    "X_test_norm = data_scaler.scale(X_test_poly)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now we can train the model:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "logistic_regression = mymllib.regression.LogisticRegression(\n",
    "    regularization_param=0.1)\n",
    "logistic_regression.fit(X_train_norm, y_train)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Lets test the model's performance both on train and test sets. Mean absolute error (MAE) and mean absolute percentage error (MAPE) are used as metrics:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train F1 score: 1.0\n",
      "Test F1 score: 0.9726830903301491\n"
     ]
    }
   ],
   "source": [
    "import mymllib.metrics.classification as metrics\n",
    "\n",
    "def test_model(model, X_train, y_train, X_test, y_test):\n",
    "    y_train_pred = model.predict(X_train)\n",
    "    y_test_pred = model.predict(X_test)\n",
    "    \n",
    "    print(\"Train F1 score:\", metrics.f1_score(\n",
    "        y_train, y_train_pred))\n",
    "    print(\"Test F1 score:\", metrics.f1_score(\n",
    "        y_test, y_test_pred))\n",
    "\n",
    "test_model(logistic_regression, X_train_norm, y_train, X_test_norm, y_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Neural network"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We will use a neural network with a hidden layer to capture non-linear dependencies, so we can skip creating polynomial features and just perform normalization:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "data_scaler = mymllib.preprocessing.DataScaler().fit(X_train)\n",
    "X_train_norm = data_scaler.scale(X_train)\n",
    "X_test_norm = data_scaler.scale(X_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Here we train and test a feed-forward fully-connected neural network with a single hidden layer:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train F1 score: 1.0\n",
      "Test F1 score: 0.9726830903301491\n"
     ]
    }
   ],
   "source": [
    "neural_net = mymllib.neural_networks.ClassificationNeuralNetwork(\n",
    "    hidden_layers=(7,),\n",
    "    regularization_param=0.1)\n",
    "neural_net.fit(X_train_norm, y_train)\n",
    "test_model(neural_net, X_train_norm, y_train, X_test_norm, y_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Conclusion"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Both models performed similarly, achieving 1.0 train and 0.97 test F1 score. Such results indicate overfitting, so the test score can be further improved by increasing a regularization coefficient, decreasing complexity of the models, etc. However, this is beyond the scope of this notebook since its main goal is to show how to use the library for classification and demonstrate that it works as expected, which was successfully done."
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.10"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}