Fake news detection.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Importing Libraries"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[nltk_data] Downloading package stopwords to\n",
      "[nltk_data]     C:\\Users\\siddh\\AppData\\Roaming\\nltk_data...\n",
      "[nltk_data]   Package stopwords is already up-to-date!\n",
      "[nltk_data] Downloading package punkt to\n",
      "[nltk_data]     C:\\Users\\siddh\\AppData\\Roaming\\nltk_data...\n",
      "[nltk_data]   Package punkt is already up-to-date!\n"
     ]
    }
   ],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "import time\n",
    "\n",
    "# For cleaning and processing\n",
    "import json\n",
    "import os, re\n",
    "from random import randint\n",
    "from pathlib import Path\n",
    "from sklearn.feature_extraction.text import TfidfVectorizer\n",
    "from nltk import sent_tokenize\n",
    "import nltk\n",
    "import string\n",
    "from nltk.corpus import stopwords\n",
    "from nltk.stem import SnowballStemmer\n",
    "nltk.download('stopwords')\n",
    "nltk.download('punkt')\n",
    "\n",
    "# For Modeling and Computing Accuracy\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.metrics import accuracy_score, classification_report, confusion_matrix\n",
    "from sklearn.metrics.pairwise import cosine_similarity\n",
    "# !pip install xgboost\n",
    "import xgboost\n",
    "from xgboost import XGBClassifier\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.svm import LinearSVC\n",
    "from sklearn.feature_selection import SelectPercentile, f_classif\n",
    "from scipy.sparse import hstack"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Importing the Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Claim: A line from George Orwell's novel 1984 predicts the power of smartphones.\n",
      "Speaker: \n",
      "Date: 2017-07-17\n",
      "Related Article Ids: [122094, 122580, 130685, 134765]\n",
      "First evidence article id: 122094\n",
      "1984: George Orwell predicted 2017 almost 70 years ago\n",
      "April, 1984. 13:00. Comrade 6079, Winston Smith, thinks a thought, starts a diary, and falls in love. But Big Brother is always watching. Set in a world where an invasive government keeps a malevolently watchful eye on its citizens, this radical and much-lauded staging explores surveillance, identity and why Orwell?s vision of the future is as relevant now as ever. Now with a brand new Australian cast, we bring adapter-directors Robert Icke and Duncan Macmillan?s thrilling vision back to Australia for a national tour.\n",
      "\n",
      "THERE was no crystal ball, no religious apparitions or voices from the dead when George Orwell wrote Nineteen Eighty-Four, accurately depicted our present almost 70 years ago.\n",
      "\n",
      "It’s no coincidence that Orwell’s 1949 novel had climbed back into multiple bestseller lists this year — people have realised its scarily accurate foretelling.\n",
      "\n",
      "In 2013 its relevance — still strong and worryingly reliable — led to the West End smash hit adaptation of the book, which is currently showing in Australia.\n",
      "\n",
      "The book is one of the greatest dystopian novels every written. It looks at a future where people are controlled into what to think, how to act and how to live by the Government known as Big Brother. It uses telescreens, fearmongering, media control and corruption to control the masses. The main protagonist is the initially apathetic Winston who craves something more than the controlled world he inhabits.\n",
      "\n",
      "Corey McMahon, the associate director of the 1984 production in Australia, said the book’s reach and resonance will always continue as it speaks to more than a literary audience.\n",
      "\n",
      "“It speaks to a world filled with unpopular policies and world leaders,” McMahon said.\n",
      "\n",
      "The production takes a deeper look into some of the core themes of the novel such as governmental surveillance through the subconscious co-operation of the public through their screens. You are reading this very line on one of your ‘screens’.\n",
      "\n",
      "McMahon said the production looks to act as a sort of catalyst for its audiences to look at the world in a different light. To question what is true and what isn’t, to look beyond the political spin.\n",
      "\n",
      "“We are so busy Facebooking, Snapchatting, Tweeting etc that we are missing what is happening in the world,” he said.\n",
      "\n",
      "He pointed to a quote from the production he said not only sticks out to him every time, but one that he think will challenge audiences to really think about the world they inhabit and how they chose to be a part of it.\n",
      "\n",
      "The line is from one of the characters that works for the Government, otherwise known as Big Brother. He says: “The people will not revolt. They will not look up from their screens long enough to notice what’s happening.”\n",
      "\n",
      "Orwell didn’t have a crystal ball, what he did have was an understanding of the human condition and its weakness. Apathy.\n",
      "\n",
      "1984 is holding shows across Australia. You can purchase tickets here.\n"
     ]
    }
   ],
   "source": [
    "# These are the file paths from where the dataset will be accesses\n",
    "METADATA_FILEPATH = Path('C:/train.json')  # To change with your filepath\n",
    "ARTICLES_FILEPATH = Path('C:/train_articles')  # To change with your filepath\n",
    "\n",
    "# Read in the metadata file.\n",
    "with open(METADATA_FILEPATH, 'r') as f:\n",
    "    claims = json.load(f)\n",
    "\n",
    "# Inspect the first claim.\n",
    "claim = claims[0]\n",
    "print('Claim:', claim['claim'])\n",
    "print('Speaker:', claim['claimant'])\n",
    "print('Date:', claim['date'])\n",
    "print('Related Article Ids:', claim['related_articles'])\n",
    "\n",
    "# Print the first evidence article.\n",
    "idx = claim['related_articles'][0]\n",
    "print('First evidence article id:', idx)\n",
    "with open(os.path.join(ARTICLES_FILEPATH, '%d.txt' % idx), 'r', encoding=\"utf8\") as f:\n",
    "    print(f.read())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Saving the claims.json as pandas dataframe"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "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>claim</th>\n",
       "      <th>claimant</th>\n",
       "      <th>date</th>\n",
       "      <th>label</th>\n",
       "      <th>related_articles</th>\n",
       "      <th>id</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>A line from George Orwell's novel 1984 predict...</td>\n",
       "      <td></td>\n",
       "      <td>2017-07-17</td>\n",
       "      <td>0</td>\n",
       "      <td>[122094, 122580, 130685, 134765]</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Maine legislature candidate Leslie Gibson insu...</td>\n",
       "      <td></td>\n",
       "      <td>2018-03-17</td>\n",
       "      <td>2</td>\n",
       "      <td>[106868, 127320, 128060]</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>A 17-year-old girl named Alyssa Carson is bein...</td>\n",
       "      <td></td>\n",
       "      <td>2018-07-18</td>\n",
       "      <td>1</td>\n",
       "      <td>[132130, 132132, 149722]</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>In 1988 author Roald Dahl penned an open lette...</td>\n",
       "      <td></td>\n",
       "      <td>2019-02-04</td>\n",
       "      <td>2</td>\n",
       "      <td>[123254, 123418, 127464]</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>When it comes to fighting terrorism, \"Another ...</td>\n",
       "      <td>Hillary Clinton</td>\n",
       "      <td>2016-03-22</td>\n",
       "      <td>2</td>\n",
       "      <td>[41099, 89899, 72543, 82644, 95344, 88361]</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                               claim         claimant  \\\n",
       "0  A line from George Orwell's novel 1984 predict...                    \n",
       "1  Maine legislature candidate Leslie Gibson insu...                    \n",
       "2  A 17-year-old girl named Alyssa Carson is bein...                    \n",
       "3  In 1988 author Roald Dahl penned an open lette...                    \n",
       "4  When it comes to fighting terrorism, \"Another ...  Hillary Clinton   \n",
       "\n",
       "         date  label                            related_articles  id  \n",
       "0  2017-07-17      0            [122094, 122580, 130685, 134765]   0  \n",
       "1  2018-03-17      2                    [106868, 127320, 128060]   1  \n",
       "2  2018-07-18      1                    [132130, 132132, 149722]   4  \n",
       "3  2019-02-04      2                    [123254, 123418, 127464]   5  \n",
       "4  2016-03-22      2  [41099, 89899, 72543, 82644, 95344, 88361]   6  "
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Claims as a dataframe\n",
    "claims_pd=pd.DataFrame(claims)\n",
    "#Create a copy of claims dataframe\n",
    "claims = claims_pd.copy(deep = True)\n",
    "# Let's take a look at the dataset\n",
    "claims_pd.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**The set of numbers in 'related_articles' column denote the respective related articles for a particular claim**"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Data Exploration"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    7408\n",
       "1    6451\n",
       "2    1696\n",
       "Name: label, dtype: int64"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Counting the frequency of unique values in the label column\n",
    "claims_pd['label'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x504 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Plotting a Pie chart for the no. of False, Partly True and True labels\n",
    "labels = 'False', 'Partly True', 'True'\n",
    "sizes = [x for x in claims_pd['label'].value_counts()]\n",
    "\n",
    "plt.figure(figsize=(10,7))\n",
    "plt.pie(sizes, labels=labels, autopct='%1.1f%%', textprops={'fontsize': 15})\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**The pie chart represents how varied the portion of false, partly true and true claims in the dataset is. We can observe that there is only 11% of true claims while there is about 48% of false claims**\n",
    "\n",
    "**Let's explore further**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Donald Trump       1233\n",
       "Bloggers            372\n",
       "Barack Obama        234\n",
       "Hillary Clinton     220\n",
       "Viral image         127\n",
       "Name: claimant, dtype: int64"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Counting the frequency of unique values in the claimant column\n",
    "claims_pd['claimant'].value_counts()[1:6] #1st entry is omitted because it is null "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Plotting a bar graph for count of the claims for the first 5 claimants\n",
    "plt.figure(figsize=(10,6))\n",
    "label = ['Donald Trump', 'Bloggers', 'Barack Obama', 'Hillary Clinton', 'Viral image']\n",
    "\n",
    "data = [claims_pd['claimant'].value_counts()[1:6][i] for i in range(5)] \n",
    "plt.bar(label, height = data, width = 0.5)\n",
    "#Setting the title\n",
    "plt.title('Counts of claims for first 5 claimants', fontsize=14)\n",
    "#Setting the x-label\n",
    "plt.xlabel('Top 5 Claimants',fontsize=14)\n",
    "#Seting the y-label\n",
    "plt.ylabel('Claims count',fontsize=14)\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**The plot above represents the top 5 counts of claims made by the claimants. Out of them Donald Trump had the highest number of claims made**\n",
    "\n",
    "**We will divide this further**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "<ipython-input-8-aae21c9f0b20>:3: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  top5['label_new'] = top5['label'].map({0: 'False', 1:'Partly True', 2: 'True'})\n"
     ]
    }
   ],
   "source": [
    "#Putting the top five claimants in a list\n",
    "top5 = claims_pd[(claims_pd['claimant'] == 'Donald Trump') | (claims_pd['claimant'] == 'Bloggers') | (claims_pd['claimant'] == 'Barack Obama') | (claims_pd['claimant'] =='Hillary Clinton') | (claims_pd['claimant'] =='Viral image')]\n",
    "top5['label_new'] = top5['label'].map({0: 'False', 1:'Partly True', 2: 'True'})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Plotting a histogram of for the top 5 claimants and the validity of their claim\n",
    "plt.figure(figsize=(12,6))\n",
    "sns.countplot(x='claimant',data=top5,hue='label_new')\n",
    "#Setting the title\n",
    "plt.title('Claim classification of Top-5 Claimants', fontsize=18)\n",
    "#Setting the x-label\n",
    "plt.xlabel('Top-5 Claimants',fontsize=15)\n",
    "plt.xticks(fontsize=13)\n",
    "#Setting the y-label\n",
    "plt.ylabel('Claims Count',fontsize=15)\n",
    "plt.yticks(fontsize=13)\n",
    "\n",
    "plt.legend(fontsize=13)\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**The above plot represents the distribution of number of false, partly true and true claims made by the top 5 claimants**\n",
    "\n",
    "**Out of the top 5 only Barack Obama and Hilary Clinton made true claims more than the false or partly true claims**\n",
    "\n",
    "**Also it is astonishing to note that Donald Trump made more false claims than 50% of the his total claims**"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Data Cleaning"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### a) Cleaning Articles"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Extracting the articles from the system and merging the articles for the respective claims below. Then, performing text cleaning on the merged article for each claim, removing stopwords, converting the letters to small case, and stemming the words**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Defining a new function to concatenate and clean the articles\n",
    "\n",
    "def new_data(ind):\n",
    "    concat_article = []\n",
    "    for i in range(len(claims_pd.iloc[ind, -2])):\n",
    "        article_number = claims_pd.iloc[ind, -2][i]\n",
    "        with open(os.path.join(ARTICLES_FILEPATH, '%d.txt' % article_number), 'r', encoding=\"utf8\") as f:\n",
    "            a = f.read()\n",
    "        concat_article.append(a)\n",
    "    new = []\n",
    "    for a in range(len(concat_article)):\n",
    "        sp = concat_article[a].split('\\n')\n",
    "        new.append(sp)\n",
    "    \n",
    "    sentences = sent_tokenize(str(new))\n",
    "    tokens = re.split(r'\\W+', str(sentences))\n",
    "    tokens = [w.lower() for w in tokens]\n",
    "    table = str.maketrans('', '', string.punctuation)\n",
    "    stripped = [w.translate(table) for w in tokens]\n",
    "    words = [word for word in stripped if word.isalpha()]\n",
    "    \n",
    "    stop_words = set(stopwords.words('english'))\n",
    "    words = [w for w in words if not w in stop_words]\n",
    "    \n",
    "    stemmatizer = SnowballStemmer('english')\n",
    "    stemmatized = [stemmatizer.stem(word) for word in words]\n",
    "    return(' '.join(stemmatized))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Execution time for concatinating and cleaning the articles =  901.9781768321991\n"
     ]
    }
   ],
   "source": [
    "#Applying the above function to our dataframe\n",
    "start = time.time()\n",
    "clean_articles = [new_data(i) for i in range(len(claims_pd))]\n",
    "print('Execution time for concatinating and cleaning the articles = ', time.time()-start)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### b) Cleaning Claims\n",
    "\n",
    "The process for text cleaning is the same as that of articles."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Defining a function to clean the claims\n",
    "\n",
    "def clean_claim(ind):\n",
    "    \n",
    "    sentences = sent_tokenize(str(claims_pd.iloc[ind, 0]))\n",
    "    tokens = re.split(r'\\W+',str(sentences))\n",
    "    tokens = [w.lower() for w in tokens]\n",
    "    table = str.maketrans('', '', string.punctuation)\n",
    "    stripped = [w.translate(table) for w in tokens]\n",
    "    words = [word for word in stripped if word.isalpha()]\n",
    "    \n",
    "    stop_words = set(stopwords.words('english'))\n",
    "    words = [w for w in words if not w in stop_words]\n",
    "    stemmatizer = SnowballStemmer('english')\n",
    "    stemmatized = [stemmatizer.stem(word) for word in words]\n",
    "    return(' '.join(stemmatized))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Execution time for cleaning the claims =  9.275229215621948\n"
     ]
    }
   ],
   "source": [
    "#Applying the above function to our dataframe\n",
    "start = time.time()\n",
    "clean_claims = [clean_claim(i) for i in range(len(claims_pd))]\n",
    "print('Execution time for cleaning the claims = ', time.time()-start)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### c) Appending cleaned articles and claims in the claims_pd dataframe."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "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>claim</th>\n",
       "      <th>claimant</th>\n",
       "      <th>date</th>\n",
       "      <th>label</th>\n",
       "      <th>related_articles</th>\n",
       "      <th>id</th>\n",
       "      <th>clean articles</th>\n",
       "      <th>clean claims</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>A line from George Orwell's novel 1984 predict...</td>\n",
       "      <td></td>\n",
       "      <td>2017-07-17</td>\n",
       "      <td>0</td>\n",
       "      <td>[122094, 122580, 130685, 134765]</td>\n",
       "      <td>0</td>\n",
       "      <td>georg orwel predict almost year ago april comr...</td>\n",
       "      <td>line georg orwel novel predict power smartphon</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Maine legislature candidate Leslie Gibson insu...</td>\n",
       "      <td></td>\n",
       "      <td>2018-03-17</td>\n",
       "      <td>2</td>\n",
       "      <td>[106868, 127320, 128060]</td>\n",
       "      <td>1</td>\n",
       "      <td>republican critic parkland student drop main h...</td>\n",
       "      <td>main legislatur candid lesli gibson insult par...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>A 17-year-old girl named Alyssa Carson is bein...</td>\n",
       "      <td></td>\n",
       "      <td>2018-07-18</td>\n",
       "      <td>1</td>\n",
       "      <td>[132130, 132132, 149722]</td>\n",
       "      <td>4</td>\n",
       "      <td>first person complet nasa passport program ast...</td>\n",
       "      <td>year old girl name alyssa carson train nasa be...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>In 1988 author Roald Dahl penned an open lette...</td>\n",
       "      <td></td>\n",
       "      <td>2019-02-04</td>\n",
       "      <td>2</td>\n",
       "      <td>[123254, 123418, 127464]</td>\n",
       "      <td>5</td>\n",
       "      <td>danger measl tweet march mani parent think mea...</td>\n",
       "      <td>author roald dahl pen open letter urg parent c...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>When it comes to fighting terrorism, \"Another ...</td>\n",
       "      <td>Hillary Clinton</td>\n",
       "      <td>2016-03-22</td>\n",
       "      <td>2</td>\n",
       "      <td>[41099, 89899, 72543, 82644, 95344, 88361]</td>\n",
       "      <td>6</td>\n",
       "      <td>remark counterterror stanford univers slogan s...</td>\n",
       "      <td>come fight terror anoth thing know work base l...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                               claim         claimant  \\\n",
       "0  A line from George Orwell's novel 1984 predict...                    \n",
       "1  Maine legislature candidate Leslie Gibson insu...                    \n",
       "2  A 17-year-old girl named Alyssa Carson is bein...                    \n",
       "3  In 1988 author Roald Dahl penned an open lette...                    \n",
       "4  When it comes to fighting terrorism, \"Another ...  Hillary Clinton   \n",
       "\n",
       "         date  label                            related_articles  id  \\\n",
       "0  2017-07-17      0            [122094, 122580, 130685, 134765]   0   \n",
       "1  2018-03-17      2                    [106868, 127320, 128060]   1   \n",
       "2  2018-07-18      1                    [132130, 132132, 149722]   4   \n",
       "3  2019-02-04      2                    [123254, 123418, 127464]   5   \n",
       "4  2016-03-22      2  [41099, 89899, 72543, 82644, 95344, 88361]   6   \n",
       "\n",
       "                                      clean articles  \\\n",
       "0  georg orwel predict almost year ago april comr...   \n",
       "1  republican critic parkland student drop main h...   \n",
       "2  first person complet nasa passport program ast...   \n",
       "3  danger measl tweet march mani parent think mea...   \n",
       "4  remark counterterror stanford univers slogan s...   \n",
       "\n",
       "                                        clean claims  \n",
       "0     line georg orwel novel predict power smartphon  \n",
       "1  main legislatur candid lesli gibson insult par...  \n",
       "2  year old girl name alyssa carson train nasa be...  \n",
       "3  author roald dahl pen open letter urg parent c...  \n",
       "4  come fight terror anoth thing know work base l...  "
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Adding our cleaned claims and cleaned articles to the orignal dataframe\n",
    "claims_pd['clean articles'] = clean_articles\n",
    "claims_pd['clean claims'] = clean_claims\n",
    "\n",
    "claims_pd.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**A column is added with the Year of the claim below**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "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>claim</th>\n",
       "      <th>claimant</th>\n",
       "      <th>date</th>\n",
       "      <th>label</th>\n",
       "      <th>related_articles</th>\n",
       "      <th>id</th>\n",
       "      <th>clean articles</th>\n",
       "      <th>clean claims</th>\n",
       "      <th>year</th>\n",
       "      <th>label_new</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>A line from George Orwell's novel 1984 predict...</td>\n",
       "      <td></td>\n",
       "      <td>2017-07-17</td>\n",
       "      <td>0</td>\n",
       "      <td>[122094, 122580, 130685, 134765]</td>\n",
       "      <td>0</td>\n",
       "      <td>georg orwel predict almost year ago april comr...</td>\n",
       "      <td>line georg orwel novel predict power smartphon</td>\n",
       "      <td>2017</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Maine legislature candidate Leslie Gibson insu...</td>\n",
       "      <td></td>\n",
       "      <td>2018-03-17</td>\n",
       "      <td>2</td>\n",
       "      <td>[106868, 127320, 128060]</td>\n",
       "      <td>1</td>\n",
       "      <td>republican critic parkland student drop main h...</td>\n",
       "      <td>main legislatur candid lesli gibson insult par...</td>\n",
       "      <td>2018</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>A 17-year-old girl named Alyssa Carson is bein...</td>\n",
       "      <td></td>\n",
       "      <td>2018-07-18</td>\n",
       "      <td>1</td>\n",
       "      <td>[132130, 132132, 149722]</td>\n",
       "      <td>4</td>\n",
       "      <td>first person complet nasa passport program ast...</td>\n",
       "      <td>year old girl name alyssa carson train nasa be...</td>\n",
       "      <td>2018</td>\n",
       "      <td>Partly True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>In 1988 author Roald Dahl penned an open lette...</td>\n",
       "      <td></td>\n",
       "      <td>2019-02-04</td>\n",
       "      <td>2</td>\n",
       "      <td>[123254, 123418, 127464]</td>\n",
       "      <td>5</td>\n",
       "      <td>danger measl tweet march mani parent think mea...</td>\n",
       "      <td>author roald dahl pen open letter urg parent c...</td>\n",
       "      <td>2019</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>When it comes to fighting terrorism, \"Another ...</td>\n",
       "      <td>Hillary Clinton</td>\n",
       "      <td>2016-03-22</td>\n",
       "      <td>2</td>\n",
       "      <td>[41099, 89899, 72543, 82644, 95344, 88361]</td>\n",
       "      <td>6</td>\n",
       "      <td>remark counterterror stanford univers slogan s...</td>\n",
       "      <td>come fight terror anoth thing know work base l...</td>\n",
       "      <td>2016</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                               claim         claimant  \\\n",
       "0  A line from George Orwell's novel 1984 predict...                    \n",
       "1  Maine legislature candidate Leslie Gibson insu...                    \n",
       "2  A 17-year-old girl named Alyssa Carson is bein...                    \n",
       "3  In 1988 author Roald Dahl penned an open lette...                    \n",
       "4  When it comes to fighting terrorism, \"Another ...  Hillary Clinton   \n",
       "\n",
       "         date  label                            related_articles  id  \\\n",
       "0  2017-07-17      0            [122094, 122580, 130685, 134765]   0   \n",
       "1  2018-03-17      2                    [106868, 127320, 128060]   1   \n",
       "2  2018-07-18      1                    [132130, 132132, 149722]   4   \n",
       "3  2019-02-04      2                    [123254, 123418, 127464]   5   \n",
       "4  2016-03-22      2  [41099, 89899, 72543, 82644, 95344, 88361]   6   \n",
       "\n",
       "                                      clean articles  \\\n",
       "0  georg orwel predict almost year ago april comr...   \n",
       "1  republican critic parkland student drop main h...   \n",
       "2  first person complet nasa passport program ast...   \n",
       "3  danger measl tweet march mani parent think mea...   \n",
       "4  remark counterterror stanford univers slogan s...   \n",
       "\n",
       "                                        clean claims  year    label_new  \n",
       "0     line georg orwel novel predict power smartphon  2017        False  \n",
       "1  main legislatur candid lesli gibson insult par...  2018         True  \n",
       "2  year old girl name alyssa carson train nasa be...  2018  Partly True  \n",
       "3  author roald dahl pen open letter urg parent c...  2019         True  \n",
       "4  come fight terror anoth thing know work base l...  2016         True  "
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Adding a new column showing the year for the made claim\n",
    "claims_pd['year'] = [int(claims_pd['date'][i][0:4]) for i in claims_pd.index]\n",
    "#Adding a new column showing the validity(label) for the made claim\n",
    "claims_pd['label_new'] = claims_pd['label'].map({0: 'False', 1:'Partly True', 2: 'True'})\n",
    "\n",
    "claims_pd.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Plotting a histogram for the no of claims over the years and their label\n",
    "plt.figure(figsize=(15,6))\n",
    "sns.countplot(x='year',data=claims_pd,hue='label_new')\n",
    "#Setting the title\n",
    "plt.title('Claims Classification over the Years', fontsize=16)\n",
    "#Setting the x-label\n",
    "plt.xlabel('Year',fontsize=15)\n",
    "plt.xticks(fontsize=13, rotation = 45)\n",
    "#Setting the y-label\n",
    "plt.ylabel('Claims Count',fontsize=15)\n",
    "plt.yticks(fontsize=13)\n",
    "\n",
    "plt.legend(fontsize=15)\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The plot depicts the bifurcation of claims made each year. As we can see the number of claims made kept on increasing until 2018 and along with it the count of false claims also kept on increasing each year.\n",
    "\n",
    "**Note: As there isn't much data for 2019 in our dataset, a low count is observed which should not be confused with low count of claims made actually in that year**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x504 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Plotting a bar chart for the no. of false claimants over the years\n",
    "fc = claims_pd[claims_pd['label']==0].groupby(['year'])['claimant'].nunique()[9:13]\n",
    "label = ['2015', '2016', '2017', '2018']\n",
    "plt.figure(figsize=(10,7))\n",
    "plt.bar(label, height = fc, width = 0.5)\n",
    "\n",
    "#Setting the title\n",
    "plt.title('No of False Claimants over the years', fontsize=17)\n",
    "#Setting the x-label\n",
    "plt.xlabel('Years',fontsize=15)\n",
    "plt.xticks(fontsize=13)\n",
    "#Seting the y-label\n",
    "plt.ylabel('Claimants Count',fontsize=15)\n",
    "plt.yticks(fontsize=13)\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The bar-chart shows how the number of false claimants have increased year-on-year since 2015 to 2018"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**For the model, we will omit the claims made by Donald Trump, Hillary Clinton and Barack Obama and then make predictions on these omitted datapoints**\n",
    "\n",
    "**Splitting the dataset below**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "big = claims_pd[~ ((claims_pd['claimant']=='Donald Trump') | (claims_pd['claimant']=='Hillary Clinton') | (claims_pd['claimant']=='Barack Obama'))] \n",
    "# Training data set without taking Donald Trump, Hillary Clinton and Barack Obama in the data set\n",
    "\n",
    "small = claims_pd[(claims_pd['claimant']=='Donald Trump') | (claims_pd['claimant']=='Hillary Clinton') | (claims_pd['claimant']=='Barack Obama')] \n",
    "# Testing data set with Donald Trump, Hillary Clinton and Barack Obama in the data set"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "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>claim</th>\n",
       "      <th>claimant</th>\n",
       "      <th>date</th>\n",
       "      <th>label</th>\n",
       "      <th>related_articles</th>\n",
       "      <th>id</th>\n",
       "      <th>clean articles</th>\n",
       "      <th>clean claims</th>\n",
       "      <th>year</th>\n",
       "      <th>label_new</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>A line from George Orwell's novel 1984 predict...</td>\n",
       "      <td></td>\n",
       "      <td>2017-07-17</td>\n",
       "      <td>0</td>\n",
       "      <td>[122094, 122580, 130685, 134765]</td>\n",
       "      <td>0</td>\n",
       "      <td>georg orwel predict almost year ago april comr...</td>\n",
       "      <td>line georg orwel novel predict power smartphon</td>\n",
       "      <td>2017</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Maine legislature candidate Leslie Gibson insu...</td>\n",
       "      <td></td>\n",
       "      <td>2018-03-17</td>\n",
       "      <td>2</td>\n",
       "      <td>[106868, 127320, 128060]</td>\n",
       "      <td>1</td>\n",
       "      <td>republican critic parkland student drop main h...</td>\n",
       "      <td>main legislatur candid lesli gibson insult par...</td>\n",
       "      <td>2018</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>A 17-year-old girl named Alyssa Carson is bein...</td>\n",
       "      <td></td>\n",
       "      <td>2018-07-18</td>\n",
       "      <td>1</td>\n",
       "      <td>[132130, 132132, 149722]</td>\n",
       "      <td>4</td>\n",
       "      <td>first person complet nasa passport program ast...</td>\n",
       "      <td>year old girl name alyssa carson train nasa be...</td>\n",
       "      <td>2018</td>\n",
       "      <td>Partly True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>In 1988 author Roald Dahl penned an open lette...</td>\n",
       "      <td></td>\n",
       "      <td>2019-02-04</td>\n",
       "      <td>2</td>\n",
       "      <td>[123254, 123418, 127464]</td>\n",
       "      <td>5</td>\n",
       "      <td>danger measl tweet march mani parent think mea...</td>\n",
       "      <td>author roald dahl pen open letter urg parent c...</td>\n",
       "      <td>2019</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>Rhode Island is \"almost dead last\" among North...</td>\n",
       "      <td>Leonidas Raptakis</td>\n",
       "      <td>2014-02-11</td>\n",
       "      <td>2</td>\n",
       "      <td>[8284, 3768, 20091, 82368, 73148, 4493]</td>\n",
       "      <td>7</td>\n",
       "      <td>lis code virginia prev next punish convict fel...</td>\n",
       "      <td>rhode island almost dead last among northeaste...</td>\n",
       "      <td>2014</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                               claim           claimant  \\\n",
       "0  A line from George Orwell's novel 1984 predict...                      \n",
       "1  Maine legislature candidate Leslie Gibson insu...                      \n",
       "2  A 17-year-old girl named Alyssa Carson is bein...                      \n",
       "3  In 1988 author Roald Dahl penned an open lette...                      \n",
       "5  Rhode Island is \"almost dead last\" among North...  Leonidas Raptakis   \n",
       "\n",
       "         date  label                         related_articles  id  \\\n",
       "0  2017-07-17      0         [122094, 122580, 130685, 134765]   0   \n",
       "1  2018-03-17      2                 [106868, 127320, 128060]   1   \n",
       "2  2018-07-18      1                 [132130, 132132, 149722]   4   \n",
       "3  2019-02-04      2                 [123254, 123418, 127464]   5   \n",
       "5  2014-02-11      2  [8284, 3768, 20091, 82368, 73148, 4493]   7   \n",
       "\n",
       "                                      clean articles  \\\n",
       "0  georg orwel predict almost year ago april comr...   \n",
       "1  republican critic parkland student drop main h...   \n",
       "2  first person complet nasa passport program ast...   \n",
       "3  danger measl tweet march mani parent think mea...   \n",
       "5  lis code virginia prev next punish convict fel...   \n",
       "\n",
       "                                        clean claims  year    label_new  \n",
       "0     line georg orwel novel predict power smartphon  2017        False  \n",
       "1  main legislatur candid lesli gibson insult par...  2018         True  \n",
       "2  year old girl name alyssa carson train nasa be...  2018  Partly True  \n",
       "3  author roald dahl pen open letter urg parent c...  2019         True  \n",
       "5  rhode island almost dead last among northeaste...  2014         True  "
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "big.head()  #Displaying the training dataset"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Model Preparation\n",
    "\n",
    "**As the datasets are ready now, dataset is vecorized using TF-IDF and the model implementation is done**"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Using TF-IDF vectorizer for converting the text to vectors.\n",
    "\n",
    "Term frequency–inverse document frequency, is a numerical statistic that is intended to reflect how important a word is to a document in a collection or corpus. The tf–idf value increases proportionally to the number of times a word appears in the document and is offset by the number of documents in the corpus that contain the word, which helps to adjust for the fact that some words appear more frequently in general.\n",
    "\n",
    "Source - https://en.wikipedia.org/wiki/Tf%E2%80%93idf"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### a) Vectorizing training dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "#Using the Tfidf Vectorizer on the training set\n",
    "\n",
    "vectorizer1 = TfidfVectorizer(use_idf=True, smooth_idf=True, norm='l2', max_features = 1400)\n",
    "vectorizer1.fit(big['clean claims']+big['clean articles'])\n",
    "array_TFIDF1 = vectorizer1.transform(big['clean claims']+big['clean articles'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### b) Vectorizing testing dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "#Vectorizing the test set\n",
    "\n",
    "array_TFIDF2 = vectorizer1.transform(small['clean claims']+small['clean articles'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Allocating the training and testing datasets\n",
    "\n",
    "X_train = array_TFIDF1\n",
    "X_test = array_TFIDF2\n",
    "y_train = big['label']\n",
    "y_test = small['label']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Model Implementation"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Let's implement algorithms on the training dataset**"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**a) Random Forest**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.5732068761114404\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.66      0.59      0.62       910\n",
      "           1       0.49      0.65      0.56       669\n",
      "           2       0.17      0.01      0.02       108\n",
      "\n",
      "    accuracy                           0.57      1687\n",
      "   macro avg       0.44      0.41      0.40      1687\n",
      "weighted avg       0.56      0.57      0.56      1687\n",
      "\n"
     ]
    }
   ],
   "source": [
    "#Using the Random Forest Algorithm\n",
    "rf = RandomForestClassifier(random_state = 0)\n",
    "#Fitting the model\n",
    "rf.fit(X_train, y_train)\n",
    "predictions0 = rf.predict(X_test)\n",
    "#Finding the accuracy of the predictions\n",
    "ACCU = accuracy_score(y_test, predictions0)\n",
    "print('Accuracy:', ACCU)\n",
    "#Printing the classification report \n",
    "print(classification_report(y_test,predictions0))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**b) XGBoost**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\siddh\\anaconda3\\lib\\site-packages\\xgboost\\sklearn.py:1224: UserWarning: The use of label encoder in XGBClassifier is deprecated and will be removed in a future release. To remove this warning, do the following: 1) Pass option use_label_encoder=False when constructing XGBClassifier object; and 2) Encode your labels (y) as integers starting with 0, i.e. 0, 1, 2, ..., [num_class - 1].\n",
      "  warnings.warn(label_encoder_deprecation_msg, UserWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:39:57] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.5.1/src/learner.cc:1115: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'multi:softprob' was changed from 'merror' to 'mlogloss'. Explicitly set eval_metric if you'd like to restore the old behavior.\n",
      "Accuracy: 0.6\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.69      0.62      0.65       910\n",
      "           1       0.51      0.66      0.58       669\n",
      "           2       0.27      0.03      0.05       108\n",
      "\n",
      "    accuracy                           0.60      1687\n",
      "   macro avg       0.49      0.43      0.43      1687\n",
      "weighted avg       0.59      0.60      0.58      1687\n",
      "\n"
     ]
    }
   ],
   "source": [
    "xgb = XGBClassifier(random_state = 0)\n",
    "xgb.fit(X_train, y_train)\n",
    "predictions = xgb.predict(X_test)\n",
    "ACCU = accuracy_score(y_test, predictions)\n",
    "print('Accuracy:', round(ACCU,2))\n",
    "print(classification_report(y_test,predictions))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**c) Support Vector Machines (SVM)**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.5275637225844695\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.63      0.49      0.55       910\n",
      "           1       0.45      0.66      0.54       669\n",
      "           2       0.42      0.05      0.08       108\n",
      "\n",
      "    accuracy                           0.53      1687\n",
      "   macro avg       0.50      0.40      0.39      1687\n",
      "weighted avg       0.55      0.53      0.52      1687\n",
      "\n"
     ]
    }
   ],
   "source": [
    "#Using SVM ( Linear ) model\n",
    "svcl = LinearSVC(random_state = 0)\n",
    "#Fitting the model\n",
    "svcl.fit(X_train, y_train)\n",
    "predictions2 = svcl.predict(X_test)\n",
    "#Finding the accuracy of the predictions\n",
    "ACCU = accuracy_score(y_test, predictions2)\n",
    "print('Accuracy:', ACCU)\n",
    "#Printing the classification report \n",
    "print(classification_report(y_test,predictions2))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**d) Logistic Regression**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.5323058684054535\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.64      0.50      0.56       910\n",
      "           1       0.46      0.65      0.54       669\n",
      "           2       0.50      0.04      0.07       108\n",
      "\n",
      "    accuracy                           0.53      1687\n",
      "   macro avg       0.53      0.40      0.39      1687\n",
      "weighted avg       0.56      0.53      0.52      1687\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\siddh\\anaconda3\\lib\\site-packages\\sklearn\\linear_model\\_logistic.py:814: ConvergenceWarning: lbfgs failed to converge (status=1):\n",
      "STOP: TOTAL NO. of ITERATIONS REACHED LIMIT.\n",
      "\n",
      "Increase the number of iterations (max_iter) or scale the data as shown in:\n",
      "    https://scikit-learn.org/stable/modules/preprocessing.html\n",
      "Please also refer to the documentation for alternative solver options:\n",
      "    https://scikit-learn.org/stable/modules/linear_model.html#logistic-regression\n",
      "  n_iter_i = _check_optimize_result(\n"
     ]
    }
   ],
   "source": [
    "#Using Logistic Regression\n",
    "model_LR = LogisticRegression()\n",
    "model_LR.fit(X_train,y_train)\n",
    "predictions3 = model_LR.predict(X_test)\n",
    "#Finding the accuracy of the predictions\n",
    "ACCUs = accuracy_score(y_test, predictions3)\n",
    "print('Accuracy:', ACCUs)\n",
    "#Printing the classification report \n",
    "print(classification_report(y_test,predictions3))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Out of the models implemented, XGBoost performs best as it gave the best combination of accuracy and macro F1 score. Therefore, selecting this model for further analysis**"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Discussions on model\n",
    "\n",
    "**Plotting to make inferences from our model and discuss the viability**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Correct      1006\n",
       "Incorrect     681\n",
       "Name: Prediction Class, dtype: int64"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Converting our y_test list to a dataframe\n",
    "x = pd.DataFrame(y_test)\n",
    "\n",
    "#Defining new columns\n",
    "x['predictions'] = predictions\n",
    "x['Year'] = small['year']\n",
    "x['claimant'] = small['claimant']\n",
    "x['label_new'] = x['label'].map({0: 'False', 1:'Partly True', 2: 'True'})\n",
    "x['Prediction Class']=['Correct' if x['label'][i]==x['predictions'][i] else 'Incorrect' for i in x.index]\n",
    "\n",
    "#Counting the no. of Correct and Incorrect predictions\n",
    "x['Prediction Class'].value_counts()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**a) Plotting to see how good our model worked**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x576 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Plotting a bar graph for the validity(label) of the actual claims\n",
    "fig, p=plt.subplots(nrows=1, ncols=2, figsize=(15,8), sharey = True)\n",
    "sns.countplot(x='label_new',data=x,ax=p[0])\n",
    "#Setting the title\n",
    "p[0].set_title('Actual Claims', fontsize = 17)\n",
    "#Setting the x-label\n",
    "p[0].set_xlabel('Claims Classification', fontsize = 17)\n",
    "#Setting the y-label\n",
    "p[0].set_ylabel('Claims Count', fontsize = 17)\n",
    "p[0].tick_params(axis='both', which='major', labelsize=15)\n",
    "\n",
    "#Plotting a bargraph for the predicted claims and their accuracy\n",
    "sns.countplot(x='label_new',data=x,hue='Prediction Class',ax=p[1])\n",
    "#Setting the title\n",
    "p[1].set_title('Claims Predicted by the model', fontsize = 17)\n",
    "#Setting the x-label\n",
    "p[1].set_xlabel('Claims Classification', fontsize = 17)\n",
    "#Setting the y-label\n",
    "p[1].set_ylabel('Claims Count', fontsize = 17)\n",
    "p[1].tick_params(axis='both', which='major', labelsize=15)\n",
    "p[1].legend(fontsize=15)\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**From the plot it can be inferred that the model correctly predicted the false claims better than the others as the data for false claims is more as compared to that of the true or partly true claims**\n",
    "\n",
    "**And with more/equivalent number of datapoints for the other two categories, our model can predict classes more accurately**"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**b) Plotting for count of claims distributed over the years**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 936x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(13,6))\n",
    "#Plotting a graph for no of claims made by Donald Trump over the years and their validity(label)\n",
    "sns.countplot(x='Year',data=x,hue='label_new')\n",
    "#Setting the title\n",
    "plt.title('Year-wise Claims Classification', fontsize=18)\n",
    "plt.legend(loc = 2, fontsize=12)\n",
    "#Setting the x-label\n",
    "plt.xlabel('Year',fontsize=15)\n",
    "plt.xticks(fontsize=13)\n",
    "#Setting the y-label\n",
    "plt.ylabel('Claims Count',fontsize=15)\n",
    "plt.yticks(fontsize=13)\n",
    "#Displaying the graph\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**The plot shows the collective number of false, partly true, true claims made by Donald Trump, Hillary Clinton and Barack Obama each year. It can be inferred that collectively they made mostly false claims and the peak was around the time of USA 2016 elections**\n",
    "\n",
    "**Let's explore further on the Correct Predictions of our testing dataset**"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**a) Plotting distribution of claims for the three politicians**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x504 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(12,7))\n",
    "ax = sns.countplot(x=\"claimant\", data=x[x['Prediction Class']=='Correct'], hue='label_new')\n",
    "plt.xlabel('Claimant',fontsize=15)\n",
    "plt.xticks(fontsize=13)\n",
    "plt.ylabel('Claims Count (Predicted Correctly)',fontsize=15)\n",
    "plt.yticks(fontsize=13)\n",
    "plt.title(\"Distribution of Claims for the Three US Politicians\", fontsize=16)\n",
    "plt.legend(fontsize=15)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**The model has correctly predicted the above count of claims made by the politicians**\n",
    "\n",
    "**It can be observed that the highest count is of false claims made by Donald Trump**\n",
    "\n",
    "**b) Plotting distribution of false claims made by Donald Trump over the years**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x504 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(12,7))\n",
    "ax = sns.countplot(x=\"Year\", data=x[(x['Prediction Class']=='Correct') & (x['claimant']=='Donald Trump') & (x['label_new']=='False')])\n",
    "plt.xlabel('Years',fontsize=16)\n",
    "plt.xticks(fontsize=14)\n",
    "plt.ylabel('False Claims Count of Donald Trump',fontsize=16)\n",
    "plt.yticks(fontsize=14)\n",
    "plt.title(\"Correctly Predicted False Claims made by Donald Trump over the years\", fontsize=18)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**It can be inferred that the count of false claims by Donald Trump increased exponentially until 2016 after which he came to power. Even after coming to power, he carried out the same philosophy and kept on making false claims**\n",
    "\n",
    "**This shows that since his ideology of making false claims worked for the better for him during the elections, he didn't stop it and relied on the same to influence the American society**"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Here are some of the examples of false claims made by Donald Trump after he became President. These claims were correctly predicted by the model as false**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The claim made -   “In fact the GDP since I’ve taken over has doubled and tripled.”  \n",
      "\n",
      "Details of the claim -\n",
      " label                          0\n",
      "predictions                    1\n",
      "Year                        2018\n",
      "claimant            Donald Trump\n",
      "label_new                  False\n",
      "Prediction Class       Incorrect\n",
      "Name: 3445, dtype: object\n"
     ]
    }
   ],
   "source": [
    "print (\"The claim made - \",claims_pd.loc[3445].claim, \"\\n\\nDetails of the claim -\\n\",x.loc[3445])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### This is a wild exaggeration. U.S. gross domestic product grew by 2.3 per cent in 2017. It grew by 2 per cent in the first quarter of 2018.\n",
    "\n",
    "Source: https://projects.thestar.com/donald-trump-fact-check/"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The claim made -   \"I don’t know if you noticed, but deficits seem to be coming down.\"  \n",
      "\n",
      "Details of the claim -\n",
      " label                          0\n",
      "predictions                    1\n",
      "Year                        2019\n",
      "claimant            Donald Trump\n",
      "label_new                  False\n",
      "Prediction Class       Incorrect\n",
      "Name: 7825, dtype: object\n"
     ]
    }
   ],
   "source": [
    "print (\"The claim made - \",claims_pd.loc[7825].claim, \"\\n\\nDetails of the claim -\\n\", x.loc[7825])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### In fact, the deficit in the first three months of fiscal year 2019 was 42 percent higher than it was for the same period last year.\n",
    "\n",
    "Source: https://www.factcheck.org/2019/02/trump-disinformation-on-deficits/"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**c) Plotting confusion matrix heatmap of our model**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAdcAAAGHCAYAAADiChg2AAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8vihELAAAACXBIWXMAAAsTAAALEwEAmpwYAAA6QElEQVR4nO3dd5wU9f3H8ddnD47eFaQKAoKgoogtGmNX7CbB2LGB5mc0UaPBGFui0USisSs2ULFgL4kaFE0soCKiUhSQ3qVzcJS7+/z+mDncO64s7N7u7N776WMftzPzne98Zo/zs9/vfOc75u6IiIhI6sQyHYCIiEiuUXIVERFJMSVXERGRFFNyFRERSTElVxERkRRTchUREUkxJVeplczsJjNbYmZuZueloL7OYV39UhBeZJnZoeF57pDpWESiTMlVIsPM2pjZ3Wb2vZltNLMFZvaWmR2X4uPsDtwIXAK0BZ5PQbXzwrompqCuSsUlt9Vm1rDctt3CbduU/MxsuJm9mWDxTwjOc/k2hC1S69TJdAAiELT8gI+BtcC1wFcEX/6OAB4COqXwcN3Cn696imZRcfdiYHEq6krQamAAMCJu3YXAXFL7WW1hZnXdfRPpPU+RrKSWq0TFA4AB/dx9lLt/5+5T3f0+oE9pITPrZGavmNna8PWymXWI236TmU0ys9PDFvBaM3u1tCVnZjcBr4TFS8zMw/Vbtd5K64pb3sPM3jOzNWG9X5nZYeG2rbqFzewQM/vUzDaEXdB3mVl+3PYPzOwBM/urmS0zs6VmNtTMEvm7HA5cEFdXXeCccH38OeSZ2WNmNsvMCs1supldU3qM8PMYCBwf1+o9NO58zjCzMWZWCFxcvls4rHuymTWIO95H29ASFslJSq6ScWbWEjgWuM/dC8pvd/eVYTkDXgXaAIcDhwHtgFfDbaU6A78CTgWOBvYGbg23DQUGhe/bhq9EPQMsAvYL67wJ2FDJObUH3gK+DMteCJwB3Fau6FlAEfAT4DfA78LYq/M0sJ+ZdQ2XTwAKgA/KlYsBC4DTgN2A64A/AueH24cCo4B3+fHz+CRu/9sIvvj0Ivjsy7scqBvWQ1h/N+ISv0htpG5hiYJuBK3WqdWUO5KgFdvV3WcDmNmZwAyC7uN3w3J1gPPcfXVYZhhhMnH3AjNbFb7f1u7NnYGh7v5tuDyjirL/R5CI/8/dS4CpZjYEeNjMrnf39WG5Ke5+Q/h+mpkNCs/l2WpiWQG8TpDEriNI3k8AZbq53X0zcEPcqtlm1pcg0T8Wfh6FwMb4zyPuu8q97v5i3PpucXXh7uvC38EnZracoEv/JHdfWk38IjlNLVeJAqu+CBC0vBaWJlYAd58JLCRoWZWaU5pYQwuB1skGCdwJPBp2k15nZj2riXVsmFhLfQTk8+M1X4Cvy+23LbE+Bgw0s47AUZTrEi5lZpeY2Xgz+8HMCoArSPy67PjqCrj7eIKegeuBYe7+VoJ1i+QsJVeJgukELa7dqilnlGuZxYlfv7mCbdX9Wy9h6yRft0wl7jfxY/foT4Cvzayy7s+ajLXUu0Ax8CQwxt3nbxWE2a+AfxIk3mOAvQi6efPLl63EuuoKhF3yB4exdC3XRS9SKym5Ssa5+wrgHeA3Zta4/HYzax6+nQK0D0cWl27bheC665Qkw/iBra+/7lVBrNPd/R53P56g5XhRJfVNAQ4sNzjpYGAT8H2SsZbGUkKQNA8NY6nIwcCn7n6fu09w9xlA13JlNgF5SYRyJdAXOAQ4ALgsibpEcoKSq0TF/xG09sab2QAz62FmPc3s1/zYdfouwS06I81sn3Bk7khgAjAmyeOPAfY2swvMrJuZXQMcVLrRzBqY2f1xI2n3J0hclSX1BwiS/gMW3H96PHA7waCt9ZXssz1uAXYEXq5k+zSgr5n1N7PuZnY98LNyZWYDu4ef+Q7hyOOEmFkfgi7hwe7+CfBr4G8W3EssUmspuUokuPssgtbPaOBvBAl1DHAScHFYxoFTCFqZHwDvE9xzeUqy96u6+zvAzQSJ4guCEccPxBUpBloQ3Ff6HcHtPGMJWm0V1bcA6E8wUngi8DjBIKU/JhNnBcfZ7O7Lyl3bjfcwwWjgZ4DPCc7rH+XKPEIwmGw8wWd7EAkws/oEX26ecfeXwnieBV4k+AJUb9vORiR3WIruoRcREZGQWq4iIiIppuQqIiKSYkquIiIiKabkKiIikmJKriIiIimWNXMLX2JNNaw5R/39sPJzGkiuaPrmB5kOQWpKw2Y1NhNXsv+/f8jXZHyWsKxJriIiUjvkQpdqLpyDiIhIpKjlKiIikRLLgWc/KLmKiEik5EKXqpKriIhESiz7G6458QVBREQkUtRyFRGRSMmFVp+Sq4iIRIoGNImIiKSYWq4iIiIppgFNIiIishW1XEVEJFJyodWn5CoiIpFiGtAkIiKSWmq5ioiIpJgGNImIiMhW1HIVEZFIyYVWn5KriIhEimZoEhERSbFcaLnmwjmIiIhEilquIiISKbkwWljJVUREIiUXulSVXEVEJFJiZH/TVclVREQiJRe6hXOh9S0iIhIparmKiEik5EKrT8lVREQiJRe6hZVcRUQkUjSgSUREJMVyoeWaC13bIiIikaKWq4iIREoutPqUXEVEJFLULSwiIpJiMSypVyLMbLaZfWNmE81sfLiupZmNNrPp4c8WceWvNbMZZvadmR1T/TmIiIjUToe5+17u3i9cHgK85+7dgffCZcysF3A60Bs4FnjAzPKqqljJVUREIiVmyb2ScDIwInw/Ajglbv1z7r7R3WcBM4D9qjyHpMIQERFJMUvylSAH/mNmX5jZ4HBdG3dfBBD+bB2ubw/Mi9t3friuUhrQJCIikZLsgKYwWQ6OWzXM3YeVK3aQuy80s9bAaDP7tqoqK1jnVcWg5CoiIpGS7AxNYSItn0zLl1kY/lxqZq8QdPMuMbO27r7IzNoCS8Pi84GOcbt3ABZWVb+6hUVEpFYxs0Zm1qT0PXA0MAl4HRgYFhsIvBa+fx043czqmVkXoDvwWVXHUMtVREQiJQ33ubYBXjEzCPLgM+7+tpl9DowyswuBucAAAHefbGajgClAEXCpuxdXdQAlVxERiZSa7lJ195lAnwrWLweOqGSfW4FbEz2GkquIiERKDkzQpOQqIiLRErPsT68a0CQiIpJiarmKiEikZH+7VclVREQiRslVqrXrzw7myg/+vdX69atWcWWLTmXWddl/X0646Vq6HNCPvLp1WTZzNm/dOpTxz79UYd3HDLmSU2+7iRkfjWXoT6t9SIOkWF6/A6l3xnnEdt4Fa9wUX72S4slfsXHEw5TMmVnhPvWvuI78E3/JptH/YsNtf9qy3tq0pf5vriGvWw+seQt8QyEls75n03PDKfrs43SdkmyjDz8ZyyPDn+T7mbNYvWYtLVu0YO8+e3DZxYPo1nWXTIeXtZRcJWHPXXY1cz6fsGW5uKiozPbdjzuGS14ZyefPvMDjZ15E0aZNtO3Vk7r161VY3w5dOtP/ut+zZsnSCrdLzbOmzSieNpVNr72Ar1pJrM1O5J9xPo3uG0HBRafhSxaVKZ/Xuw91jzwOL1i7dV0NGuCrV7Hx8fsp+WEJ1rAxdY8/lYa338f6G6+i6MMx6Tot2QarV6+h9267ceaAX9KyRQsWLl7MI0+M4LSBF/LGqGdo365tpkOUDFFyTZPFU79j1qefV7itXuPGnPvEA/z3gUd54YohW9Z/+94HldZ3xoN38dnIUezUozuxOvo1ZkLRmLcpGvP2luVioPjbSTQe8Sp1DzmSTS889WPhvDrUv/JPbHz6MfJP/MVWdZXMnsmGoTeXrX/ch+Q98yZ1jz1JyTWiTuh/DCf0L9trtOfuvel/6gDeeXcMF5x7VoYiy26m0cKSCvsMOIWmrXfk3X/cm1D5fc8YQKe+fXj12purLyxp5atXB2/K9Uzk/+pciMXY9MKTiVdWUoyvK9iqLom25s2aAVBHX3q3W5qeilOjlFzT5IKRj/JA0UqGLpvNBSMfo0XHDlu2dT34QAqWr6D9Hr25/uux3L95BX+dO4XjbxiCxcr+iho2b86Au27j5WuuZ/3Klek+DalILAZ16hBr34n6V/6JkuU/sPn9H1u01q4D9c6+iA1331Z9ojSDWB7WohX5Zw8i1mFnNr06qoZPQJJVXFzMps2bmT1nLjfechs77tCK4489KtNhZa1Ykq8o0FerGla4eg2jh97DtP9+xIY1a+m4dx+O/eNV/GHsu9y698Gs/WEZzdu1Jb9hAy545lH+/Ze/M/eLifQ88jCOu/4aGjZvxgtXXrulvp/f8ReWTJvB2OEjM3hWEq/R/U+R16MXAMXz57L+qovxVT9+8Wnwu+vY/OEYiieOr7auehf/jnqnnQuAr19H4S1DKP6yyvnBJQIGnHM+k6cGTyzbuWNHRjz8AK1atsxwVNkrB3qF05tczawhcBXQyd0HmVl3oIe7v5nOONJp3sSvmTfx6y3L0//3MdP/9zFDPnufwy7/Na9f/xcsFiO/QQNeu+7PvHfX/QBM++9HNGrVgp9dOog3brqNDWvW0O3gAzng3DP4a9+fZup0pAKFt/0Ja9goaKGedg4N73iQdZefjy9ZRN0jjyOvZ28KB56aUF2bXhzJ5jHvEGvZirpHn0CD6/5K4U1XUzTuwxo+C0nGHbfcTMG6dcybv4DHnxrJ+b++jGeeGEaHdu0yHZpkSLpb0E8AG4EDw+X5wC2VFTazwWY23szGT2FTOuJLi3lffsXSaTPYed++AKxbvgKAqaPfL1Nu6n/GUCc/n3a9ewJw1sN38/FjT7Jy/kIaNGtGg2bNiNWpQywvjwbNmlEnPz+9JyIAlMydRfG3kyga8zbrrroYa9CQemdcAPUbUO/XV7Hx2eH4po3QqHHwMsPq1Ane55X9fuvLllIybQpF4z6k8M9/oHjKN9S75IoMnZkkqusuXeizx+6c0P8Yhj98P+vXr2fY4yMyHVbWsiT/i4J0dwt3dfdfmdkZAO5eaFUMC4t/4O0l1rTKp75nHTPw4JQWTp4arHPfugzgJcH6tr160rZXT37264u2qu6uVfMY9bshjLn7gZqLWaq3roCSBfOIte+INWtOrEVL6g+6jPqDLitTLNamLXUPO4b1119B0ccfVFpd8bQp5P/izBoOWlKpaZMmdOrYgbnz5mc6lKwVjfSYnHQn101m1gBwADPrStCSrVU67bM3bXbtxhejXgHgq1ff5ORbrqf3sUf+mGiBXsccwabCQhZOmgLAnYcet1VdA/55O7G8PJ6/7GqWzqh44gJJH2vRklinzmx+99/4iuWsu2LrL0INrr+dkpkz2DjyUUpmfV9FZUbe7ntRslD/k84my5YvZ9bsOZx43LGZDiVrKbluuxuBt4GOZjYSOAg4L80xpNUFTz/KsllzmDthIoWrVtNx7z059tqrWLVgIe/f+xAQtFw/eeJpTvzzdVgsxtwJX9HzyEM5+KKB/Psvf2fjunVAcB22vMJVq4nVqVPhNqlZDf78D4qnfUvJzOn4+gJiHXYm/5dnQXFxcI/r5k0Uf/XF1jtu2kTJyuVlttUbeDE0aUbxpIn4iuVYy1bkH3cKeT13p/DWP6bxrGRbXHrl1fTq2ZMe3bvRuHEjZs+Zy/CRz5KXl8f556jHYXul4WHpNS6tydXdR5vZBOAAgi8nv3X3ZemMId0WTJrCvmf8ksMuG0x+w4asXryEL19+nTdu/OuWa60AIy/+LasWLOLQyy6maZvWLJ89lxev/CNj7nkwg9FLVYqnfEPdQ48idto5UKcOJT8soXjieNY/8/hWszNVW9f0b8n/xZnUPewYrFFjfOVyir+fxvrfXkDx5K9q6AwkWX322J23R7/HE0+NZHPRZnZq04b9++3D4AsGajBTLWde/jpfTR7M7CBgoruvM7Ozgb7A3e4+p7p9c+6aq2zx98O6ZjoEqSFN3/wg0yFITWnYrMbal2/t0C6p/9/3X7Yw423fdI8WfhBYb2Z9gKuBOcA2TFkjIiK5TjM0bbsiD5rKJwP3uPvdQJM0xyAiIhFmltwrCtI9oGmtmV0LnA0cYmZ5QN00xyAiIlKj0t1y/RXBrTcXuvtioD1wR5pjEBGRCMuFbuF0jxZeDNwZtzwXXXMVEZE4scikyO2XluRqZmsJJ44ovwlwd2+ajjhERCT6sj+1pim5ursGLYmISEKiMigpGRl55JyZtQbqly6H3cMiIiI5Ia0DmszsJDObDswC/gvMBt5KZwwiIhJtuTCgKd2jhf9CMPXhNHfvAhwBfJzmGEREJMJy4ZFz6U6um919ORAzs5i7vw/sleYYREQkwmKW3CsK0n3NdZWZNQb+B4w0s6VAUZpjEBGRCItIfkxKWlquZtYpfHsysB64guDRc98DJ6YjBhERkXRJV8v1VaBv+DScl9z9F8CINB1bRESySC60XNOVXOM/q13SdEwREclCURmUlIx0JVev5L2IiEgZmkQicX3MbA1BC7ZB+B40/aGIiOSgdE1/mJeO44iISPZL9z2iNSEj0x+KiIhUJgd6hZVcRUQkWiwHLroquYqISKRkf2rNja5tERGRSFHLVUREIiUXWq5KriIiEim65ioiIpJiUXmyTTJ0zVVERCTF1HIVEZFIsRxouiq5iohIpOTAJVclVxERiRYlVxERkRTLhdHCGtAkIiKSYmq5iohIpORAw1UtVxERiRYzS+qV4DHyzOxLM3szXG5pZqPNbHr4s0Vc2WvNbIaZfWdmxyRSv5KriIhEillyrwT9FpgatzwEeM/duwPvhcuYWS/gdKA3cCzwgJlV+4xyJVcREYmUmFlSr+qYWQfgeODRuNUnAyPC9yOAU+LWP+fuG919FjAD2K/ac0j8dEVERKLPzAab2fi41+ByRf4JXAOUxK1r4+6LAMKfrcP17YF5ceXmh+uqpAFNIiISKckOaHL3YcCwiuu2E4Cl7v6FmR2aSDgVHaK6nZRcRUQkUmr4PteDgJPM7DigPtDUzJ4GlphZW3dfZGZtgaVh+flAx7j9OwALqzuIuoVFRCRSLJbcqyrufq27d3D3zgQDlca4+9nA68DAsNhA4LXw/evA6WZWz8y6AN2Bz6o7B7VcRURE4HZglJldCMwFBgC4+2QzGwVMAYqAS929uLrKlFxFRCRS0jX9obt/AHwQvl8OHFFJuVuBW7elbiVXERGJlFyYoUnJVUREIiUXJu5XchURkUjJgdyq0cIiIiKpppariIhESiJTGEadkquIiERKDuRWJVcREYkWDWhKo/vvODfTIUgNGXvHq5kOQWrIwZkOQLJSDuRWDWgSERFJtaxpuYqISO2QCy1XJVcREYkUi2V/dlVyFRGRSMmFlquuuYqIiKSYWq4iIhIpmkRCREQkxXIgtyq5iohItGgSCRERkRTLgdyqAU0iIiKpppariIhEirqFRUREUiwHcquSq4iIRItariIiIilmOTAaKAdOQUREJFrUchURkUhRt7CIiEiq6ak4IiIiKZYDLVddcxUREUkxtVxFRCRSdM1VREQk1XTNVUREJMVyueVqZj9PtBJ3fzk14YiISG1nOd5yfTHBOhzIS0EsIiIiOaHS5OruGkksIiLpl8vdwiIiIpmQC93CCbdOzay/mf3LzKaaWcdw3UVmdkTNhSciIrWOWXKvCEgouZrZWcAoYBrQGagbbsoDrqmRyEREpHaKWXKvCEi05XoNMMjdrwCK4taPA/ZKdVAiIiLZLNFrrt2BsRWsLwCapi4cERGp7WrTDE0LgV2BOeXWHwJ8n9KIRESkdotI124yEu0WHgbcY2YHhcsdzWwg8HfgwRqJTEREaqccGNCUUMvV3f9uZs2A0UB94H1gIzDU3e+vwfhERESyTsL3ubr7dWZ2K9CLoMU7xd0LaiwyERGplSwHpjDa1kkkHNgQvi9OcSwiIiKR6dpNRqL3udYzs38CK4CvgK+BFWZ2t5nVr8H4RESklrGYJfWKgkRbrg8CRwMX8eMtOQcCtwFNgAtSH5qIiNRKOdByTTS5DgB+7u6j49bNNLOlwEsouYqIiGyRaHJdByyoYP0CoDB14YiISK0Xka7dZCQ6Jute4EYza1C6Inx/fbhNREQkJcwsqVcUVNpyNbPXy606FFhgZl+Hy3uE+zeqmdBERKRWyoGWa1XdwsvLLb9UbnlWimMRERGp8QFN4V0u/wPqEeTBF939RjNrCTxP8PS32cBp7r4y3Oda4EKC21Avd/d3qjpGpcnV3c9PwTmIiIhEzUbgcHcvMLO6wEdm9hbwc+A9d7/dzIYAQ4A/mFkv4HSgN9AOeNfMdnX3Sud7yIF5MEREJJfU9DVXD5TOMFg3fDlwMjAiXD8COCV8fzLwnLtvdPdZwAxgv6qOkfAMTWZ2PnAG0AnILxfoLonWIyIiUqUkr7ma2WBgcNyqYe4+rFyZPOALoBtwv7t/amZt3H0RgLsvMrPWYfH2BM8vLzU/XFephJKrmV0NXAs8TPCYuQfCgA4BhiZSh4iISCKSHfEbJtJh1ZQpBvYys+bAK2a2e1UhVVRFVfUn2i08CBjs7tcCm4H73P0k4B/AzgnWISIiEinuvgr4ADgWWGJmbQHCn0vDYvOBjnG7dSB4znmlEu0W7gB8Fr4vBJqG758N1w9KsJ7aZ9e9iPXsBzt1ggaNYe1KfPpX+Lh3YPPGoEynXbHdD8DadYFGzWDdanz2t/gn/4L1ZR88lPf7+yo8TPGI2+CHiub5kHTp/ezTtDj8UObedTdzb78DgEZ77kHna/9Aw916UrdFc4rWrKHg60nMu+ufrB0/ocz+9Tp1pMsNf6L5IQdjdeuy9suJzL75Fgq++rqiw0kEfDr+C84d9Out1jdp3JjxH47JQEQ5ooZvxTGzHYHN7r4qnLPhSOBvwOvAQOD28Odr4S6vA8+Y2Z0EA5q682NOrFCiyXUxsAMwF5hDMK/wRIKu4SqbxrVdrN8R+NqV8OHr+NpVWOsO2E+Owzp2p+SZOwEn1udgyK+Hj3sHX7UMa7Ej9pPjsc67UTLir7B5U5k6SyaNw7/6qOyBVi5FMmeHU0+mUe/dtlpfp1lTCmfNZsnzo9i0ZCl1d2hF+4sHsccrL/L1ST+n4MuJQbkWzdnz9VcoLihgxtVDKCkspN0lg9n95VF8dewJFE6fkeYzkm3xp2uuYo/evbYs5+XlZTCaHFDzE0G0BUaE111jwCh3f9PMxgKjzOxCgnw3AMDdJ5vZKGAKUARcWtVIYUg8uY4BTgImAI8Bd5nZaUBfYNS2n1ftUfLKw1D4Y+vT58+ADeuJHXcudOwO86ZR8u6orcr4yqXknX4F1qMvPmlc2UrXroJFs9NzAlKtvKZN2eXPNzLzhpvp+dD9Zbat/vBjVn/4cZl1q8Z8wP5Tv6b1gF9sSa5tzzuX/B134ItTfsmG2bODch99TL/PPqHT1Vfx3eCtW0cSHV27dGGvPffIdBg5o6afbOPuXwN7V7B+OXBEJfvcCtya6DESTa6DCa/PuvtDZrYSOIhgYomHEz1YrVS49fPkffEcAKxJs6DZX0EZFs8NfjZuXmOhSWp0ueE61n83jWWvvAblkmtFitevp2TTJnzz5i3rmuzTl8KZs7YkVoCS9YWsGfcZLY86EvLyoFiPUJZaIiJTGCYjoeTq7iVASdzy8wSzWGwzM2vk7uu2Z99cYR27A+DLl1ReqEO3oMyKxVvvv9dPsX2PAHdYNIuSj/8NC76vkVilak3325fWA37Bl4cfXXVBMywvj/w2relw2aUALB757JbNXlxcJtmWKtm0ibyGDWjQeWcKv5+Z0tgldX5/3Q2sXLWKpk0ac/CBB3DV5b+hXdudMh2WZFBVcwv3TbQSd59QXRkz+wnwKNAY6GRmfYCL3f3/Ej1OTmjcDDvoeHz2t7BkbsVl6tYjdtgv8OWLYHrZwSwlkz+DmZPwgtVY05bYvkcQO+1ySl68D+ZNT8MJSCmrU4eud9zO/Acfrjbx9XzkIXY48XgANv3wA1POPJfCaT/+vgpnzKT5IYdQp0VzilauCg9gNNl7LyC4JivR06RxYy445yz23acvjRs1Ysp33/HwY8P5bOCFvPrcU7Rq2TLTIWanHJ9beDzBYKXqztKBRK7e3wUcQzDqCnf/yswOSSTInFE3n9gpF0NJCSVvP11xGYsRO+F8aNI8GPDkJWU2+1tP/vh+wff4jK+JnX8dsYNOoOS5u2oyeimnw2X/R16D+sz/Z/UPhpr1l1uZf98D1GvXjrYXDKTX08OZNOCMLSOBFz35FO0uOp9d772bmdfdQHFhIR1/dzn1OwWj/71E4wajqFfPHvTq2WPL8n79+rJv370ZcM75PPns81xxqa6Vb4+oPNkmGVXd59oF2CX8WdUr4dmZ3H1euVVVXkQys8FmNt7Mxj8ybnKih4mmvDrETr0YmrWi5MX7oWBVBYUM638O7NyDkleHwbIqb6MKbN6Iz5wU3OojaVOvfTs6/PZy5vxtKJafT17TpuQ1De5Qi+XXC97Hfvzz2jhnLgUTv2L5v99i8hnnsHnZcnYeck2Z7d/93+U07rMH/T77mP2/mUDTfn1Z8PCjAGxeUsUlBImU3rv1pHOnTkyaPCXToWSvmCX3ioCqJu6fk+JjzQu7ht3M8oHLgalV7RA/y0bx0N9k71f3WIzYyRfBTp0peeHeSpOmHXU61rMvJa8/BnOnbcMBovGPqTapv/PO5DWoT48Htm61drj0EjpceglfHn406yr4H6xv3sy6KVNptHvvMuuX/+vfLH/rbRp03QXftJkNc+bQ9W9/ZcP8BWxckMAXLYkMx3NiUE7G5MBnl/DcwilwCXA3wXyM84H/AJem8fgZYtjx50GnHpS8/GClt9DYoadiex6Iv/UUzNiGSQPy62O79NatOWlWMGky35w6YKv1e7zyAktfeIklzzxH4azZFe4ba1Cfxn32pPD7CgahlZRsuac1v00bdjj5JBY88GAqQ5ca9s3kKcyeM5f+Rx2Z6VAkg9KWXN19GXBWuo4XFXbkacR69KVk7NvBZBBtO/+4ce0qKFiF7XcksX5HUPLNJ/jKH8qWWV8Aq5cFdfU7Alq2hrnT8XXhgKZ+R0CjppT8awSSPsVr1rD6k7EVbtswf/6WbV3vuJ2iVasomPg1m1esoH6H9rS98Hzy27Rm2m9+u2Ufq1OHzjdcx+qx4yheu5aGPXrQ4fJLWf/ddyx4sMopUiWDrvrj9XRo347ePXvSpEljpn47jYefGE6b1jty9umnZTq87KWWa+LM7AkqmM3J3S9IVwyZYF2CWVtiBx4LBx5bZlvJJ//GP/k31iXoHozt8RPY4ydly0wah4eDn3zFEmLd94RufbB6DWDTBlgwk5J3RsLiVPfiSyoUTPiSNmedwU5nn0VewwZsXLyYtRMmMv2K37N+6rdbyrk7DXbpwo4/P4U6TZuycdEiljz7PPPvvrfCW3QkGnbt2pU33/kPTz83ig0bNrBDq1YcffhhXHbJYFpqhPf2y4Hkau7puZRpZr+IW6wPnAosdPfLE9k/q6+5SpXG3vFqpkOQGnLwrCwfiCiVa9isxjJg0RWnJvX/+zp3vZLx7LxNLVcz2wHoCkx0943bsq+7v1SurmeBd7elDhERkWyQ0CPnzKxJOGnxUuATwofEmtlDZnbTdh67O8GD10VERH5kltwrAhJ9nuvfCBJqX4JHzpV6k6B7t1pmttbM1pS+gDeAP2xLsCIiUgvkQHJNtFv4JOBUd59oZvF94VNJYBIJC6bb6O3ulcz3JyIiEopIgkxGosm1BbC8gvVNqGaWJQB3dzN7BdhnG2ITEZHaKJZop2p0JXoGnxO0XkuVtl4vJrgGm4hxZrZvooGJiIhkq0Rbrn8E3jGz3uE+V4bv9wOqnHzfzH7j7vcBhwEXm9kcYB3BnH3u7ntud/QiIpJ7aku3sLt/Es4L/Hvge4IntU8ADnT3b6rZ/QLgPqB/MoGKiEgtUVuSK0CYRAdu74Fq4EEAIiKSi2pLcjWzKp/46+4rqti8Z3jrzVbVBrt600RiEBERyRaJtlyXUcG8wHGqelj6N+6+d+IhiYhIrZYDo4UTTa6HlVuuC+wN/Br4U0ojEhGR2q22dAu7+38rWP2umc0ELgKeqWL3F7YnMBERqaVqS3KtwkSquRXH3f+a5DFERKQ2yYHkut0d22bWGPgdMC9l0YiIiOSAREcLr6XsgCYDGhJMBnFWgnXkuXu1UyWKiEjtZrVoQNNvyi2XAD8An7r7ygTrmGFmLwJPuPuURAMUEZFaJge6hatNrmZWB2gEvOruC5M41p7A6cCjZhYDHgeec/eK7oEVEZHaKgeSa7Vtb3cvAu4guP1mu7n7Wnd/xN1/AlwD3AgsMrMRZtYtmbpFRCSH5MDzXBPt2B5Hko+LM7M8MzspfPTc3cA/CJ4F+wbw72TqFhERiZJEr7k+Agw1s07AFwQDmbZw9wkJ1DEdeB+4w93jH1P3oplVeTuPiIjUIrk+oMnMHie43aZ0kog7KyjmVD39Yak93b2gog3ufnkC+4uISG0Qka7dZFTXch0IDAG6bO8BzOxewtt4rIIPTIlVRETKqAXJ1SDpx8WNT2JfERGRrJPINdeqnoZT/c7uIwDMbIC7l5ln2MwGJFO3iIjkoBxouSZy1XixmRVX9UrwWNcmuE5ERGqzWCy5VwQk0nIdDKza3gOYWX/gOKC9md0Tt6kpULS99YqISI7KgZZrIsn1DXdfmsQxFhJcdz2J4DaeUmuBK5KoV0REclEtSK5JXW8FcPevzGwScHTp9VcREZFcltBo4WS5e7GZtTKzfHfflIo6RUQkR0Xkumkyqkyu7p7KM5wDfGxmrxM3w5O7VzQxhYiI1Fa1oFs4lRaGrxjQJI3HFRGRbKLkmjh3vzldxxIRkSym5Jo4M9uR4FFzvYH6pevd/fB0xSAiIpIO6bxqPBL4lmCe4puB2cDnaTy+iIhkgxyYRCKdUbRy98eAze7+X3e/ADggjccXEZFskAMPS0/ngKbN4c9FZnY8weCmDmk8voiIZIOIJMhkpDO53mJmzYCrgHsJpj/UDE0iIpJzajy5mll94BKgG9AeeMzdD6vp44qISJayaFw3TUY6zmAE0A/4BugP/CMNxxQRkWwVs+Re1TCzjmb2vplNNbPJZvbbcH1LMxttZtPDny3i9rnWzGaY2Xdmdkx1x0hHt3Avd98DwMweAz5LwzFFRCRb1XzLtQi4yt0nmFkT4AszGw2cB7zn7reb2RBgCPAHM+sFnE5wK2k74F0z29XdK33kajparqUDmXB3PWJORESqVsOjhd19kbtPCN+vBaYSXLY8maC3lfDnKeH7k4Hn3H2ju88CZgD7VXWMdLRc+5jZmvC9AQ3CZQPc3ZumIQYREaklzGwwwbPISw1z92GVlO0M7A18CrRx90UQJGAzax0Waw+Mi9ttfriuUjWeXN09r6aPISIiOSTJiSDCRFphMo1nZo2Bl4Dfufsaq7zVW9GGKh/Jms5bcURERKqXhvtczawuQWId6e4vh6uXmFnbsNXaFlgarp8PdIzbvQPBXA2Vyv7xziIiklssltyruuqDJupjwNRyjz19HRgYvh8IvBa3/nQzq2dmXYDuVDM4Vy1XERGJlppvuR4EnAN8Y2YTw3V/BG4HRpnZhcBcYACAu082s1HAFIKRxpdWNVIYlFxFRKSWcfePqPg6KsARlexzK3BrosfImuSad/6QTIcgNeSg0y7JdAgiEiURebJNMrImuYqISC2hiftFRERSTHMLi4iISHlquYqISLQkMPl+1Cm5iohItORAt7CSq4iIRIsGNImIiKRYDrRcs/8MREREIkYtVxERiRYNaBIREUkxXXMVERFJsRy45qrkKiIi0ZID3cLZ//VAREQkYtRyFRGRaFG3sIiISIppQJOIiEiK5UDLNfvPQEREJGLUchURkWjJgdHCSq4iIhItOdAtrOQqIiLRogFNIiIiKRbL/pZr9p+BiIhIxKjlKiIi0aJuYRERkRTTgCYREZEUU8tVREQkxTSgSURERMpTy1VERKJF3cIiIiIppgFNIiIiKZYDLdfs/3ogIiISMWq5iohItKhbWEREJMX0yDkREZEUU8tVREQkxTSgSURERMpTy1VERKJF3cIiIiKpZTnQLazkKiIi0aKWq4iISIrlQHLN/jMQERGJGLVcRUQkWjSJhIiISIrlQLewkquIiERLDowWzv6vByIiIhGjlquIiESLuoVle3w47nMeefo5vp89h9VrC2jZvBl779Gbyy48l25dOle4zw1/u5PnX/sXJx59BENv+mN6A5aEnXPVDXz+9eQKtx3cby8eve16CtYXcv9To5g07XumzJjJuvWFjBh6M/v32T3N0UqyPvxkLI8Mf5LvZ85i9Zq1tGzRgr377MFlFw+iW9ddMh1e9sqBbmEl1wxYvWYtvXvuypk/P4mWLZqzcPFSHnn6WU4bdBlvPPUo7du2KVN+wjeTeeM/79G4UaMMRSyJuvHyQRSsKyyzbuLU77j9oeEcfuC+AKxas5aX3xlDr25d+EnfPRn90aeZCFVSYPXqNfTebTfOHPBLWrZowcLFi3nkiRGcNvBC3hj1DO3btc10iNlJLVfZHiccfTgnHH14mXV79upJ/zPO4533/8sFZ562Zf3moiJuuP1OLhl4Fs+/+ma6Q5Vt1G3njlute+Gt0dStW4fjDj0IgPZtduTTl0cA8MmEr5Rcs9gJ/Y/hhP7HlFm35+696X/qAN55dwwXnHtWhiLLcjV8K46ZPQ6cACx1993DdS2B54HOwGzgNHdfGW67FrgQKAYud/d3qjtG9n89yBHNmzUFoE6dst93Hhv5PMUlJZx/xoBMhCVJ2rBxI2//byyHHdCP5k2bALkxb6pUrnmzZsDWf8sSKcOBY8utGwK85+7dgffCZcysF3A60Dvc5wEzy6vuAPrtZ1BxcTHFJSUsXLyEfzz4KDu2asnxRx62Zfvc+Qt5cPhIHh56K/l162YwUtleoz/6lHXrCznlqEMzHYrUoC1/ywsX8Y977mfHHVpx/LFHZTqs7FXD3cLu/j8z61xu9cnAoeH7EcAHwB/C9c+5+0ZglpnNAPYDxlZ1jLQlVwu+rp8F7OLufzazTsBO7v5ZumKImgGDfsPkb6cBsHOH9oy4dyitWrbYsv3GO+7iqJ8dzAH77J2pECVJr47+gFbNm3HIfn0zHYrUoAHnnM/kqd8CsHPHjox4+AFatWyZ4aiyWJK9O2Y2GBgct2qYuw+rZrc27r4IwN0XmVnrcH17YFxcufnhuiqls+X6AFACHA78GVgLvATsm8YYIuWOG4ZQsG498xYs4vFnR3H+b6/hmYfupkPbnXjt7dF8M/U73np2eKbDlO20ZNkKxn75Deecehx18qrtRZIsdsctN1Owbh3z5i/g8adGcv6vL+OZJ4bRoV27TIeWnZJsuYaJtLpkmqiKMr1Xt1M6r7nu7+6XAhsAwgvF+VXtYGaDzWy8mY0fNmJkOmJMq66dd6ZP79044ejDGX7PUNYXFjLsqWdZt76Q2+95iEFnn069/HzWrC1gzdoCStwpKi5mzdoCNhcVZTp8qcYb7/2PkpISTj3qsOoLS1bruksX+uyxOyf0P4bhD9/P+vXrGfb4iEyHJdtmiZm1BQh/Lg3XzwfiRyp2ABZWV1k6W66bw4vADmBmOxK0ZCtV5tvH8vnVflPIZk2bNKZTh/bMnb+QlatXs2LVKu586DHufOixMuUWLVnKW+99wP233cyRPzs4Q9FKIl579wN67tKZnl07ZzoUSaOmTZrQqWMH5s6bn+lQsldmBv29DgwEbg9/vha3/hkzuxNoB3QHqr2cmc7keg/wCtDazG4Ffgn8KY3Hj7RlK1Ywa85cTjz6CHZs2ZIn7/vHVmWuvOEWdu3ahUsGnkX3XbpkIEpJ1DffzWD67HkMueS8TIciabZs+XJmzZ7DiceVH4wqCavhAU1m9izB4KUdzGw+cCNBUh1lZhcCc4EBAO4+2cxGAVOAIuBSdy+u7hhpS67uPtLMvgCOIOjDPsXdp6br+FFy6ZAb6NWjOz267kLjRg2ZPW8+w597iby8PM4/YwD16uWzf9+9ttqvXn4+rVq0qHCbRMtr7/6XOnl5nHD4Tyvc/r/PJrB+w0amzZoDwOdfT2Hl6rU0rF9Pg5+yyKVXXk2vnj3p0b0bjRs3YvacuQwf+Wzwt3zOmZkOL3vFany08BmVbDqikvK3ArduyzHSOVq4E7AeeCN+nbvPTVcMUdGn9268Pea/PPHsC2zeXMRObXZk/737MPjcM+nQdqdMhydJ2lxUxL/e/4iD992LHVo0r7DMTfcMY+GSH7Ys3/fk8wC0a7MjY55+KB1hSgr02WN33h79Hk88NZLNRZvZqU0b9u+3D4MvGKjBTEnIhXvBzT09lzLN7BuC660G1Ae6AN+5e++EKsjxa661ma9blekQpIbYDlvPWCU5omGzGsuA/t24pP5/bz0OyHh2Tme38B7xy2bWF7g4XccXEZEsobmFt5+7TzCzWnuPq4iIVCIHuoXTec31yrjFGNAX+KGS4iIiUlup5bpNmsS9LwL+RTBDk4iIyI/Uck1MOHlEY3e/Oh3HExERyaQaT65mVsfdi8IBTCIiIlWr4ftc0yEdLdfPCK6vTjSz14EXgHWlG9395TTEICIi2ULdwtukJbCc4Kk4pfe7OqDkKiIiP9KApoS0DkcKT+LHpFpKE0OIiEjOSUdyzQMas53PxBMRkVpG3cIJWeTuf07DcUREJCcouSYi+z8lERFJH7VcE1LhI3xEREQqlAPJtcaHZLn7ipo+hoiISJRkbOJ+ERGRimV/y1XJVUREoiUHuoWVXEVEJFqyP7cquYqISNRkf3bN/jmmREREIkYtVxERiRZdcxUREUkxJVcREZFUy/7kqmuuIiIiKaaWq4iIRIu6hUVERFJNyVVERCS11HIVERFJsRxIrhrQJCIikmJquYqISMRkf8tVyVVERCLFcqBbWMlVRESiRclVREQk1bI/uWpAk4iISIqp5SoiItGibmEREZEUU3IVERFJtexPrrrmKiIikmJquYqISLSoW1hERCTFsj+3KrmKiEjUZH92VXIVEZFoyYFuYQ1oEhERSTG1XEVEJFpyoOWq5CoiIhGj5CoiIpJaarmKiIikWA4kVw1oEhERSTG1XEVEJGKyv+Wq5CoiItGSA93C5u6ZjkEqYGaD3X1YpuOQ1NPvNnfpdyuldM01ugZnOgCpMfrd5i79bgVQchUREUk5JVcREZEUU3KNLl23yV363eYu/W4F0IAmERGRlFPLVUREJMWUXNPEzIrNbGLcq3MVZQvSGJqE4n5Hk8zsBTNruA377mVmx8Utn2dm9yW473Vx/y7i/51cvj3nIallZq3ifieLzWxB3HJ+puOTaFK3cJqYWYG7N051WUmd+M/dzEYCX7j7nQnsVwc4G+jn7r8J150Xv7w9McStM4K/1ZJtqUtSz8xuAgrcfWjcujruXpS5qCSK1HLNEDNrbGbvmdkEM/vGzE6uoExbM/tfXGvqp+H6o81sbLjvC2amRJx6HwLdzOxEM/vUzL40s3fNrA0E/5M1s2Fm9h/gSeDPwK/C39WvSisxsyZmNsvM6obLTc1sdulyZcyss5lNNbMHgAlAx/geDTP7pZkND9/vaGYvmdnn4eugFH8WUo6ZDTezO83sfeBv4b+H38dtn1TaO2VmZ5vZZ+G/jYfNLC9TcUv6KLmmT4O4rqRXgA3Aqe7eFzgM+EfYQol3JvCOu+8F9AEmmtkOwJ+AI8N9xwNXpu0saoGwJdof+Ab4CDjA3fcGngOuiSu6D3Cyu58J3AA87+57ufvzpQXcfS3wAXB8uOp04CV335xAKD2AJ919b3efU0W5u4G73H1f4BfAownULcnbleDv8KrKCpjZbsCvgIPCv+Ni4Kz0hCeZpLmF06cw/OMCIGy5/NXMDgFKgPZAG2Bx3D6fA4+HZV9194lm9jOgF/BxmIvzgbHpOYWc18DMJobvPwQeI0hwz5tZW4LPelZc+dfdvTCBeh8lSMqvAucDgxKMZ467j0ug3JFAr7jvZk3NrEmY2KXmvODuxdWUOYLgS9jn4e+nAbC0pgOTzFNyzZyzgB2Bfdx9s5nNBurHF3D3/4XJ93jgKTO7A1gJjHb3M9IdcC1Q5gsQgJndC9zp7q+b2aHATXGb1yVSqbt/HHbz/gzIc/dJCcZTvv74ARLx/1ZiwIEJJnpJnfjfTxFlewJLfz8GjHD3a9MWlUSCuoUzpxmwNEyshwE7ly9gZjuHZR4haEX1BcYBB5lZt7BMQzPbNY1x1zbNgAXh+4FVlFsLNKli+5PAs8ATScSyxMx2M7MYcGrc+v8AWwZOmdleSRxDts9sgr9PzKwv0CVc/x7wSzNrHW5rGf5dS45Tcs2ckUA/MxtP0Ir9toIyhxJcZ/2S4Fra3e7+A3Ae8KyZfU2QbHumJeLa6SbgBTP7EFhWRbn3CbpmywxoijMSaEGQYLfXEOBNYAywKG795QT/lr42synAJUkcQ7bPS0DL8LLCr4FpAO4+hWCMxH/Cv9fRQNtMBSnpo1txRNLAzH5JMPjpnEzHIiI1T9dcRWpYeN22P3BcdWVFJDeo5SoiIpJiuuYqIiKSYkquIiIiKabkKiIikmJKriLlhPPG3hK+/6mZfZem43rp/csVbPvAzC5KsJ7ZZnbkdsaw3fuKyI+UXCUrhUmg0MwKzGyJmT1REw8wcPcP3b1HAvGcZ2Yfpfr4IpKdlFwlm50YPp6tL7Avwc36ZYST8IuIpJWSq2Q9d18AvAXsDlu6Vy81s+nA9HDdCeHsSavM7BMz27N0fzPbO3x831oze564eXvN7FAzmx+33NHMXjazH8xsuZndFz755CHgwLAlvSosW8/MhprZ3LB1/ZCZNYir62ozW2RmC83sgkTP18y6mtmY8PjLzGykmTUvV2xfM5tiZivDVn38OVX6WZQ7zn5mNt7M1oTxV/tsWxEJKLlK1jOzjgQTNHwZt/oUYH+CKQn7Ao8DFwOtgIeB18Pkl0/wtJqngJbACwRTTVZ0nDyC6QfnAJ0JnmT0nLtPJZhycKy7N3b35uEufyN4LNleQLew/A1hXccCvweOAroTPNkm4VMGbgPaAbsBHSn7QAEIptQ8BugaxvCn8LiVfhYVHOdugik3m4b1jNqGGEVqNSVXyWavhq3Ej4D/An+N23abu68InxQzCHjY3T9192J3HwFsBA4IX3WBf7r7Znd/keBRfxXZjyChXe3u69x9g7tXeJ3VgueLDQKuCONYG8Z3eljkNOAJd5/k7uvYOjlWyt1nuPtod98YzjV9J/CzcsXuc/d57r4CuBUofYpSVZ9FeZsJHhi/g7sXJPj4OxFB0x9KdjvF3d+tZNu8uPc7AwPN7LK4dfkEidKBBV52qrLKHkzekeAZq0UJxLYj0BD4Iu45qwbkhe/bAV8kcMythE9YuQf4KcGTeGIEjyKMF3/+c8LjQdWfRXkXAn8GvjWzWcDN7v5monGK1GZquUquik+W84Bb3b153Kuhuz9L8HSZ9haXAYFOldQ5D+hUySCp8vOILgMKgd5xx2wWDsAiPG7HBI5ZkdvC4+0ZdtmeTZC445Wve2HcOVT2WZQ9Iffp4XODWxN0cb9oZo22IU6RWkvJVWqDR4BLzGx/CzQys+PNrAkwluBB15ebWR0z+zlB929FPiNIireHddQ3s4PCbUuADuE1XNy9JDzuXXHP8mxvZseE5UcB55lZLzNrCNy4DefTBCgAVplZe+DqCspcamYdzKwl8Efg+QQ+izLM7Gwz2zE8l1Xh6uJtiFOk1lJylZzn7uMJrjXeR9B9OoPgmbi4+ybg5+HySuBXwMuV1FMMnEgwOGkuMD8sD8EzVicDi82s9LmvfwiPNc7M1gDvAj3Cut4C/hnuNyP8maibCW4/Wg38q5J4nyF4iPrM8HVLdZ9FBY4FJptZAcHgptPdfcM2xClSa+mpOCIiIimmlquIiEiKKbmKiIikmJKriIhIiim5ioiIpJiSq4iISIopuYqIiKSYkquIiEiKKbmKiIikmJKriIhIiv0/9vysu8HzsDEAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 576x432 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#https://stackoverflow.com/questions/19233771/sklearn-plot-confusion-matrix-with-labels/48018785\n",
    "#Plotting a confusion matrix for the Actual vs Predicted labels\n",
    "plt.figure(figsize=(8,6))\n",
    "labels = ['False', 'Partly True', 'True']\n",
    "cm = confusion_matrix(y_test, predictions)\n",
    "ax= plt.subplot()\n",
    "sns.heatmap(cm, annot=True, annot_kws={\"size\": 16},fmt='g',cmap='Reds') #annot=True to annotate cells\n",
    "#Setting the x-label\n",
    "ax.set_xlabel('Predicted labels',fontsize=12)\n",
    "#Setting the y-label\n",
    "ax.set_ylabel('True label',fontsize=14)\n",
    "#Setting the title\n",
    "ax.set_title('Confusion Matrix',fontsize=14) \n",
    "ax.set_xticklabels(labels)\n",
    "ax.set_yticklabels(labels)\n",
    "#Displaying the matrix\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**From the confusion matrix heatmap, we can determine that our model works quite good for False claims. This emphasizes the need of more datapoints for the other two categories to build a model upon**"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Cosine Similarity Implementation with text summarization"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The part for concatinating and cleaning of the articles is taking a lot of time as we are directly performing the text preprocessing on the full article, to speed up the process we can just take the top 5 sentences having the highest cosine similarity with the claim. After we get the top 5 sentences then we preprocess those to get the clean article column. Part for cleaning the claims is same as before. Articles, claims and claimants are vectorized together and then split for model implementation. But here I will only show XGBoost model as it performed the best among all the models implemented."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Defining a new function to concatenate and select top 5 important sentences in the articles\n",
    "\n",
    "def imp_data(ind):\n",
    "    concat_article = []\n",
    "    \n",
    "    for i in range(len(claims.iloc[ind, -2])):\n",
    "        article_number = claims.iloc[ind, -2][i]\n",
    "        with open(os.path.join(ARTICLES_FILEPATH, '%d.txt' % article_number), 'r', encoding=\"utf8\") as f:\n",
    "            a = f.read()\n",
    "        concat_article.append(a)\n",
    "    new = []\n",
    "    for a in range(len(concat_article)):\n",
    "        sp = concat_article[a].split('\\n')\n",
    "        new.append(sp)\n",
    "\n",
    "    sentences = sent_tokenize(str(new))\n",
    "\n",
    "    # append claim to articles\n",
    "    sentences.append(claims.iloc[ind, 0])\n",
    "\n",
    "    # vectorize sentences based on tf-idf\n",
    "    vectorizer = TfidfVectorizer()\n",
    "    X = vectorizer.fit_transform(sentences)\n",
    "\n",
    "    # measure similarity between claim and each sentence\n",
    "    if X.shape[0] == 2:\n",
    "        similarity = cosine_similarity(X[1], X[0]) \n",
    "    else:\n",
    "        similarity = cosine_similarity(X[-1], X[:-2])\n",
    "\n",
    "    # find top 5 sentences with greatest similarity\n",
    "    sorted_index = np.argsort(similarity)[0][::-1]\n",
    "    top_sentences = []\n",
    "    for i in range(0,min(5, len(sorted_index))):\n",
    "        top_sentences.append(sentences[sorted_index[i]])\n",
    "    \n",
    "    return(' '.join(top_sentences))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Execution time for concatinating and summarizing the articles =  294.30055713653564\n"
     ]
    }
   ],
   "source": [
    "#Applying the above function to our dataframe\n",
    "start = time.time()\n",
    "important_sentences = [imp_data(i) for i in range(len(claims))]\n",
    "print('Execution time for concatinating and summarizing the articles = ', time.time()-start)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Defining a function to clean the articles\n",
    "\n",
    "def clean_data(new):\n",
    "    \n",
    "    table = str.maketrans(string.punctuation, ' '*len(string.punctuation)) #map punctuation to space\n",
    "    stripped = new.translate(table)\n",
    "    tokens = stripped.lower().split()\n",
    "    \n",
    "    words = [word for word in tokens if word.isalpha()]\n",
    "    stop_words = set(stopwords.words('english'))\n",
    "    words = list(filter(lambda x: x not in stop_words, words))\n",
    "            \n",
    "    stemmatizer = SnowballStemmer('english')\n",
    "    stemmatized = [stemmatizer.stem(word) for word in words]\n",
    "    return(' '.join(stemmatized))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Execution time for cleaning the articles =  22.572611093521118\n"
     ]
    }
   ],
   "source": [
    "#Applying the above function to our important sentences\n",
    "start = time.time()\n",
    "clean_articles = [clean_data(important_sentences[i]) for i in range(len(claims))]\n",
    "print('Execution time for cleaning the articles = ', time.time()-start)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Execution time for cleaning the claims =  7.119403600692749\n"
     ]
    }
   ],
   "source": [
    "#Applying the above function to our claims\n",
    "start = time.time()\n",
    "clean_articles = [clean_data(claims.iloc[i, 0]) for i in range(len(claims))]\n",
    "print('Execution time for cleaning the claims = ', time.time()-start)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "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>claim</th>\n",
       "      <th>claimant</th>\n",
       "      <th>date</th>\n",
       "      <th>label</th>\n",
       "      <th>related_articles</th>\n",
       "      <th>id</th>\n",
       "      <th>clean articles</th>\n",
       "      <th>clean claims</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>A line from George Orwell's novel 1984 predict...</td>\n",
       "      <td></td>\n",
       "      <td>2017-07-17</td>\n",
       "      <td>0</td>\n",
       "      <td>[122094, 122580, 130685, 134765]</td>\n",
       "      <td>0</td>\n",
       "      <td>line georg orwel novel predict power smartphon</td>\n",
       "      <td>line georg orwel novel predict power smartphon</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Maine legislature candidate Leslie Gibson insu...</td>\n",
       "      <td></td>\n",
       "      <td>2018-03-17</td>\n",
       "      <td>2</td>\n",
       "      <td>[106868, 127320, 128060]</td>\n",
       "      <td>1</td>\n",
       "      <td>main legislatur candid lesli gibson insult par...</td>\n",
       "      <td>main legislatur candid lesli gibson insult par...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>A 17-year-old girl named Alyssa Carson is bein...</td>\n",
       "      <td></td>\n",
       "      <td>2018-07-18</td>\n",
       "      <td>1</td>\n",
       "      <td>[132130, 132132, 149722]</td>\n",
       "      <td>4</td>\n",
       "      <td>year old girl name alyssa carson train nasa be...</td>\n",
       "      <td>year old girl name alyssa carson train nasa be...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>In 1988 author Roald Dahl penned an open lette...</td>\n",
       "      <td></td>\n",
       "      <td>2019-02-04</td>\n",
       "      <td>2</td>\n",
       "      <td>[123254, 123418, 127464]</td>\n",
       "      <td>5</td>\n",
       "      <td>author roald dahl pen open letter urg parent c...</td>\n",
       "      <td>author roald dahl pen open letter urg parent c...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>When it comes to fighting terrorism, \"Another ...</td>\n",
       "      <td>Hillary Clinton</td>\n",
       "      <td>2016-03-22</td>\n",
       "      <td>2</td>\n",
       "      <td>[41099, 89899, 72543, 82644, 95344, 88361]</td>\n",
       "      <td>6</td>\n",
       "      <td>come fight terror anoth thing know work base l...</td>\n",
       "      <td>come fight terror anoth thing know work base l...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                               claim         claimant  \\\n",
       "0  A line from George Orwell's novel 1984 predict...                    \n",
       "1  Maine legislature candidate Leslie Gibson insu...                    \n",
       "2  A 17-year-old girl named Alyssa Carson is bein...                    \n",
       "3  In 1988 author Roald Dahl penned an open lette...                    \n",
       "4  When it comes to fighting terrorism, \"Another ...  Hillary Clinton   \n",
       "\n",
       "         date  label                            related_articles  id  \\\n",
       "0  2017-07-17      0            [122094, 122580, 130685, 134765]   0   \n",
       "1  2018-03-17      2                    [106868, 127320, 128060]   1   \n",
       "2  2018-07-18      1                    [132130, 132132, 149722]   4   \n",
       "3  2019-02-04      2                    [123254, 123418, 127464]   5   \n",
       "4  2016-03-22      2  [41099, 89899, 72543, 82644, 95344, 88361]   6   \n",
       "\n",
       "                                      clean articles  \\\n",
       "0     line georg orwel novel predict power smartphon   \n",
       "1  main legislatur candid lesli gibson insult par...   \n",
       "2  year old girl name alyssa carson train nasa be...   \n",
       "3  author roald dahl pen open letter urg parent c...   \n",
       "4  come fight terror anoth thing know work base l...   \n",
       "\n",
       "                                        clean claims  \n",
       "0     line georg orwel novel predict power smartphon  \n",
       "1  main legislatur candid lesli gibson insult par...  \n",
       "2  year old girl name alyssa carson train nasa be...  \n",
       "3  author roald dahl pen open letter urg parent c...  \n",
       "4  come fight terror anoth thing know work base l...  "
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Adding our cleaned claims and cleaned articles to the orignal dataframe\n",
    "claims['clean articles'] = clean_articles\n",
    "claims['clean claims'] = clean_claims\n",
    "\n",
    "claims.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Vectorizing claims, articles, claimants"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [],
   "source": [
    "y = claims['label']\n",
    "X_train, X_test, y_train, y_test = train_test_split(claims[['clean claims', 'clean articles', 'claimant']], y, \n",
    "                                                    random_state=0, test_size=0.3, stratify = y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "#Using the Tfidf Vectorizer on the claims, articles, claimants\n",
    "\n",
    "vectorizer = TfidfVectorizer(use_idf=True, smooth_idf=True, norm='l2')\n",
    "vectorizer1 = TfidfVectorizer(use_idf=True, smooth_idf=True, norm='l2')\n",
    "vectorizer2 = TfidfVectorizer(use_idf=True, smooth_idf=True, norm='l2')\n",
    "\n",
    "vectorizer.fit(X_train['clean claims'])#.values.astype('U'))\n",
    "vectorizer1.fit(X_train['clean articles'])#.values.astype('U'))\n",
    "vectorizer2.fit(X_train['claimant'])#.values.astype('U'))\n",
    "\n",
    "X1 = vectorizer.transform(X_train['clean claims'])#.values.astype('U'))\n",
    "X2 = vectorizer1.transform(X_train['clean articles'])#.values.astype('U'))\n",
    "X3 = vectorizer2.transform(X_train['claimant'])#.values.astype('U'))\n",
    "X_model_train = hstack((X1, X2, X3))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [],
   "source": [
    "X1_test = vectorizer.transform(X_test['clean claims'])\n",
    "X2_test = vectorizer1.transform(X_test['clean articles'])\n",
    "X3_test = vectorizer2.transform(X_test['claimant'])\n",
    "X_model_test = hstack((X1_test, X2_test, X3_test))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Model Implementation"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### XGBoost"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\siddh\\anaconda3\\lib\\site-packages\\xgboost\\sklearn.py:1224: UserWarning: The use of label encoder in XGBClassifier is deprecated and will be removed in a future release. To remove this warning, do the following: 1) Pass option use_label_encoder=False when constructing XGBClassifier object; and 2) Encode your labels (y) as integers starting with 0, i.e. 0, 1, 2, ..., [num_class - 1].\n",
      "  warnings.warn(label_encoder_deprecation_msg, UserWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:50:09] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.5.1/src/learner.cc:1115: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'multi:softprob' was changed from 'merror' to 'mlogloss'. Explicitly set eval_metric if you'd like to restore the old behavior.\n",
      "Accuracy: 0.6087422326976645\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.63      0.74      0.68      2223\n",
      "           1       0.59      0.61      0.60      1935\n",
      "           2       0.38      0.02      0.04       509\n",
      "\n",
      "    accuracy                           0.61      4667\n",
      "   macro avg       0.53      0.46      0.44      4667\n",
      "weighted avg       0.58      0.61      0.58      4667\n",
      "\n"
     ]
    }
   ],
   "source": [
    "xgb = XGBClassifier(random_state = 0)\n",
    "xgb.fit(X_model_train, y_train)\n",
    "predictions = xgb.predict(X_model_test)\n",
    "ACCU = accuracy_score(y_test, predictions)\n",
    "print('Accuracy:', ACCU)\n",
    "print(classification_report(y_test,predictions))"
   ]
  }
 ],
 "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": "3.8.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}