spam-classifier.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Load dataset\n",
    "\n",
    "SMS Spam Collection Dataset: <https://www.kaggle.com/datasets/uciml/sms-spam-collection-dataset>\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(5572, 5)"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "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>v1</th>\n",
       "      <th>v2</th>\n",
       "      <th>Unnamed: 2</th>\n",
       "      <th>Unnamed: 3</th>\n",
       "      <th>Unnamed: 4</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>ham</td>\n",
       "      <td>Go until jurong point, crazy.. Available only ...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>ham</td>\n",
       "      <td>Ok lar... Joking wif u oni...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>spam</td>\n",
       "      <td>Free entry in 2 a wkly comp to win FA Cup fina...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>ham</td>\n",
       "      <td>U dun say so early hor... U c already then say...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>ham</td>\n",
       "      <td>Nah I don't think he goes to usf, he lives aro...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     v1                                                 v2 Unnamed: 2  \\\n",
       "0   ham  Go until jurong point, crazy.. Available only ...        NaN   \n",
       "1   ham                      Ok lar... Joking wif u oni...        NaN   \n",
       "2  spam  Free entry in 2 a wkly comp to win FA Cup fina...        NaN   \n",
       "3   ham  U dun say so early hor... U c already then say...        NaN   \n",
       "4   ham  Nah I don't think he goes to usf, he lives aro...        NaN   \n",
       "\n",
       "  Unnamed: 3 Unnamed: 4  \n",
       "0        NaN        NaN  \n",
       "1        NaN        NaN  \n",
       "2        NaN        NaN  \n",
       "3        NaN        NaN  \n",
       "4        NaN        NaN  "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 5572 entries, 0 to 5571\n",
      "Data columns (total 5 columns):\n",
      " #   Column      Non-Null Count  Dtype \n",
      "---  ------      --------------  ----- \n",
      " 0   v1          5572 non-null   object\n",
      " 1   v2          5572 non-null   object\n",
      " 2   Unnamed: 2  50 non-null     object\n",
      " 3   Unnamed: 3  12 non-null     object\n",
      " 4   Unnamed: 4  6 non-null      object\n",
      "dtypes: object(5)\n",
      "memory usage: 217.8+ KB\n"
     ]
    }
   ],
   "source": [
    "df = pd.read_csv('./data/spam.csv', encoding='latin-1')\n",
    "\n",
    "display(df.shape)\n",
    "display(df.head())\n",
    "df.info()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Clean data\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Transform columns\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(5572, 2)"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "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>target</th>\n",
       "      <th>text</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>ham</td>\n",
       "      <td>Go until jurong point, crazy.. Available only ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>ham</td>\n",
       "      <td>Ok lar... Joking wif u oni...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>spam</td>\n",
       "      <td>Free entry in 2 a wkly comp to win FA Cup fina...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>ham</td>\n",
       "      <td>U dun say so early hor... U c already then say...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>ham</td>\n",
       "      <td>Nah I don't think he goes to usf, he lives aro...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  target                                               text\n",
       "0    ham  Go until jurong point, crazy.. Available only ...\n",
       "1    ham                      Ok lar... Joking wif u oni...\n",
       "2   spam  Free entry in 2 a wkly comp to win FA Cup fina...\n",
       "3    ham  U dun say so early hor... U c already then say...\n",
       "4    ham  Nah I don't think he goes to usf, he lives aro..."
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "df = df.drop(columns=['Unnamed: 2', 'Unnamed: 3', 'Unnamed: 4'])\n",
    "df = df.rename(columns={'v1': 'target', 'v2': 'text'})\n",
    "\n",
    "display(df.shape)\n",
    "display(df.head())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Encode labels"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['ham', 'spam'], dtype=object)"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "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>target</th>\n",
       "      <th>text</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>Go until jurong point, crazy.. Available only ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>Ok lar... Joking wif u oni...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>Free entry in 2 a wkly comp to win FA Cup fina...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "      <td>U dun say so early hor... U c already then say...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>Nah I don't think he goes to usf, he lives aro...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   target                                               text\n",
       "0       0  Go until jurong point, crazy.. Available only ...\n",
       "1       0                      Ok lar... Joking wif u oni...\n",
       "2       1  Free entry in 2 a wkly comp to win FA Cup fina...\n",
       "3       0  U dun say so early hor... U c already then say...\n",
       "4       0  Nah I don't think he goes to usf, he lives aro..."
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.preprocessing import LabelEncoder\n",
    "\n",
    "encoder = LabelEncoder()\n",
    "encoder.fit(df['target'])\n",
    "display(encoder.classes_)\n",
    "\n",
    "df['target'] = encoder.transform(df['target'])\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Check for null values\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "target    0\n",
       "text      0\n",
       "dtype: int64"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "target    0\n",
       "text      0\n",
       "dtype: int64"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "display(df.isnull().sum())\n",
    "\n",
    "if df.isnull().to_numpy().any():\n",
    "    df = df.dropna()\n",
    "    display('Removed rows with missing values.')\n",
    "\n",
    "display(df.isnull().sum())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Remove duplicates\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "403"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "'Removed duplicated rows.'"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "0"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "display(df.duplicated().sum())\n",
    "\n",
    "if df.duplicated().any():\n",
    "    df = df.drop_duplicates()\n",
    "    display('Removed duplicated rows.')\n",
    "\n",
    "display(df.duplicated().sum())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Perform EDA (Exploratory Data Analysis)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "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>target</th>\n",
       "      <th>text</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>Go until jurong point, crazy.. Available only ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>Ok lar... Joking wif u oni...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>Free entry in 2 a wkly comp to win FA Cup fina...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "      <td>U dun say so early hor... U c already then say...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>Nah I don't think he goes to usf, he lives aro...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   target                                               text\n",
       "0       0  Go until jurong point, crazy.. Available only ...\n",
       "1       0                      Ok lar... Joking wif u oni...\n",
       "2       1  Free entry in 2 a wkly comp to win FA Cup fina...\n",
       "3       0  U dun say so early hor... U c already then say...\n",
       "4       0  Nah I don't think he goes to usf, he lives aro..."
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "\n",
    "plt.pie(\n",
    "    df['target'].value_counts(),\n",
    "    labels=encoder.classes_.tolist(),\n",
    "    autopct='%0.2f%%'\n",
    ")\n",
    "plt.title('Target Distribution')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### NLTK"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Install package"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Requirement already satisfied: nltk in /home/ushka/anaconda3/lib/python3.11/site-packages (3.8.1)\n",
      "Requirement already satisfied: click in /home/ushka/anaconda3/lib/python3.11/site-packages (from nltk) (8.1.7)\n",
      "Requirement already satisfied: joblib in /home/ushka/anaconda3/lib/python3.11/site-packages (from nltk) (1.2.0)\n",
      "Requirement already satisfied: regex>=2021.8.3 in /home/ushka/anaconda3/lib/python3.11/site-packages (from nltk) (2023.10.3)\n",
      "Requirement already satisfied: tqdm in /home/ushka/anaconda3/lib/python3.11/site-packages (from nltk) (4.65.0)\n",
      "Note: you may need to restart the kernel to use updated packages.\n"
     ]
    }
   ],
   "source": [
    "%pip install nltk"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Download *Punkt Sentence Tokenizer*\n",
    "\n",
    "The Punkt Sentence Tokenizer is used to segment text into sentences, enabling more detailed analysis of text structure."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[nltk_data] Downloading package punkt to /home/ushka/nltk_data...\n",
      "[nltk_data]   Package punkt is already up-to-date!\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "True"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import nltk\n",
    "\n",
    "nltk.download('punkt')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Feature engineering"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Add features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "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>target</th>\n",
       "      <th>text</th>\n",
       "      <th>characters_count</th>\n",
       "      <th>words_count</th>\n",
       "      <th>sentences_count</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>Go until jurong point, crazy.. Available only ...</td>\n",
       "      <td>111</td>\n",
       "      <td>24</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>Ok lar... Joking wif u oni...</td>\n",
       "      <td>29</td>\n",
       "      <td>8</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>Free entry in 2 a wkly comp to win FA Cup fina...</td>\n",
       "      <td>155</td>\n",
       "      <td>37</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "      <td>U dun say so early hor... U c already then say...</td>\n",
       "      <td>49</td>\n",
       "      <td>13</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>Nah I don't think he goes to usf, he lives aro...</td>\n",
       "      <td>61</td>\n",
       "      <td>15</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   target                                               text  \\\n",
       "0       0  Go until jurong point, crazy.. Available only ...   \n",
       "1       0                      Ok lar... Joking wif u oni...   \n",
       "2       1  Free entry in 2 a wkly comp to win FA Cup fina...   \n",
       "3       0  U dun say so early hor... U c already then say...   \n",
       "4       0  Nah I don't think he goes to usf, he lives aro...   \n",
       "\n",
       "   characters_count  words_count  sentences_count  \n",
       "0               111           24                2  \n",
       "1                29            8                2  \n",
       "2               155           37                2  \n",
       "3                49           13                1  \n",
       "4                61           15                1  "
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['characters_count'] = df['text'].apply(len)\n",
    "df['words_count'] = df['text'].apply(lambda x: len(nltk.word_tokenize(x)))\n",
    "df['sentences_count'] = df['text'].apply(lambda x: len(nltk.sent_tokenize(x)))\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_hams(df):\n",
    "    return df[df['target'] == 0]\n",
    "\n",
    "\n",
    "def get_spams(df):\n",
    "    return df[df['target'] == 1]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Hams & Spams"
   ]
  },
  {
   "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>characters_count</th>\n",
       "      <th>words_count</th>\n",
       "      <th>sentences_count</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>5169.000000</td>\n",
       "      <td>5169.000000</td>\n",
       "      <td>5169.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>78.977945</td>\n",
       "      <td>18.455794</td>\n",
       "      <td>1.965564</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>58.236293</td>\n",
       "      <td>13.324758</td>\n",
       "      <td>1.448541</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>2.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>36.000000</td>\n",
       "      <td>9.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>60.000000</td>\n",
       "      <td>15.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>117.000000</td>\n",
       "      <td>26.000000</td>\n",
       "      <td>2.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>910.000000</td>\n",
       "      <td>220.000000</td>\n",
       "      <td>38.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       characters_count  words_count  sentences_count\n",
       "count       5169.000000  5169.000000      5169.000000\n",
       "mean          78.977945    18.455794         1.965564\n",
       "std           58.236293    13.324758         1.448541\n",
       "min            2.000000     1.000000         1.000000\n",
       "25%           36.000000     9.000000         1.000000\n",
       "50%           60.000000    15.000000         1.000000\n",
       "75%          117.000000    26.000000         2.000000\n",
       "max          910.000000   220.000000        38.000000"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "display(df[['characters_count', 'words_count', 'sentences_count']].describe())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Hams"
   ]
  },
  {
   "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>characters_count</th>\n",
       "      <th>words_count</th>\n",
       "      <th>sentences_count</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>4516.000000</td>\n",
       "      <td>4516.000000</td>\n",
       "      <td>4516.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>70.459256</td>\n",
       "      <td>17.123782</td>\n",
       "      <td>1.820195</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>56.358207</td>\n",
       "      <td>13.493970</td>\n",
       "      <td>1.383657</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>2.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>34.000000</td>\n",
       "      <td>8.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>52.000000</td>\n",
       "      <td>13.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>90.000000</td>\n",
       "      <td>22.000000</td>\n",
       "      <td>2.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>910.000000</td>\n",
       "      <td>220.000000</td>\n",
       "      <td>38.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       characters_count  words_count  sentences_count\n",
       "count       4516.000000  4516.000000      4516.000000\n",
       "mean          70.459256    17.123782         1.820195\n",
       "std           56.358207    13.493970         1.383657\n",
       "min            2.000000     1.000000         1.000000\n",
       "25%           34.000000     8.000000         1.000000\n",
       "50%           52.000000    13.000000         1.000000\n",
       "75%           90.000000    22.000000         2.000000\n",
       "max          910.000000   220.000000        38.000000"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "display(\n",
    "    get_hams(df)[\n",
    "        [\n",
    "            'characters_count',\n",
    "            'words_count',\n",
    "            'sentences_count'\n",
    "        ]\n",
    "    ].describe()\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Spams"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "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>characters_count</th>\n",
       "      <th>words_count</th>\n",
       "      <th>sentences_count</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>653.000000</td>\n",
       "      <td>653.000000</td>\n",
       "      <td>653.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>137.891271</td>\n",
       "      <td>27.667688</td>\n",
       "      <td>2.970904</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>30.137753</td>\n",
       "      <td>7.008418</td>\n",
       "      <td>1.488425</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>13.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>132.000000</td>\n",
       "      <td>25.000000</td>\n",
       "      <td>2.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>149.000000</td>\n",
       "      <td>29.000000</td>\n",
       "      <td>3.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>157.000000</td>\n",
       "      <td>32.000000</td>\n",
       "      <td>4.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>224.000000</td>\n",
       "      <td>46.000000</td>\n",
       "      <td>9.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       characters_count  words_count  sentences_count\n",
       "count        653.000000   653.000000       653.000000\n",
       "mean         137.891271    27.667688         2.970904\n",
       "std           30.137753     7.008418         1.488425\n",
       "min           13.000000     2.000000         1.000000\n",
       "25%          132.000000    25.000000         2.000000\n",
       "50%          149.000000    29.000000         3.000000\n",
       "75%          157.000000    32.000000         4.000000\n",
       "max          224.000000    46.000000         9.000000"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "display(\n",
    "    get_spams(df)[\n",
    "        [\n",
    "            'characters_count',\n",
    "            'words_count',\n",
    "            'sentences_count'\n",
    "        ]\n",
    "    ].describe()\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Plots"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAkkAAAGxCAYAAAB2qSLdAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjguMCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy81sbWrAAAACXBIWXMAAA9hAAAPYQGoP6dpAABRlElEQVR4nO3de1xUdf4/8NfIZUCuBshFARGzVMgSSsH1VokLu6615pq2aqbsl1gvgH5LNNdLFmZK5FcBM0rJ9VKZbdtSQpasCdbi/cIaqwiIg8B4GbnN9fz+8OdsE2dgBoEZ4PV8PObxYD7nfc75zGGKl5/zOedIBEEQQEREREQGelm6A0RERETWiCGJiIiISARDEhEREZEIhiQiIiIiEQxJRERERCIYkoiIiIhEMCQRERERiWBIIiIiIhJha+kOdFU6nQ7Xrl2Di4sLJBKJpbtDREREJhAEAXfu3IGfnx969Wp5rIghqY2uXbsGf39/S3eDiIiI2qCiogL9+/dvsYYhqY1cXFwA3D3Irq6uFu4NERERmUKhUMDf31//d7wlDEltdO8Um6urK0MSERFRF2PKVBlO3CYiIiISwZBEREREJIIhiYiIiEgE5yQRERFZgFarhVqttnQ3uh07OzvY2Ni0y7YYkoiIiDqRIAioqqrCrVu3LN2Vbsvd3R0+Pj73fR9DhiQiIqJOdC8g9e3bF7179+YNiduRIAhoaGhAdXU1AMDX1/e+tseQRERE1Em0Wq0+IHl4eFi6O92So6MjAKC6uhp9+/a9r1NvnLhNRETUSe7NQerdu7eFe9K93Tu+9zvniyGJiIiok/EUW8dqr+PLkEREREQkgnOSiIiIrEBNTQ0UCkWn7c/V1RVeXl6dtr+uiCGJiIjIwmpqajBz5suQy5Wdtk8PDyl2784wOSi9+OKLuHXrFj7//HOD9sOHD2PChAm4efMm3N3d27+jFsSQREREZGEKhQJyuRJS6RI4Ovp3+P4aGysgl2+CQqHgaFILGJKIiIishKOjP5ycgjtlX8oOGLSSy+VYsGABjhw5ghs3biA4OBjLly/HjBkz9DXjx49HaGgobGxssHPnTtjb2+P111/HCy+8gAULFuDTTz9F3759sWXLFkRHR7d/J83AidtERG1UU1ODS5cutfiqqamxdDeJOk1TUxPCwsLw5Zdf4ty5c/jTn/6EWbNm4YcffjCo27lzJzw9PfHjjz9i4cKFePnllzFt2jRERkbixIkTmDRpEmbNmoWGhgYLfZK7OJJERNQGNTU1eHnmTCjl8hbrpB4eyNi9m6c0qFv48ssv4ezsbNCm1Wr1P/fr1w9Lly7Vv1+4cCG+/vprfPLJJxg5cqS+ffjw4XjttdcAAMnJyVi/fj08PT0RGxsLAPjLX/6CjIwMnDlzBqNGjerIj9QihiQiojZQKBRQyuVYIpXC///f4feXKhobsUku57wP6jYmTJiAjIwMg7YffvgBf/zjHwHcDUzr16/Hvn37UFlZCaVSCaVSCScnJ4N1HnnkEf3PNjY28PDwQGhoqL7N29sbAPSPF7EUhiQiovvg7+iI4F/8ATDQERM/iCzEyckJgwYNMmi7evWq/udNmzbhnXfeQVpaGkJDQ+Hk5ISEhASoVCqDdezs7AzeSyQSg7Z7N4PU6XTt/RHMwpBERERE7eLIkSOYMmWKfmRJp9OhpKQEQ4YMsXDP2oYhiYiIyEo0NlZ06f0MGjQI+/fvR0FBAfr06YPU1FRUVVUxJBEREVHbuLq6wsNDCrl8U6edofXwkMLV1bVdt7ly5UqUlpZi0qRJ6N27N/70pz/hmWeewe3bt9t1P52FIYmIiMjCvLy8sHt3hlU/lmTHjh2i7ePHj4cgCPr3v7wj9y8dPny4WduVK1eatf18m5bCkERERGQFvLy8eBWkleHNJImIiIhEMCQRERERiWBIIiIiIhLBkEREREQkgiGJiIiISARDEhEREZEIhiQiIiIiEQxJRERERCJ4M0kiIiIrUFNTY9V33AaA6upqrFy5El999RWuX7+OPn36YPjw4Vi9ejUiIiI6qKeWY/GQlJ6ejrfffhsymQzDhg1DWloaxowZY7Q+Pz8fSUlJOH/+PPz8/PDKK68gLi5OtHbv3r2YMWMGpkyZ0uw26ebul4iIqKPU1NTg5ZkzoZTLO22fUg8PZOzebVZQmjp1KtRqNXbu3ImBAwfi+vXrOHToEG7cuNGBPbUci4akffv2ISEhAenp6Rg9ejS2bduG6OhoXLhwAQEBAc3qS0tLERMTg9jYWOzatQtHjx5FfHw8vLy8MHXqVIPasrIyLF26VDT4mLtfIiKijqRQKKCUy7FEKoW/o2OH76+isRGb5HIoFAqTQ9KtW7fw/fff4/Dhwxg3bhwAIDAwEE888YS+RiKRID09HV988QUOHz4MHx8fbNiwAdOmTdPXvPrqqzhw4ACuXr0KHx8fvPDCC/jLX/4COzs7AMDq1avx+eefY9GiRVi9ejVu3LiBWbNmYcuWLdi0aRNSU1Oh0+mwePFirFixoh2PSnMWDUmpqamYN28e5s+fDwBIS0vDwYMHkZGRgZSUlGb1mZmZCAgIQFpaGgBgyJAhKCoqwsaNGw1CklarxQsvvIA1a9bgyJEjuHXr1n3tl4iIqDP4Ozoi2Mmpc3amVJpV7uzsDGdnZ3z++ecYNWoUpFKpaN3KlSuxfv16vPvuu/joo48wY8YMhISEYMiQIQAAFxcX7NixA35+fjh79ixiY2Ph4uKCV155Rb+NS5cu4auvvsLXX3+NS5cu4bnnnkNpaSkGDx6M/Px8FBQU4KWXXsJTTz2FUaNGtf0YtMJiE7dVKhWOHz+OqKgog/aoqCgUFBSIrlNYWNisftKkSSgqKoJarda3rV27Fl5eXpg3b1677BcAlEolFAqFwYuIiKinsLW1xY4dO7Bz5064u7tj9OjRWL58Oc6cOWNQN23aNMyfPx+DBw/G66+/jvDwcPzf//2ffvlrr72GyMhIDBgwAJMnT8aSJUvw8ccfG2xDp9Phgw8+wNChQzF58mRMmDABFy9eRFpaGh566CHMnTsXDz30EA4fPtyhn9liIam2thZarRbe3t4G7d7e3qiqqhJdp6qqSrReo9GgtrYWAHD06FFkZWVh+/bt7bZfAEhJSYGbm5v+5e/v3+pnJCIi6k6mTp2Ka9eu4YsvvsCkSZNw+PBhjBgxAjt27NDX/HICd0REBIqLi/XvP/30U/zqV7+Cj48PnJ2dsXLlSpSXlxusM2DAALi4uOjfe3t7Y+jQoejVq5dBW3V1dTt/QkMWvwWARCIxeC8IQrO21urvtd+5cwd//OMfsX37dnh6erbrfpOTk3H79m39q6KiosXtExERdUcODg6YOHEi/vKXv6CgoAAvvvgiVq1a1eI69/6+Hjt2DM8//zyio6Px5Zdf4uTJk1ixYgVUKpVB/b35ST9fX6xNp9O1wycyzmJzkjw9PWFjY9Ns9Ka6urrZKM89Pj4+ovW2trbw8PDA+fPnceXKFUyePFm//N4BtLW1xcWLF+Hv72/2fgFAKpUaPf9KRETUUw0dOtTgCvJjx45h9uzZBu8fe+wxAHfP9gQGBhpMuC4rK+u0vprLYiNJ9vb2CAsLQ15enkF7Xl4eIiMjRdeJiIhoVp+bm4vw8HDY2dnh4YcfxtmzZ3Hq1Cn963e/+x0mTJiAU6dOwd/fv037JSIi6unkcjmefPJJ7Nq1C2fOnEFpaSk++eQTbNiwAVOmTNHXffLJJ/jggw/w008/YdWqVfjxxx+xYMECAMCgQYNQXl6OvXv34tKlS9i8eTMOHDhgqY/UKote3ZaUlIRZs2YhPDwcEREReO+991BeXq6/71FycjIqKyuRnZ0NAIiLi8OWLVuQlJSE2NhYFBYWIisrC3v27AFwdwgwJCTEYB/u7u4AYNDe2n6JiIgsoaKx0Wr34+zsjJEjR+Kdd97BpUuXoFar4e/vj9jYWCxfvlxft2bNGuzduxfx8fHw8fHBX//6VwwdOhQAMGXKFCQmJmLBggVQKpX4zW9+g5UrV2L16tXt9dHalUVD0vTp0yGXy7F27VrIZDKEhIQgJycHgYGBAACZTGYwmSsoKAg5OTlITEzE1q1b4efnh82bNze7R9L97peIiKgzubq6QurhgU1yudmX5reV1MMDrq6uptdLpUhJSWn1Vjl+fn7Izc01unzDhg3YsGGDQVtCQoL+59WrVzcLTT+fGH5PR1/ZBljBHbfj4+MRHx8vukzsoIwbNw4nTpwwefti22htv0RERJ3Jy8sLGbt3W/1jSXoai4ckIiIiuhuUGFqsC0MSERERtYt7t+XpLix+nyQiIiIia8SQRERERCSCIYmIiKiTdbfTUtamvY4vQxIREVEnufdojYaGBgv3pHu7d3x/+SgTc3HiNhERUSexsbGBu7u7/sGsvXv3bvG5oWQeQRDQ0NCA6upquLu7w8bG5r62x5BERETUiXx8fACgw59g35O5u7vrj/P9YEgiIiLqRBKJBL6+vujbty/UarWlu9Pt2NnZ3fcI0j0MSURERBZgY2PTbn/MqWNw4jYRERGRCIYkIiIiIhEMSUREREQiGJKIiIiIRDAkEREREYlgSCIiIiISwZBEREREJIIhiYiIiEgEQxIRERGRCIYkIiIiIhEMSUREREQiGJKIiIiIRDAkEREREYlgSCIiIiISwZBEREREJIIhiYiIiEgEQxIRERGRCIYkIiIiIhEMSUREREQiGJKIiIiIRFg8JKWnpyMoKAgODg4ICwvDkSNHWqzPz89HWFgYHBwcMHDgQGRmZhos/+yzzxAeHg53d3c4OTnh0UcfxUcffWRQs3r1akgkEoOXj49Pu382IiIi6rosGpL27duHhIQErFixAidPnsSYMWMQHR2N8vJy0frS0lLExMRgzJgxOHnyJJYvX45FixZh//79+poHHngAK1asQGFhIc6cOYO5c+di7ty5OHjwoMG2hg0bBplMpn+dPXu2Qz8rERERdS22ltx5amoq5s2bh/nz5wMA0tLScPDgQWRkZCAlJaVZfWZmJgICApCWlgYAGDJkCIqKirBx40ZMnToVADB+/HiDdRYvXoydO3fi+++/x6RJk/Tttra2HD0iIiIioyw2kqRSqXD8+HFERUUZtEdFRaGgoEB0ncLCwmb1kyZNQlFREdRqdbN6QRBw6NAhXLx4EWPHjjVYVlJSAj8/PwQFBeH555/H5cuX7/MTERERUXdisZGk2tpaaLVaeHt7G7R7e3ujqqpKdJ2qqirReo1Gg9raWvj6+gIAbt++jX79+kGpVMLGxgbp6emYOHGifp2RI0ciOzsbgwcPxvXr17Fu3TpERkbi/Pnz8PDwEN23UqmEUqnUv1coFG363ERERNQ1WPR0GwBIJBKD94IgNGtrrf6X7S4uLjh16hTq6upw6NAhJCUlYeDAgfpTcdHR0fra0NBQREREIDg4GDt37kRSUpLoflNSUrBmzRqzPhsRERF1XRYLSZ6enrCxsWk2alRdXd1stOgeHx8f0XpbW1uDEaBevXph0KBBAIBHH30UxcXFSElJaTZf6R4nJyeEhoaipKTEaH+Tk5MNApRCoYC/v3+Ln5GIiIi6LovNSbK3t0dYWBjy8vIM2vPy8hAZGSm6TkRERLP63NxchIeHw87Ozui+BEEwOFX2S0qlEsXFxfrTdWKkUilcXV0NXkRERNR9WfR0W1JSEmbNmoXw8HBERETgvffeQ3l5OeLi4gDcHb2prKxEdnY2ACAuLg5btmxBUlISYmNjUVhYiKysLOzZs0e/zZSUFISHhyM4OBgqlQo5OTnIzs5GRkaGvmbp0qWYPHkyAgICUF1djXXr1kGhUGDOnDmdewCIiIjIalk0JE2fPh1yuRxr166FTCZDSEgIcnJyEBgYCACQyWQG90wKCgpCTk4OEhMTsXXrVvj5+WHz5s36y/8BoL6+HvHx8bh69SocHR3x8MMPY9euXZg+fbq+5urVq5gxYwZqa2vh5eWFUaNG4dixY/r9EhEREUmEezOfySwKhQJubm64ffs2T70R9UCXLl1CwrRpSHN3R7CTk3hNfT0Sbt1C2iefIDg4uJN7SERizPn7bfHHkhARERFZI4YkIiIiIhEMSUREREQiGJKIiIiIRDAkEREREYlgSCIiIiISwZBEREREJIIhiYiIiEgEQxIRERGRCIYkIiIiIhEMSUREREQiGJKIiIiIRDAkEREREYlgSCIiIiISwZBEREREJIIhiYiIiEgEQxIRERGRCIYkIiIiIhEMSUREREQiGJKIiIiIRDAkEREREYlgSCIiIiISwZBEREREJIIhiYiIiEgEQxIRERGRCIYkIiIiIhEMSUREREQiGJKIiIiIRDAkEREREYlgSCIiIiISYfGQlJ6ejqCgIDg4OCAsLAxHjhxpsT4/Px9hYWFwcHDAwIEDkZmZabD8s88+Q3h4ONzd3eHk5IRHH30UH3300X3vl4iIiHoWi4akffv2ISEhAStWrMDJkycxZswYREdHo7y8XLS+tLQUMTExGDNmDE6ePInly5dj0aJF2L9/v77mgQcewIoVK1BYWIgzZ85g7ty5mDt3Lg4ePNjm/RIREVHPIxEEQbDUzkeOHIkRI0YgIyND3zZkyBA888wzSElJaVb/6quv4osvvkBxcbG+LS4uDqdPn0ZhYaHR/YwYMQK/+c1v8Prrr7dpv2IUCgXc3Nxw+/ZtuLq6mrQOEXUfly5dQsK0aUhzd0ewk5N4TX09Em7dQtonnyA4OLiTe0hEYsz5+22xkSSVSoXjx48jKirKoD0qKgoFBQWi6xQWFjarnzRpEoqKiqBWq5vVC4KAQ4cO4eLFixg7dmyb90tEREQ9j62ldlxbWwutVgtvb2+Ddm9vb1RVVYmuU1VVJVqv0WhQW1sLX19fAMDt27fRr18/KJVK2NjYID09HRMnTmzzfgFAqVRCqVTq3ysUCtM/LBEREXU5FgtJ90gkEoP3giA0a2ut/pftLi4uOHXqFOrq6nDo0CEkJSVh4MCBGD9+fJv3m5KSgjVr1rT6eYiIiKh7sFhI8vT0hI2NTbPRm+rq6majPPf4+PiI1tva2sLDw0Pf1qtXLwwaNAgA8Oijj6K4uBgpKSkYP358m/YLAMnJyUhKStK/VygU8Pf3N+3DEhERUZdjsTlJ9vb2CAsLQ15enkF7Xl4eIiMjRdeJiIhoVp+bm4vw8HDY2dkZ3ZcgCPpTZW3ZLwBIpVK4uroavIiIiKj7sujptqSkJMyaNQvh4eGIiIjAe++9h/LycsTFxQG4O3pTWVmJ7OxsAHevZNuyZQuSkpIQGxuLwsJCZGVlYc+ePfptpqSkIDw8HMHBwVCpVMjJyUF2drbBlWyt7ZeIiIjIoiFp+vTpkMvlWLt2LWQyGUJCQpCTk4PAwEAAgEwmM7h3UVBQEHJycpCYmIitW7fCz88PmzdvxtSpU/U19fX1iI+Px9WrV+Ho6IiHH34Yu3btwvTp003eLxEREZFF75PUlfE+SUQ9G++TRNQ1dYn7JBERERFZM4YkIiIiIhEMSUREREQiGJKIiIiIRDAkEREREYlgSCIiIiISwZBEREREJIIhiYiIiEgEQxIRERGRCIYkIiIiIhEMSUREREQiGJKIiIiIRDAkEREREYlgSCIiIiISwZBEREREJIIhiYiIiEgEQxIRERGRCIYkIiIiIhEMSUREREQiGJKIiIiIRDAkEREREYlgSCIiIiISwZBEREREJIIhiYiIiEiE2SGptLS0I/pBREREZFXMDkmDBg3ChAkTsGvXLjQ1NXVEn4iIiIgszuyQdPr0aTz22GNYsmQJfHx88D//8z/48ccfO6JvRERERBZjdkgKCQlBamoqKisr8eGHH6Kqqgq/+tWvMGzYMKSmpqKmpqYj+klERETUqdo8cdvW1hbPPvssPv74Y7z11lu4dOkSli5div79+2P27NmQyWTt2U8iIiKiTtXmkFRUVIT4+Hj4+voiNTUVS5cuxaVLl/Dtt9+isrISU6ZMac9+EhEREXUqW3NXSE1NxYcffoiLFy8iJiYG2dnZiImJQa9ed/NWUFAQtm3bhocffrjdO0tE1FlqamqgUCiMLi8rK4NGo+nEHhFRZzM7JGVkZOCll17C3Llz4ePjI1oTEBCArKwsk7aXnp6Ot99+GzKZDMOGDUNaWhrGjBljtD4/Px9JSUk4f/48/Pz88MorryAuLk6/fPv27cjOzsa5c+cAAGFhYXjzzTfxxBNP6GtWr16NNWvWGGzX29sbVVVVJvWZiLq3mpoavDxzJpRyudGaeqUS1ysqoHRz68SeEVFnMjsklZSUtFpjb2+POXPmtFq3b98+JCQkID09HaNHj8a2bdsQHR2NCxcuICAgoFl9aWkpYmJiEBsbi127duHo0aOIj4+Hl5cXpk6dCgA4fPgwZsyYgcjISDg4OGDDhg2IiorC+fPn0a9fP/22hg0bhm+++Ub/3sbGxpSPT0Q9gEKhgFIuxxKpFP6OjqI1x27exBsaDbQcTSLqtswOSR9++CGcnZ0xbdo0g/ZPPvkEDQ0NJoWje1JTUzFv3jzMnz8fAJCWloaDBw8iIyMDKSkpzeozMzMREBCAtLQ0AMCQIUNQVFSEjRs36kPSX//6V4N1tm/fjk8//RSHDh3C7Nmz9e22trZGR8KIiADA39ERwU5OosvKGhs7uTdE1NnMnri9fv16eHp6Nmvv27cv3nzzTZO3o1KpcPz4cURFRRm0R0VFoaCgQHSdwsLCZvWTJk1CUVER1Gq16DoNDQ1Qq9V44IEHDNpLSkrg5+eHoKAgPP/887h8+XKL/VUqlVAoFAYvIiIi6r7MDkllZWUICgpq1h4YGIjy8nKTt1NbWwutVgtvb2+D9pbmBlVVVYnWazQa1NbWiq6zbNky9OvXD08//bS+beTIkcjOzsbBgwexfft2VFVVITIyEvIW5h+kpKTAzc1N//L39zf1oxIREVEXZHZI6tu3L86cOdOs/fTp0/Dw8DC7AxKJxOC9IAjN2lqrF2sHgA0bNmDPnj347LPP4ODgoG+Pjo7G1KlTERoaiqeffhr/+Mc/AAA7d+40ut/k5GTcvn1b/6qoqGj9wxEREVGXZfacpOeffx6LFi2Ci4sLxo4dC+DuFWeLFy/G888/b/J2PD09YWNj02zUqLq6utlo0T0+Pj6i9ba2ts0C2saNG/Hmm2/im2++wSOPPNJiX5ycnBAaGtripHSpVAqpVNridoiIiKj7MHskad26dRg5ciSeeuopODo6wtHREVFRUXjyySfNmpNkb2+PsLAw5OXlGbTn5eUhMjJSdJ2IiIhm9bm5uQgPD4ednZ2+7e2338brr7+Or7/+GuHh4a32RalUori4GL6+vib3n4iIiLo3s0eS7O3tsW/fPrz++us4ffo0HB0dERoaisDAQLN3npSUhFmzZiE8PBwRERF47733UF5err/vUXJyMiorK5GdnQ0AiIuLw5YtW5CUlITY2FgUFhYiKysLe/bs0W9zw4YNWLlyJXbv3o0BAwboR56cnZ3h7OwMAFi6dCkmT56MgIAAVFdXY926dVAoFGZdmUdERETdm9kh6Z7Bgwdj8ODB97Xz6dOnQy6XY+3atZDJZAgJCUFOTo4+cMlkMoPJ4EFBQcjJyUFiYiK2bt0KPz8/bN68WX/5P3D35pQqlQrPPfecwb5WrVqF1atXAwCuXr2KGTNmoLa2Fl5eXhg1ahSOHTvWpqBHRERE3ZPZIUmr1WLHjh04dOgQqqurodPpDJZ/++23Zm0vPj4e8fHxost27NjRrG3cuHE4ceKE0e1duXKl1X3u3bvX1O4RERFRD2V2SFq8eDF27NiB3/zmNwgJCWnxSjQiIiKirsrskLR37158/PHHiImJ6Yj+EBEREVkFs69us7e3x6BBgzqiL0RERERWw+yQtGTJErz77rv6mzgSERERdUdmn277/vvv8d133+Grr77CsGHDDO5PBACfffZZu3WOiIiIyFLMDknu7u549tlnO6IvRERERFbD7JD04YcfdkQ/iIiIiKyK2XOSAECj0eCbb77Btm3bcOfOHQDAtWvXUFdX166dIyIiIrIUs0eSysrK8Otf/xrl5eVQKpWYOHEiXFxcsGHDBjQ1NSEzM7Mj+klERETUqcweSVq8eDHCw8Nx8+ZNODo66tufffZZHDp0qF07R0RERGQpbbq67ejRo7C3tzdoDwwMRGVlZbt1jIiIiMiSzB5J0ul00Gq1zdqvXr0KFxeXdukUERERkaWZHZImTpyItLQ0/XuJRIK6ujqsWrWKjyohIiKibsPs023vvPMOJkyYgKFDh6KpqQkzZ85ESUkJPD09sWfPno7oIxEREVGnMzsk+fn54dSpU9izZw9OnDgBnU6HefPm4YUXXjCYyE1ERETUlZkdkgDA0dERL730El566aX27g8RERGRVTA7JGVnZ7e4fPbs2W3uDBEREZG1MDskLV682OC9Wq1GQ0MD7O3t0bt3b4YkIiIi6hbMvrrt5s2bBq+6ujpcvHgRv/rVrzhxm4iIiLqNNj277ZcefPBBrF+/vtkoExEREVFX1S4hCQBsbGxw7dq19tocERERkUWZPSfpiy++MHgvCAJkMhm2bNmC0aNHt1vHiIiIiCzJ7JD0zDPPGLyXSCTw8vLCk08+iU2bNrVXv4iIiIgsyuyQpNPpOqIfRERERFal3eYkEREREXUnZo8kJSUlmVybmppq7uaJiIiIrILZIenkyZM4ceIENBoNHnroIQDATz/9BBsbG4wYMUJfJ5FI2q+XRERERJ3M7JA0efJkuLi4YOfOnejTpw+AuzeYnDt3LsaMGYMlS5a0eyeJiIiIOpvZc5I2bdqElJQUfUACgD59+mDdunW8uo2IiIi6DbNDkkKhwPXr15u1V1dX486dO+3SKSIiIiJLMzskPfvss5g7dy4+/fRTXL16FVevXsWnn36KefPm4fe//73ZHUhPT0dQUBAcHBwQFhaGI0eOtFifn5+PsLAwODg4YODAgcjMzDRYvn37dowZMwZ9+vRBnz598PTTT+PHH3+87/0SERFRz2J2SMrMzMRvfvMb/PGPf0RgYCACAwPxwgsvIDo6Gunp6WZta9++fUhISMCKFStw8uRJjBkzBtHR0SgvLxetLy0tRUxMDMaMGYOTJ09i+fLlWLRoEfbv36+vOXz4MGbMmIHvvvsOhYWFCAgIQFRUFCorK9u8XyIiIup5JIIgCG1Zsb6+HpcuXYIgCBg0aBCcnJzM3sbIkSMxYsQIZGRk6NuGDBmCZ555BikpKc3qX331VXzxxRcoLi7Wt8XFxeH06dMoLCwU3YdWq0WfPn2wZcsWzJ49u037FaNQKODm5obbt2/D1dXVpHWIqGu4dOkSEqZNQ5q7O4KN/L/t29paLDh9GvuGD0eop6f4durrkXDrFtI++QTBwcEd2WUiMpE5f7/bfDNJmUwGmUyGwYMHw8nJCeZmLZVKhePHjyMqKsqgPSoqCgUFBaLrFBYWNqufNGkSioqKoFarRddpaGiAWq3GAw880Ob9AoBSqYRCoTB4ERERUfdldkiSy+V46qmnMHjwYMTExEAmkwEA5s+fb9bl/7W1tdBqtfD29jZo9/b2RlVVleg6VVVVovUajQa1tbWi6yxbtgz9+vXD008/3eb9AkBKSgrc3Nz0L39//1Y/IxEREXVdZoekxMRE2NnZoby8HL1799a3T58+HV9//bXZHfjlTScFQWjxRpRi9WLtALBhwwbs2bMHn332GRwcHO5rv8nJybh9+7b+VVFRYbSWiIiIuj6zbyaZm5uLgwcPon///gbtDz74IMrKykzejqenJ2xsbJqN3lRXVzcb5bnHx8dHtN7W1hYeHh4G7Rs3bsSbb76Jb775Bo888sh97RcApFIppFKpSZ+NiIiIuj6zR5Lq6+sNRpDuqa2tNStE2NvbIywsDHl5eQbteXl5iIyMFF0nIiKiWX1ubi7Cw8NhZ2enb3v77bfx+uuv4+uvv0Z4ePh975eIiIh6HrND0tixY5Gdna1/L5FIoNPp8Pbbb2PChAlmbSspKQnvv/8+PvjgAxQXFyMxMRHl5eWIi4sDcPcU170r0oC7V7KVlZUhKSkJxcXF+OCDD5CVlYWlS5fqazZs2IDXXnsNH3zwAQYMGICqqipUVVWhrq7O5P0SERERmX267e2338b48eNRVFQElUqFV155BefPn8eNGzdw9OhRs7Y1ffp0yOVyrF27FjKZDCEhIcjJyUFgYCCAu1fQ/fzeRUFBQcjJyUFiYiK2bt0KPz8/bN68GVOnTtXXpKenQ6VS4bnnnjPY16pVq7B69WqT9ktERETUpvskVVVVISMjA8ePH4dOp8OIESPw5z//Gb6+vh3RR6vE+yQRdV+8TxJR92XO32+zRpLUajWioqKwbds2rFmz5r46SURERGTNzJqTZGdnh3PnzrV4qTwRERFRd2D2nKTZs2cjKysL69ev74j+EBF1uJqamhbvml9WVgaNRtOJPSIia2R2SFKpVHj//feRl5eH8PDwZs9sS01NbbfOERG1t5qaGrw8cyaUcrnRmnqlEtcrKqB0c+vEnhGRtTEpJJ05cwYhISHo1asXzp07hxEjRgAAfvrpJ4M6noYjImunUCiglMuxRCqFv6OjaM2xmzfxhkYDLUeTiHo0k0LSY489BplMhr59+6KsrAz/+te/mt3hmoioK/F3dDR65VpZY2Mn94aIrJFJE7fd3d1RWloKALhy5Qp0Ol2HdoqIiIjI0kwaSZo6dSrGjRsHX19fSCQShIeHw8bGRrT28uXL7dpBIiIiIkswKSS99957+P3vf4///Oc/WLRoEWJjY+Hi4tLRfSMiIiKyGJOvbvv1r38NADh+/DgWL17MkERERETdmtm3APjwww87oh9EREREVsWsO24TERER9RRmjyRRz9TaHYoBwNXVFV5eXp3UIyIioo7FkEStqqmpwcyZL0MuV7ZY5+Ehxe7dGQxKRETULTAkUasUCgXkciWk0iVwdPQXrWlsrIBcvgkKhYIhiYiIugWGJDKZo6M/nJyCjS5XtjzQRERE1KVw4jYRERGRCIYkIiIiIhE83UbtRq1WoqysrMUaXgFHRERdBUMStQuVSo6ysstYuHA9pFKp0TpeAUdERF0FQxK1C622DhqNPeztE+HuPli0hlfAERFRV8KQRO3KwaE/r4AjIqJugRO3iYiIiEQwJBERERGJYEgiIiIiEsGQRERERCSCIYmIiIhIBEMSERERkQiGJCIiIiIRvE8SdSo+uoSIiLoKi48kpaenIygoCA4ODggLC8ORI0darM/Pz0dYWBgcHBwwcOBAZGZmGiw/f/48pk6digEDBkAikSAtLa3ZNlavXg2JRGLw8vHxac+PRSJ+/uiSadMSjL5mznwZNTU1lu4uERH1cBYdSdq3bx8SEhKQnp6O0aNHY9u2bYiOjsaFCxcQEBDQrL60tBQxMTGIjY3Frl27cPToUcTHx8PLywtTp04FADQ0NGDgwIGYNm0aEhMTje572LBh+Oabb/TvbWxs2v8DkgE+uoSIiLoSi4ak1NRUzJs3D/PnzwcApKWl4eDBg8jIyEBKSkqz+szMTAQEBOhHh4YMGYKioiJs3LhRH5Ief/xxPP744wCAZcuWGd23ra0tR48shI8uISKirsBip9tUKhWOHz+OqKgog/aoqCgUFBSIrlNYWNisftKkSSgqKoJarTZr/yUlJfDz80NQUBCef/55XL582bwPQERERN2axUJSbW0ttFotvL29Ddq9vb1RVVUluk5VVZVovUajQW1trcn7HjlyJLKzs3Hw4EFs374dVVVViIyMhFwuN7qOUqmEQqEweBEREVH3ZfGJ2xKJxOC9IAjN2lqrF2tvSXR0NKZOnYrQ0FA8/fTT+Mc//gEA2Llzp9F1UlJS4Obmpn/5+/ubvD8iIiLqeiwWkjw9PWFjY9Ns1Ki6urrZaNE9Pj4+ovW2trbw8PBoc1+cnJwQGhqKkpISozXJycm4ffu2/lVRUdHm/REREZH1s9jEbXt7e4SFhSEvLw/PPvusvj0vLw9TpkwRXSciIgJ///vfDdpyc3MRHh4OOzu7NvdFqVSiuLgYY8aMMVojlUohlUrbvA8iun81NTWtnurmfbaIqL1Y9Oq2pKQkzJo1C+Hh4YiIiMB7772H8vJyxMXFAbg7elNZWYns7GwAQFxcHLZs2YKkpCTExsaisLAQWVlZ2LNnj36bKpUKFy5c0P9cWVmJU6dOwdnZGYMGDQIALF26FJMnT0ZAQACqq6uxbt06KBQKzJkzp5OPABGZqqamBi/PnAllC3MHAUDq4YGM3bsZlIjovlk0JE2fPh1yuRxr166FTCZDSEgIcnJyEBgYCACQyWQoLy/X1wcFBSEnJweJiYnYunUr/Pz8sHnzZv3l/wBw7do1PPbYY/r3GzduxMaNGzFu3DgcPnwYAHD16lXMmDEDtbW18PLywqhRo3Ds2DH9fonI+igUCijlciyRSuHv6ChaU9HYiE1yOe+zRUTtwuKPJYmPj0d8fLzosh07djRrGzduHE6cOGF0ewMGDNBP5jZm7969ZvWRiKyHv6Mjgp2cjBfwRltE1E4sfnUbERERkTViSCIiIiISwZBEREREJIIhiYiIiEgEQxIRERGRCIYkIiIiIhEMSUREREQiGJKIiIiIRDAkEREREYlgSCIiIiISYfHHkhD9klqtRFlZWYs1fNI7ERF1NIYksioqlRxlZZexcOF6SKVSo3UeHlLs3p3BoERERB2GIYmsilZbB43GHvb2iXB3Hyxa09hYAbl8E5/0TkREHYohiaySg0N/ODkFG13OB72TMUq1usXTtWVlZdBoNJ3YIyLqqhiSiKjbkKtUuFxWhvULFxo9XVuvVOJ6RQWUbm6d3Dsi6moYkoio26jTamGv0SDR3h6D3d1Fa47dvIk3NBpoOZpERK1gSCKibqe/gwOCnZxEl5U1NnZyb4ioq+J9koiIiIhEMCQRERERiWBIIiIiIhLBkEREREQkgiGJiIiISARDEhEREZEIhiQiIiIiEQxJRERERCIYkoiIiIhEMCQRERERiWBIIiIiIhLBkEREREQkgiGJiIiISARDEhEREZEIi4ek9PR0BAUFwcHBAWFhYThy5EiL9fn5+QgLC4ODgwMGDhyIzMxMg+Xnz5/H1KlTMWDAAEgkEqSlpbXLfomIiKhnsWhI2rdvHxISErBixQqcPHkSY8aMQXR0NMrLy0XrS0tLERMTgzFjxuDkyZNYvnw5Fi1ahP379+trGhoaMHDgQKxfvx4+Pj7tsl+yPmq1EmVlZbh06ZLRV01NjaW7SUREXZitJXeempqKefPmYf78+QCAtLQ0HDx4EBkZGUhJSWlWn5mZiYCAAP3o0JAhQ1BUVISNGzdi6tSpAIDHH38cjz/+OABg2bJl7bJfsi4qlRxlZZexcOF6SKVSo3UeHlLs3p0BLy+vTuwdERF1FxYLSSqVCsePH28WZKKiolBQUCC6TmFhIaKiogzaJk2ahKysLKjVatjZ2XXIfgFAqVRCqVTq3ysUilb3RR1Dq62DRmMPe/tEuLsPFq1pbKyAXL4JCoWCIYmIiNrEYiGptrYWWq0W3t7eBu3e3t6oqqoSXaeqqkq0XqPRoLa2Fr6+vh2yXwBISUnBmjVrWt0+dR4Hh/5wcgo2uvxnmZaIiMhsFp+4LZFIDN4LgtCsrbV6sfb23m9ycjJu376tf1VUVJi1PyIiIupaLDaS5OnpCRsbm2ajN9XV1c1Gee7x8fERrbe1tYWHh0eH7RcApFJpi/NfiIiIqHux2EiSvb09wsLCkJeXZ9Cel5eHyMhI0XUiIiKa1efm5iI8PNyk+Uht3S8RERH1PBa9ui0pKQmzZs1CeHg4IiIi8N5776G8vBxxcXEA7p7iqqysRHZ2NgAgLi4OW7ZsQVJSEmJjY1FYWIisrCzs2bNHv02VSoULFy7of66srMSpU6fg7OyMQYMGmbRfIiIiIouGpOnTp0Mul2Pt2rWQyWQICQlBTk4OAgMDAQAymczg3kVBQUHIyclBYmIitm7dCj8/P2zevFl/+T8AXLt2DY899pj+/caNG7Fx40aMGzcOhw8fNmm/RERERBYNSQAQHx+P+Ph40WU7duxo1jZu3DicOHHC6PYGDBign8zd1v0SERERWfzqNiIiIiJrxJBEREREJIIhiYiIiEiExeckkeXV1NS0+JiVsrIyaDSaTuwRERGR5TEk9XA1NTWYOfNlyOXGn+GhVNajouI63Ny61nM+1GolysrKWqxxdXXls92IiEgUQ1IPp1AoIJcrIZUugaOjv2jNzZvHoNG8AY1G28m9azuVSo6ysstYuHB9i3dK9/CQYvfuDAYlIiJqhiGJAACOjv5GHxbb2NjyaIw10mrroNHYw94+Ee7ug0VrGhsrIJdvgkKhYEgiIqJmGJKoW3Nw6G80/AGAsmudQSQiok7Eq9uIiIiIRHAkiYioAynVal5AQNRFMSQREXUQuUqFy2VlWL9wYYsXEEg9PJCxezeDEpGVYUgiIuogdVot7DUaJNrbY7C7u2hNRWMjNsnlvICAyAoxJBERdbD+Dg4IdnIyXsArCIisEiduExEREYlgSCIiIiISwdNt1KPx0SVERGQMQxL1WHx0CRERtYQhiXosPrqEiIhawpBEPV57PLqkpqYGCoWixRqetiMi6loYkoha0dq8Jblcjv/933W4c0docTs8bUdE1LUwJBG1wJR5S0plPSoqruOhh96Bi4v4iBRP2xERdT0MSUQtMGXe0s2bx6DRvAFbW9/7Pm1HRETWgyGJyAQtzVtqbGz5FgJERNQ18WaSRERERCIYkoiIiIhEMCQRERERieCcJCLqcLyPFBF1RQxJRNShampq8PLMmVDK5S3WST08kLF7N4MSEVkNhiQiK9IdR1wUCgWUcjmWSKXwd3QUralobMQmuZz3kSIiq8KQRGQlampqMHPmy5DLW76hUle9c7e/oyOCnZyMF/BGUkRkZSw+cTs9PR1BQUFwcHBAWFgYjhw50mJ9fn4+wsLC4ODggIEDByIzM7NZzf79+zF06FBIpVIMHToUBw4cMFi+evVqSCQSg5ePj0+7fi4icykUCsjlSkilS+Dunib6kkqXQC5XtjraRERE98+iIWnfvn1ISEjAihUrcPLkSYwZMwbR0dEoLy8XrS8tLUVMTAzGjBmDkydPYvny5Vi0aBH279+vryksLMT06dMxa9YsnD59GrNmzcIf/vAH/PDDDwbbGjZsGGQymf519uzZDv2sRKZydPSHk1Ow6MvR0d/S3SMi6jEserotNTUV8+bNw/z58wEAaWlpOHjwIDIyMpCSktKsPjMzEwEBAUhLSwMADBkyBEVFRdi4cSOmTp2q38bEiRORnJwMAEhOTkZ+fj7S0tKwZ88e/bZsbW05ekRkRZRqdYsPEi4rK4NGo+nEHhFRT2exkKRSqXD8+HEsW7bMoD0qKgoFBQWi6xQWFiIqKsqgbdKkScjKyoJarYadnR0KCwuRmJjYrOZesLqnpKQEfn5+kEqlGDlyJN58800MHDjQaH+VSiWUP5szwdMdZC61WskQYIRcpcLlsjKsX7jQ6IOE65VKXK+ogNLNrZN7R0Q9lcVCUm1tLbRaLby9vQ3avb29UVVVJbpOVVWVaL1Go0FtbS18fX2N1vx8myNHjkR2djYGDx6M69evY926dYiMjMT58+fh4eEhuu+UlBSsWbOmLR+VCCqVHGVll7Fw4XqjIUCprEdFxXW4ufW8Ccx1Wi3sNRok2ttjsLu7aM2xmzfxhkYDbQ8NkkTU+Sx+dZtEIjF4LwhCs7bW6n/Z3to2o6Oj9T+HhoYiIiICwcHB2LlzJ5KSkkT3m5ycbLBMoVDA35/zQ8g0Wm0dNBp72Nsnwt19sGjNzZvHoNG8AY1G28m9sx79HRyMXgFX1tjYyb0hop7OYiHJ09MTNjY2zUaNqqurm40E3ePj4yNab2trqx8BMlZjbJsA4OTkhNDQUJSUlBitkUqlRkcAiEzl4NAfTk7BossaG42fiiMios5nsavb7O3tERYWhry8PIP2vLw8REZGiq4TERHRrD43Nxfh4eGws7NrscbYNoG7842Ki4vh6+vblo9CRERE3ZBFbwGQlJSE999/Hx988AGKi4uRmJiI8vJyxMXFAbh7imv27Nn6+ri4OJSVlSEpKQnFxcX44IMPkJWVhaVLl+prFi9ejNzcXLz11lv497//jbfeegvffPMNEhIS9DVLly5Ffn4+SktL8cMPP+C5556DQqHAnDlzOu2zExERkXWz6Jyk6dOnQy6XY+3atZDJZAgJCUFOTg4CAwMBADKZzOCeSUFBQcjJyUFiYiK2bt0KPz8/bN68WX/5PwBERkZi7969eO2117By5UoEBwdj3759GDlypL7m6tWrmDFjBmpra+Hl5YVRo0bh2LFj+v0SERERWXzidnx8POLj40WX7dixo1nbuHHjcOLEiRa3+dxzz+G5554zunzv3r1m9ZGIiIh6Hos/loSIiIjIGjEkEREREYlgSCIiIiISwZBEREREJMLiE7eJyDytPQMOAFxdXeHl5XXf+6qpqWn1OYXttS8iImvDkETUhZjyDDgA8PCQYvfujPsKLzU1NXh55kwo5fIW66QeHsjYvZtBiYi6HYYkoi7ElGfANTZWoKrqTZw9e7bFe3+pVCrY29sbXV5WVob669fxqpMT/B0dRWsqGhuxSS6HQqHokSFJEAQ0NTWhvr5edHlTU5P++ZJE1PUwJBF1QS09A06lkuNm2UmsX7jQ6GiTUq3G5atXEdC3L2xtxf830KBS4YZMhr6PPWb0obN3N6Y0u/9dQWsB6E5dHVQqFX766Qoaba+J1vykUUKlUkGtVndkV4mogzAkEXUzWm0dHDUqJNrbY7C7u2jNkZoavK5Q4HdNOngbCUkXNGrsVjWiqbERMLId4G7gammOVFlZGTQajTkfweJUanWrAeiSqhE6nQBIfGBnJ/7cR4n2OgTh39BqtR3ZXSLqIAxJRN2QIAjwBOBjZLmbVgtBAPra9scAqbdoTbXQ+h94uUqFy2VlLY5a1SuVuF5RAaWbm5mfwnJ0///4SHr5wc5O/PhINBUA6iCR2MPGRnykrVcv46czicj6MSQRdTMq1S2oVEr89FOZ0VGQUlUjBEGABLb39Qe+TquFvUbT4qjVsZs3sU6tRn1dHert7ERrrHXujk1LAUgi/lmIqPtgSCLqZnS6BgiCBJJe/rCz8xCtuTcKYkosUSqVLU5M1ul0LY5aOWs0Js/dqaurQ72RSeLWGqSIqPtiSCLqpmwk0vsaBREEDQRBhyulldBVit8G4N/qJiiVSly8WIpGu7bP3YHmKnS627j0n3Koy6tFS0wNUjqdjlecEVG7YEgiIlGCoAUgAXr5GZ+YrKkAcAeQ+LZS0/LcHYnEttV9mRKkTAltpoQtZTe9Yo+IzMOQREQt6tWr9Xk5ptTc775MCVKmhDZTwta9OVs6QWdy34mo+2FIImqFIGjQ1FSB+vreosubmq5Bp2tqtUYQutZl8NbqfkOb6WHLtDlbRNR9MSR1c609e6sr3sOmM6nVt+CoqkT9T0ugtDUyD0Z9C07KMtRdTIDSzlm0RqO5A0fVNajVNzqyu2SG9hr9IqLuiyGpG6upqcHMmS9DLjc+v0KprEdFxXW4uXEOhhittgFOghYLJfboZ+suWnNKrcBHEPBniQ0CjNRUapuQotOgoeE/qK/vL1rD0SYiIuvCkNSNKRQKyOVKSKVL4OjoL1pz8+YxaDRvQKPhHYFb0reXFP1txUcdKv7//YT69rI3WiNX2kALNVRXUlBVuVW0hqNNRETWhSGpB3B09Df6nK/GRuOPk7BmpswTsqZRmQZo0RvAAoldi6NNGwQttFrxS9eJiKhzMSRRl2PKPCFrHZVpabRJrRZ/rAcREVkGQxJZndZGiRoaSuEkaFqcJ9RVR2W0ggClUob6+kuiy61thIzaR2sPCQYAV1dXeHl5dVKPiAhgSCIrY8ooUYPqBiSCGp4Sm241KqPQaThvqQcy5SHBACD18EDG7t0MSkSdiCGJrIrpV5MBArrXZHNT5y21dpWcUnkd4B1+ugxTHhJc0diITXI5FAoFQxJRJ2JI6sK68z2QTLmarLu636vk7o206Xharkvp7+CAYCfx3zsAgI9KIep0DEldlDXeA6mrXXEGtN5naxuVMWW0qbuOtHVlSqXyvh+4y3lLRJ2PIamLsrZ7IHXFK87Mmf9kbaMyLY02dfeRtq5EEDQQBB2ulFZCVykXrbn3wF21Wm10O5y3RGQZDEldnLXcA8mUuUTWNp+mJ89/os4hCFq0+pw47XUIwr+h1Rr/jpk6b+nNqiqcPXsWgYGBRrfF0SYi0zEkUbtqaS6Rtc6n6cnzn6hztPicuP//HTPllFxL85Y42kTU/hiSqNNwPg1Rc+11Sq5Oq4WdWo0/CwIGGglJV5uasPn6dV4lR2SiXpbuQHp6OoKCguDg4ICwsDAcOXKkxfr8/HyEhYXBwcEBAwcORGZmZrOa/fv3Y+jQoZBKpRg6dCgOHDhw3/vtbDU1Nbh06ZLRV9e+cu3ufBqxlwdHbqiHMTwlN0z0JZH0gyAADQ0NqK+vF33dqauDSqVCXZkMNy9cFn3d+k85Sq+UQy4XD2NEZMiiI0n79u1DQkIC0tPTMXr0aGzbtg3R0dG4cOECAgICmtWXlpYiJiYGsbGx2LVrF44ePYr4+Hh4eXlh6tSpAIDCwkJMnz4dr7/+Op599lkcOHAAf/jDH/D9999j5MiRbdpvZ2vPK9dMueJMp2u67xpruwoMaP3u1dbYZ+q5WjolJ5H0anW06ZKqETqdAEh8jM5/6qW6BY36P/jpp5/g4eFhtC9VVVVQtnLLAalUCh8fnxZrVCoV7O1b/odPe82Rau2WKF11X52tO3+2trBoSEpNTcW8efMwf/58AEBaWhoOHjyIjIwMpKSkNKvPzMxEQEAA0tLSAABDhgxBUVERNm7cqA9JaWlpmDhxIpKTkwEAycnJyM/PR1paGvbs2dOm/Xa29rpyzZSrt5rUt+CkLEPdxQQo7ZzbXGNtV4GZcvdqa+szkTEmTQDXVACog0TSUthSQKVSYs2aTLi47BOtaWy8g5r//Ah3W1tIJBIj/RFwS6NB3wcj4OAg/g8ntVqJa9dK0a/fINjaGv9T4+Ehxe7dGS3+0W3tD7dcLse6//1fCHfuGK0BAImLC157++0WA2JrwU4ul+N//3cd7txp+R9Ypnwua2PKP9CB9vmdAZ0botvKYiFJpVLh+PHjWLZsmUF7VFQUCgoKRNcpLCxEVFSUQdukSZOQlZUFtVoNOzs7FBYWIjExsVnNvWDVlv1ayv1euWb61VsC/iyxaWWekCk11jOXiPOfqDtqcQK4xK7V9QVBC0GQwM7uJbi7jxOtaWr6Gxy1x7DAvh/62buL1lSqarFJWwHgBaPbuXnzGBob34CNzSK4uw8WrWlsrEBV1ZstXpFnSihRKuuhKTuFjUEBGOAs/g+5K3V1WHr8FObOXQap1NjjjFoPdvdG8R966B24uBj7/3PrnwswLSS0V40pdWVlZbh+vR5OTq8a/Qd6e/3O2jNEdySLhaTa2lpotVp4e3sbtHt7e6Oqqkp0naqqKtF6jUaD2tpa+Pr6Gq25t8227Be4e+XJz4efb9++DQCtJuW2uHPnDpTKBtTWfos7dy6I1igUZ6HVNuHGjX9Co7kiWnP79knoBAH1ukbc0Yr/C6tRp4IOAhq1ynapKVPfgVopfhyvauo6vaYr9pk1rOmoGplGAY2gQ1NTudH/tzQ1VUInCGgStGgwMsraJOha3U5DQxm02iY0NFyGra34durrS1BWdgGxscthZyf+x1utbkRV1Q34+MxG7959RWvu3DkNXcMRnC8pQ7WRP7g1Gg0USiXs7oSiV6+Hjfanru4/qKt7Ak5O4qcSGxtLoFR+DoWiBID4iItpn0uFmppK9O0bAFtbmw6tMX1bd491YGAxNBrx/2+21+/MlGOtUt2AUvktKisrW7xi01z3/m6bchNXCBZSWVkpABAKCgoM2tetWyc89NBDous8+OCDwptvvmnQ9v333wsABJlMJgiCINjZ2Qm7d+82qNm1a5cglUrbvF9BEIRVq1YJuDuBhS+++OKLL7746uKvioqKVrOKxUaSPD09YWNj02z0prq6utkozz0+Pj6i9ba2tvpzzMZq7m2zLfsF7s5tSkpK0r/X6XS4ceMGPDw8jJ63byuFQgF/f39UVFTA1dW1Xbfd3fBYmY7HynQ8VqbjsTIdj5V5Oup4CYKAO3fuwM/Pr9Vai4Uke3t7hIWFIS8vD88++6y+PS8vD1OmTBFdJyIiAn//+98N2nJzcxEeHg47Ozt9TV5ensG8pNzcXERGRrZ5v8DdKzl+OdznbuTOt+3F1dWV/yGZiMfKdDxWpuOxMh2Plel4rMzTEcfLzc3NpDqLXt2WlJSEWbNmITw8HBEREXjvvfdQXl6OuLg4AHdHbyorK5GdnQ0AiIuLw5YtW5CUlITY2FgUFhYiKytLf9UaACxevBhjx47FW2+9hSlTpuBvf/sbvvnmG3z//fcm75eIiIjIoiFp+vTpkMvlWLt2LWQyGUJCQpCTk6OfMS+TyVBeXq6vDwoKQk5ODhITE7F161b4+flh8+bN+sv/ASAyMhJ79+7Fa6+9hpUrVyI4OBj79u3T3yPJlP0SERERWWziNhnX1NQkrFq1SmhqarJ0V6wej5XpeKxMx2NlOh4r0/FYmccajpdEEEy5Bo6IiIioZ7H4s9uIiIiIrBFDEhEREZEIhiQiIiIiEQxJViY9PR1BQUFwcHBAWFgYjhw5YukuWdzq1ashkUgMXj9/+rggCFi9ejX8/Pzg6OiI8ePH4/z58xbscef55z//icmTJ8PPzw8SiQSff/65wXJTjo1SqcTChQvh6ekJJycn/O53v8PVq1c78VN0jtaO1YsvvtjsezZq1CiDmp5yrFJSUvD444/DxcUFffv2xTPPPIOLFy8a1PC7dZcpx4rfrbsyMjLwyCOP6O97FBERga+++kq/3Bq/UwxJVmTfvn1ISEjAihUrcPLkSYwZMwbR0dEGt0HoqYYNGwaZTKZ/nT17Vr9sw4YNSE1NxZYtW/Cvf/0LPj4+mDhxIu608kTw7qC+vh7Dhw/Hli1bRJebcmwSEhJw4MAB7N27F99//z3q6urw29/+Flpt93rwb2vHCgB+/etfG3zPcnJyDJb3lGOVn5+PP//5zzh27Bjy8vKg0WgQFRWF+vp6fQ2/W3eZcqwAfrcAoH///li/fj2KiopQVFSEJ598ElOmTNEHIav8Tlnsujpq5oknnhDi4uIM2h5++GFh2bJlFuqRdVi1apUwfPhw0WU6nU7w8fER1q9fr29ramoS3NzchMzMzE7qoXUAIBw4cED/3pRjc+vWLcHOzk7Yu3evvqayslLo1auX8PXXX3da3zvbL4+VIAjCnDlzhClTphhdp6ceK0EQhOrqagGAkJ+fLwgCv1st+eWxEgR+t1rSp08f4f3337fa7xRHkqyESqXC8ePHERUVZdAeFRWFgoICC/XKepSUlMDPzw9BQUF4/vnncfnyZQBAaWkpqqqqDI6bVCrFuHHjevxxM+XYHD9+HGq12qDGz88PISEhPfL4HT58GH379sXgwYMRGxuL6upq/bKefKxu374NAHjggQcA8LvVkl8eq3v43TKk1Wqxd+9e1NfXIyIiwmq/UwxJVqK2thZarbbZQ3a9vb2bPYy3pxk5ciSys7Nx8OBBbN++HVVVVYiMjIRcLtcfGx635kw5NlVVVbC3t0efPn2M1vQU0dHR+Otf/4pvv/0WmzZtwr/+9S88+eSTUCqVAHrusRIEAUlJSfjVr36FkJAQAPxuGSN2rAB+t37u7NmzcHZ2hlQqRVxcHA4cOIChQ4da7XfKoo8loeYkEonBe0EQmrX1NNHR0fqfQ0NDERERgeDgYOzcuVM/+ZHHzbi2HJueePymT5+u/zkkJATh4eEIDAzEP/7xD/z+9783ul53P1YLFizAmTNnDJ5/eQ+/W4aMHSt+t/7roYcewqlTp3Dr1i3s378fc+bMQX5+vn65tX2nOJJkJTw9PWFjY9MsDVdXVzdL1j2dk5MTQkNDUVJSor/KjcetOVOOjY+PD1QqFW7evGm0pqfy9fVFYGAgSkpKAPTMY7Vw4UJ88cUX+O6779C/f399O79bzRk7VmJ68nfL3t4egwYNQnh4OFJSUjB8+HC8++67VvudYkiyEvb29ggLC0NeXp5Be15eHiIjIy3UK+ukVCpRXFwMX19fBAUFwcfHx+C4qVQq5Ofn9/jjZsqxCQsLg52dnUGNTCbDuXPnevzxk8vlqKiogK+vL4CedawEQcCCBQvw2Wef4dtvv0VQUJDBcn63/qu1YyWmJ3+3fkkQBCiVSuv9TnXIdHBqk7179wp2dnZCVlaWcOHCBSEhIUFwcnISrly5YumuWdSSJUuEw4cPC5cvXxaOHTsm/Pa3vxVcXFz0x2X9+vWCm5ub8Nlnnwlnz54VZsyYIfj6+goKhcLCPe94d+7cEU6ePCmcPHlSACCkpqYKJ0+eFMrKygRBMO3YxMXFCf379xe++eYb4cSJE8KTTz4pDB8+XNBoNJb6WB2ipWN1584dYcmSJUJBQYFQWloqfPfdd0JERITQr1+/HnmsXn75ZcHNzU04fPiwIJPJ9K+GhgZ9Db9bd7V2rPjd+q/k5GThn//8p1BaWiqcOXNGWL58udCrVy8hNzdXEATr/E4xJFmZrVu3CoGBgYK9vb0wYsQIg8tIe6rp06cLvr6+gp2dneDn5yf8/ve/F86fP69frtPphFWrVgk+Pj6CVCoVxo4dK5w9e9aCPe483333nQCg2WvOnDmCIJh2bBobG4UFCxYIDzzwgODo6Cj89re/FcrLyy3waTpWS8eqoaFBiIqKEry8vAQ7OzshICBAmDNnTrPj0FOOldhxAiB8+OGH+hp+t+5q7Vjxu/VfL730kv7vm5eXl/DUU0/pA5IgWOd3SiIIgtAxY1REREREXRfnJBERERGJYEgiIiIiEsGQRERERCSCIYmIiIhIBEMSERERkQiGJCIiIiIRDElEREREIhiSiIiIiEQwJBFRu7hy5QokEglOnTpl6a4QEbULhiQi6hHGjx+PhIQES3fDKhw+fBgSiQS3bt2ydFeIrBpDEhFZNZVKZekuGLC2/hBRx2FIIiKz6HQ6vPXWWxg0aBCkUikCAgLwxhtv6JdfvnwZEyZMQO/evTF8+HAUFhbql8nlcsyYMQP9+/dH7969ERoaij179hhsf/z48ViwYAGSkpLg6emJiRMnAgBSU1MRGhoKJycn+Pv7Iz4+HnV1dQbrHj16FOPGjUPv3r3Rp08fTJo0CTdv3sSLL76I/Px8vPvuu5BIJJBIJLhy5QoA4MKFC4iJiYGzszO8vb0xa9Ys1NbWttqf1atXIyAgAFKpFH5+fli0aJFJx0+pVOKVV16Bv78/pFIpHnzwQWRlZemX5+fn44knnoBUKoWvry+WLVsGjUajXz5gwACkpaUZbPPRRx/F6tWr9e8lEgnef/99PPvss+jduzcefPBBfPHFFwDunhadMGECAKBPnz6QSCR48cUXTeo7UU/DkEREZklOTsZbb72FlStX4sKFC9i9eze8vb31y1esWIGlS5fi1KlTGDx4MGbMmKH/I9/U1ISwsDB8+eWXOHfuHP70pz9h1qxZ+OGHHwz2sXPnTtja2uLo0aPYtm0bAKBXr17YvHkzzp07h507d+Lbb7/FK6+8ol/n1KlTeOqppzBs2DAUFhbi+++/x+TJk6HVavHuu+8iIiICsbGxkMlkkMlk8Pf3h0wmw7hx4/Doo4+iqKgIX3/9Na5fv44//OEPLfbn008/xTvvvINt27ahpKQEn3/+OUJDQ006frNnz8bevXuxefNmFBcXIzMzE87OzgCAyspKxMTE4PHHH8fp06eRkZGBrKwsrFu3zuzf05o1a/CHP/wBZ86cQUxMDF544QXcuHED/v7+2L9/PwDg4sWLkMlkePfdd83ePlGPIBARmUihUAhSqVTYvn17s2WlpaUCAOH999/Xt50/f14AIBQXFxvdZkxMjLBkyRL9+3HjxgmPPvpoq335+OOPBQ8PD/37GTNmCKNHjzZaP27cOGHx4sUGbStXrhSioqIM2ioqKgQAwsWLF432Z9OmTcLgwYMFlUrVaj9/7uLFiwIAIS8vT3T58uXLhYceekjQ6XT6tq1btwrOzs6CVqsVBEEQAgMDhXfeecdgveHDhwurVq3SvwcgvPbaa/r3dXV1gkQiEb766itBEAThu+++EwAIN2/eNKv/RD0NR5KIyGTFxcVQKpV46qmnjNY88sgj+p99fX0BANXV1QAArVaLN954A4888gg8PDzg7OyM3NxclJeXG2wjPDy82Xa/++47TJw4Ef369YOLiwtmz54NuVyO+vp6AP8dSTLH8ePH8d1338HZ2Vn/evjhhwEAly5dMtqfadOmobGxEQMHDkRsbCwOHDhgcErMmFOnTsHGxgbjxo0TXV5cXIyIiAhIJBJ92+jRo1FXV4erV6+a9dl+/ntwcnKCi4uL/vdARKZhSCIikzk6OrZaY2dnp//53h97nU4HANi0aRPeeecdvPLKK/j2229x6tQpTJo0qdlkaCcnJ4P3ZWVliImJQUhICPbv34/jx49j69atAAC1Wm1y335Jp9Nh8uTJOHXqlMGrpKQEY8eONdoff39/XLx4EVu3boWjoyPi4+MxduxYfV+Maa2PgiAYBKR7bcB/j2WvXr30bfeI7ffnv4d769/7PRCRaRiSiMhkDz74IBwdHXHo0KE2rX/kyBFMmTIFf/zjHzF8+HAMHDgQJSUlra5XVFQEjUaDTZs2YdSoURg8eDCuXbtmUPPII4+02C97e3totVqDthEjRuD8+fMYMGAABg0aZPD6ZTD6JUdHR/zud7/D5s2bcfjwYRQWFuLs2bMtrhMaGgqdTof8/HzR5UOHDkVBQYFBCCooKICLiwv69esHAPDy8oJMJtMvVygUKC0tbXG/v2Rvbw8AzY4HERliSCIikzk4OODVV1/FK6+8guzsbFy6dAnHjh0zuDqrJYMGDUJeXh4KCgpQXFyM//mf/0FVVVWr6wUHB0Oj0eD//u//cPnyZXz00UfIzMw0qElOTsa//vUvxMfH48yZM/j3v/+NjIwM/ZVqAwYMwA8//IArV66gtrYWOp0Of/7zn3Hjxg3MmDEDP/74Iy5fvozc3Fy89NJLLQaIHTt2ICsrC+fOndP3x9HREYGBgS1+jgEDBmDOnDl46aWX8Pnnn6O0tBSHDx/Gxx9/DACIj49HRUUFFi5ciH//+9/429/+hlWrViEpKQm9et393/WTTz6Jjz76CEeOHMG5c+cwZ84c2NjYtHoMfy4wMBASiQRffvklampqml0lSER3MSQRkVlWrlyJJUuW4C9/+QuGDBmC6dOnmzzXZeXKlRgxYgQmTZqE8ePHw8fHB88880yr6z366KNITU3FW2+9hZCQEPz1r39FSkqKQc3gwYORm5uL06dP44knnkBERAT+9re/wdbWFgCwdOlS2NjYYOjQofDy8kJ5eTn8/Pxw9OhRaLVaTJo0CSEhIVi8eDHc3Nz0oUSMu7s7tm/fjtGjR+tHsP7+97/Dw8Oj1c+SkZGB5557DvHx8Xj44YcRGxurn1fVr18/5OTk4Mcff8Tw4cMRFxeHefPm4bXXXtOvn5ycjLFjx+K3v/0tYmJi8MwzzyA4OLjV/f5cv379sGbNGixbtgze3t5YsGCBWesT9RQS4Zcnt4mIiIiII0lEREREYhiSiIjayZEjRwxuJ/DLFxF1LTzdRkTUThobG1FZWWl0+aBBgzqxN0R0vxiSiIiIiETwdBsRERGRCIYkIiIiIhEMSUREREQiGJKIiIiIRDAkEREREYlgSCIiIiISwZBEREREJIIhiYiIiEjE/wMGt+cMfTAQFwAAAABJRU5ErkJggg==",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "cols = ['characters_count', 'words_count', 'sentences_count']\n",
    "bins = [50, 25, 15]\n",
    "xlims = [(0, 300), (0, 100), (0, 15)]\n",
    "\n",
    "for col, bin, xlim in zip(cols, bins, xlims):\n",
    "    plt.hist(\n",
    "        get_hams(df)[col],\n",
    "        bins=bin,\n",
    "        alpha=0.7,\n",
    "        label='Ham',\n",
    "        color='blue',\n",
    "        density=True,\n",
    "        edgecolor='black',\n",
    "        range=xlim\n",
    "    )\n",
    "    plt.hist(\n",
    "        get_spams(df)[col],\n",
    "        bins=bin,\n",
    "        alpha=0.7,\n",
    "        label='Spam',\n",
    "        color='red',\n",
    "        density=True,\n",
    "        edgecolor='black',\n",
    "        range=xlim\n",
    "    )\n",
    "    plt.xlabel(col)\n",
    "    plt.ylabel('frequency')\n",
    "    plt.legend()\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "import seaborn as sns\n",
    "import warnings\n",
    "\n",
    "warnings.simplefilter(action='ignore', category=FutureWarning)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 808.875x750 with 12 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.pairplot(df, hue='target')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "corr = df.corr(numeric_only=True)\n",
    "mask = np.triu(np.ones_like(corr, dtype=bool))\n",
    "\n",
    "sns.heatmap(\n",
    "    corr,\n",
    "    cmap='coolwarm',\n",
    "    mask=mask,\n",
    ")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.preprocessing import MinMaxScaler\n",
    "\n",
    "scaler = MinMaxScaler()\n",
    "df['characters_count_scaled'] = scaler.fit_transform(df[['characters_count']])\n",
    "df['words_count_scaled'] = scaler.fit_transform(df[['words_count']])\n",
    "df['sentences_count_scaled'] = scaler.fit_transform(df[['sentences_count']])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Data preprocessing\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Install Stopwords\n",
    "\n",
    "The NLTK stopwords package is used to filter out common words (e.g., \"the\", \"is\", \"and\") from text, facilitating more focused analysis of content."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[nltk_data] Downloading package stopwords to /home/ushka/nltk_data...\n",
      "[nltk_data]   Package stopwords is already up-to-date!\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "True"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import string\n",
    "from nltk.stem.porter import PorterStemmer\n",
    "\n",
    "nltk.download('stopwords')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Stemming is the process of reducing words to their root or base form, often by removing suffixes and prefixes, to normalize variations of words for analysis or indexing in natural language processing tasks."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "stemmer = PorterStemmer()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Text transformation\n",
    "\n",
    "1. Lowercase\n",
    "1. Tokenize\n",
    "1. Remove special characters\n",
    "1. Remove stopwords and punctuation\n",
    "1. Stemming\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "def text_transform(text: str) -> str:\n",
    "    result = text.lower()\n",
    "    result = nltk.word_tokenize(result)\n",
    "\n",
    "    temp = []\n",
    "    for word in result:\n",
    "        if word.isalnum():\n",
    "            temp.append(word)\n",
    "    result = temp\n",
    "\n",
    "    temp = []\n",
    "    for word in result:\n",
    "        if word not in nltk.corpus.stopwords.words('english') and word not in string.punctuation:\n",
    "            temp.append(word)\n",
    "    result = temp\n",
    "\n",
    "    temp = []\n",
    "    for word in result:\n",
    "        temp.append(stemmer.stem(word))\n",
    "    result = temp\n",
    "\n",
    "    return \" \".join(result)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Original = FreeMsg Hey there darling it's been 3 week's now and no word back! I'd like some fun you up for it still? Tb ok! XxX std chgs to send, å£1.50 to rcv\n",
      "Transformed = freemsg hey darl 3 week word back like fun still tb ok xxx std chg send rcv\n"
     ]
    }
   ],
   "source": [
    "example_text = df['text'][5]\n",
    "\n",
    "print(\"Original =\", example_text)\n",
    "print(\"Transformed =\",  text_transform(example_text))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "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>text</th>\n",
       "      <th>text_transformed</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Go until jurong point, crazy.. Available only ...</td>\n",
       "      <td>go jurong point crazi avail bugi n great world...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Ok lar... Joking wif u oni...</td>\n",
       "      <td>ok lar joke wif u oni</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Free entry in 2 a wkly comp to win FA Cup fina...</td>\n",
       "      <td>free entri 2 wkli comp win fa cup final tkt 21...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>U dun say so early hor... U c already then say...</td>\n",
       "      <td>u dun say earli hor u c alreadi say</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Nah I don't think he goes to usf, he lives aro...</td>\n",
       "      <td>nah think goe usf live around though</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                text  \\\n",
       "0  Go until jurong point, crazy.. Available only ...   \n",
       "1                      Ok lar... Joking wif u oni...   \n",
       "2  Free entry in 2 a wkly comp to win FA Cup fina...   \n",
       "3  U dun say so early hor... U c already then say...   \n",
       "4  Nah I don't think he goes to usf, he lives aro...   \n",
       "\n",
       "                                    text_transformed  \n",
       "0  go jurong point crazi avail bugi n great world...  \n",
       "1                              ok lar joke wif u oni  \n",
       "2  free entri 2 wkli comp win fa cup final tkt 21...  \n",
       "3                u dun say earli hor u c alreadi say  \n",
       "4               nah think goe usf live around though  "
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['text_transformed'] = df['text'].apply(text_transform)\n",
    "df[['text', 'text_transformed']].head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Plot most popular words"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from collections import Counter\n",
    "\n",
    "ham_corpus = get_hams(df)['text_transformed'].str.split(\n",
    "    expand=True\n",
    ").stack()\n",
    "spam_corpus = get_spams(df)['text_transformed'].str.split(\n",
    "    expand=True\n",
    ").stack()\n",
    "\n",
    "n = 30\n",
    "ham_counter = Counter(ham_corpus)\n",
    "spam_counter = Counter(spam_corpus)\n",
    "\n",
    "plt.bar(\n",
    "    *zip(*ham_counter.most_common(n)),\n",
    "    color='blue'\n",
    ")\n",
    "plt.xticks(rotation=90)\n",
    "plt.ylabel('frequency')\n",
    "plt.title(f'Ham Top {n} Words')\n",
    "plt.show()\n",
    "\n",
    "plt.bar(\n",
    "    *zip(*spam_counter.most_common(n)),\n",
    "    color='red'\n",
    ")\n",
    "plt.xticks(rotation=90)\n",
    "plt.ylabel('frequency')\n",
    "plt.title(f'Spam Top {n} Words')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Data Augmentation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[nltk_data] Downloading package wordnet to /home/ushka/nltk_data...\n",
      "[nltk_data]   Package wordnet is already up-to-date!\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "True"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import random\n",
    "from nltk.tokenize import word_tokenize\n",
    "from nltk.corpus import wordnet\n",
    "\n",
    "nltk.download('wordnet')"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Synonim Replecement"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "def synonym_replacement(words, n):\n",
    "    new_words = words.copy()\n",
    "    random_word_list = list(set(\n",
    "        [word for word in words if word not in nltk.corpus.stopwords.words('english')]))\n",
    "    random.shuffle(random_word_list)\n",
    "    num_replaced = 0\n",
    "    for random_word in random_word_list:\n",
    "        synonyms = get_synonyms(random_word)\n",
    "        if len(synonyms) >= 1:\n",
    "            synonym = random.choice(synonyms)\n",
    "            new_words = [synonym if word ==\n",
    "                         random_word else word for word in new_words]\n",
    "            num_replaced += 1\n",
    "        if num_replaced >= n:\n",
    "            break\n",
    "\n",
    "    return new_words\n",
    "\n",
    "\n",
    "def get_synonyms(word):\n",
    "    synonyms = []\n",
    "    for syn in wordnet.synsets(word):\n",
    "        for lemma in syn.lemmas():\n",
    "            synonyms.append(lemma.name())\n",
    "    return list(set(synonyms))"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Random words inserting"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [],
   "source": [
    "def random_insertion(words, n):\n",
    "    new_words = words.copy()\n",
    "    for _ in range(n):\n",
    "        add_word(new_words)\n",
    "    return new_words\n",
    "\n",
    "\n",
    "def add_word(new_words):\n",
    "    synonyms = []\n",
    "    counter = 0\n",
    "    while len(synonyms) < 1:\n",
    "        random_word = new_words[random.randint(0, len(new_words)-1)]\n",
    "        synonyms = get_synonyms(random_word)\n",
    "        counter += 1\n",
    "        if counter >= 10:\n",
    "            return\n",
    "    random_synonym = synonyms[0]\n",
    "    random_idx = random.randint(0, len(new_words)-1)\n",
    "    new_words.insert(random_idx, random_synonym)"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Function to randomly swap words"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "def random_swap(words, n):\n",
    "    new_words = words.copy()\n",
    "    for _ in range(n):\n",
    "        new_words = swap_word(new_words)\n",
    "    return new_words\n",
    "\n",
    "\n",
    "def swap_word(new_words):\n",
    "    random_idx_1 = random.randint(0, len(new_words)-1)\n",
    "    random_idx_2 = random_idx_1\n",
    "    counter = 0\n",
    "    while random_idx_2 == random_idx_1:\n",
    "        random_idx_2 = random.randint(0, len(new_words)-1)\n",
    "        counter += 1\n",
    "        if counter > 3:\n",
    "            return new_words\n",
    "    new_words[random_idx_1], new_words[random_idx_2] = new_words[random_idx_2], new_words[random_idx_1]\n",
    "    return new_words"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Random Deletion "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "def random_deletion(words, p):\n",
    "    if len(words) == 1:\n",
    "        return words\n",
    "\n",
    "    new_words = []\n",
    "    for word in words:\n",
    "        r = random.uniform(0, 1)\n",
    "        if r > p:\n",
    "            new_words.append(word)\n",
    "\n",
    "    if len(new_words) == 0:\n",
    "        rand_int = random.randint(0, len(words)-1)\n",
    "        return [words[rand_int]]\n",
    "\n",
    "    return new_words"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Function to augment a sentence"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "def augment_sentence(sentence):\n",
    "    words = word_tokenize(sentence)\n",
    "    aug_methods = [synonym_replacement,\n",
    "                   random_insertion, random_swap, random_deletion]\n",
    "    random.shuffle(aug_methods)\n",
    "    for method in aug_methods:\n",
    "        if method == random_deletion:\n",
    "            words = method(words, p=0.1)\n",
    "        else:\n",
    "            words = method(words, n=1)\n",
    "    return ' '.join(words)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "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>text</th>\n",
       "      <th>text_augmented</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Go until jurong point, crazy.. Available only ...</td>\n",
       "      <td>only until jurong point , crazy Go in bugis n ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Ok lar... Joking wif u oni...</td>\n",
       "      <td>lar jest Joking ... wif u ONI ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Free entry in 2 a wkly comp to win FA Cup fina...</td>\n",
       "      <td>entry two in a wkly comp to win FA final tkts ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>U dun say so early hor... U c already then say...</td>\n",
       "      <td>U then so say dun early hor U c already then s...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Nah I don't think he goes to usf, he lives aro...</td>\n",
       "      <td>do n't cerebrate he goes to usf , lives he kno...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                text  \\\n",
       "0  Go until jurong point, crazy.. Available only ...   \n",
       "1                      Ok lar... Joking wif u oni...   \n",
       "2  Free entry in 2 a wkly comp to win FA Cup fina...   \n",
       "3  U dun say so early hor... U c already then say...   \n",
       "4  Nah I don't think he goes to usf, he lives aro...   \n",
       "\n",
       "                                      text_augmented  \n",
       "0  only until jurong point , crazy Go in bugis n ...  \n",
       "1                  lar jest Joking ... wif u ONI ...  \n",
       "2  entry two in a wkly comp to win FA final tkts ...  \n",
       "3  U then so say dun early hor U c already then s...  \n",
       "4  do n't cerebrate he goes to usf , lives he kno...  "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "df['text_augmented'] = df['text'].apply(augment_sentence)\n",
    "display(df[['text', 'text_augmented']].head())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Data preparation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.feature_extraction.text import TfidfVectorizer\n",
    "from sklearn.metrics import accuracy_score, confusion_matrix, precision_score, recall_score\n",
    "from sklearn.model_selection import train_test_split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [],
   "source": [
    "RS = 42"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Train/test split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(\n",
    "    df['text_transformed'],\n",
    "    df['target'],\n",
    "    test_size=0.2,\n",
    "    random_state=RS\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### CountVectorizer vs TfidfVectorizer\n",
    "\n",
    "TF-IDF is better than Count Vectorizers because it not only focuses on the frequency of words present in the corpus but also provides the importance of the words. We can then remove the words that are less important for analysis, hence making the model building less complex by reducing the input dimensions.\n",
    "\n",
    "<https://www.linkedin.com/pulse/count-vectorizers-vs-tfidf-natural-language-processing-sheel-saket/>\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Vectorizing\n",
    "\n",
    "Vectorizing is the process of converting text or categorical data into numerical vectors that machine learning models can understand and process."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'exact': 629,\n",
       " 'intent': 901,\n",
       " 'one': 1257,\n",
       " 'ok': 1250,\n",
       " 'next': 1202,\n",
       " 'time': 1782,\n",
       " 'free': 709,\n",
       " 'video': 1872,\n",
       " 'camera': 368,\n",
       " 'phone': 1329,\n",
       " 'half': 784,\n",
       " 'price': 1407,\n",
       " 'line': 1009,\n",
       " 'rental': 1483,\n",
       " '12': 15,\n",
       " 'mth': 1159,\n",
       " '500': 50,\n",
       " 'ntwk': 1228,\n",
       " 'min': 1105,\n",
       " '100': 11,\n",
       " 'txt': 1829,\n",
       " 'call': 360,\n",
       " 'mobileupd8': 1125,\n",
       " 'tick': 1778,\n",
       " 'could': 468,\n",
       " 'die': 536,\n",
       " 'know': 963,\n",
       " 'pout': 1386,\n",
       " 'stomp': 1680,\n",
       " 'feet': 667,\n",
       " 'need': 1191,\n",
       " 'tone': 1794,\n",
       " 'club': 429,\n",
       " 'sub': 1695,\n",
       " 'expir': 641,\n",
       " 'repli': 1485,\n",
       " 'mono': 1137,\n",
       " 'poli': 1366,\n",
       " 'weekli': 1914,\n",
       " '150p': 19,\n",
       " 'per': 1318,\n",
       " 'week': 1912,\n",
       " 'stop': 1681,\n",
       " 'msg': 1156,\n",
       " 'pleas': 1352,\n",
       " 'text': 1752,\n",
       " 'anymor': 162,\n",
       " 'noth': 1224,\n",
       " 'els': 599,\n",
       " 'say': 1537,\n",
       " 'pl': 1345,\n",
       " 'find': 678,\n",
       " 'aunt': 220,\n",
       " 'bruce': 328,\n",
       " 'amp': 147,\n",
       " 'mom': 1131,\n",
       " 'car': 376,\n",
       " 'ca': 351,\n",
       " 'park': 1301,\n",
       " 'long': 1028,\n",
       " 'stori': 1683,\n",
       " 'miss': 1112,\n",
       " 'allah': 132,\n",
       " 'tell': 1746,\n",
       " 'claim': 421,\n",
       " 'prize': 1411,\n",
       " 'enter': 610,\n",
       " 'ur': 1853,\n",
       " 'mobil': 1124,\n",
       " 'person': 1324,\n",
       " 'detail': 529,\n",
       " 'care': 379,\n",
       " 'stand': 1669,\n",
       " 'somebodi': 1634,\n",
       " 'set': 1568,\n",
       " 'websit': 1908,\n",
       " 'play': 1350,\n",
       " 'hold': 823,\n",
       " 'em': 601,\n",
       " 'use': 1860,\n",
       " 'eve': 620,\n",
       " 'onlin': 1259,\n",
       " 'cant': 374,\n",
       " 'keep': 949,\n",
       " 'talk': 1728,\n",
       " 'peopl': 1317,\n",
       " 'sure': 1715,\n",
       " 'pay': 1310,\n",
       " 'want': 1894,\n",
       " 'realli': 1459,\n",
       " 'buy': 345,\n",
       " 'much': 1161,\n",
       " 'will': 1935,\n",
       " 'dont': 563,\n",
       " 'worri': 1958,\n",
       " 'last': 975,\n",
       " 'hour': 838,\n",
       " 'might': 1102,\n",
       " 'get': 742,\n",
       " 'lucki': 1047,\n",
       " 'kiss': 959,\n",
       " 'feel': 666,\n",
       " 'haha': 781,\n",
       " 'awesom': 232,\n",
       " 'omw': 1256,\n",
       " 'back': 237,\n",
       " 'congratul': 455,\n",
       " 'award': 230,\n",
       " 'cd': 390,\n",
       " 'voucher': 1884,\n",
       " 'guarante': 773,\n",
       " 'entri': 612,\n",
       " 'wkli': 1948,\n",
       " 'draw': 568,\n",
       " 'music': 1166,\n",
       " '87066': 72,\n",
       " 'england': 606,\n",
       " 'news': 1201,\n",
       " 'nation': 1181,\n",
       " 'team': 1740,\n",
       " '87077': 73,\n",
       " 'eg': 594,\n",
       " 'tri': 1813,\n",
       " 'wale': 1889,\n",
       " 'scotland': 1542,\n",
       " 'way': 1903,\n",
       " 'make': 1058,\n",
       " 'train': 1808,\n",
       " 'street': 1687,\n",
       " 'hill': 813,\n",
       " 'fun': 722,\n",
       " 'night': 1207,\n",
       " 'caus': 387,\n",
       " 'wait': 1886,\n",
       " 'come': 440,\n",
       " 'home': 827,\n",
       " 'lone': 1027,\n",
       " 'today': 1789,\n",
       " 'os': 1277,\n",
       " 'run': 1522,\n",
       " 'without': 1944,\n",
       " 'instal': 896,\n",
       " 'hard': 792,\n",
       " 'copi': 464,\n",
       " 'import': 873,\n",
       " 'file': 674,\n",
       " 'system': 1723,\n",
       " 'give': 747,\n",
       " 'shop': 1581,\n",
       " 'sorri': 1646,\n",
       " 'help': 807,\n",
       " 'argh': 189,\n",
       " 'fuck': 719,\n",
       " 'nobodi': 1213,\n",
       " 'town': 1806,\n",
       " 'awak': 229,\n",
       " 'oh': 1247,\n",
       " 'rington': 1504,\n",
       " 'real': 1454,\n",
       " '1st': 24,\n",
       " 'join': 934,\n",
       " 'wk': 1946,\n",
       " 'gr8': 765,\n",
       " 'nokia': 1215,\n",
       " '8007': 63,\n",
       " 'hit': 816,\n",
       " 'love': 1039,\n",
       " 'south': 1649,\n",
       " 'recent': 1464,\n",
       " 'start': 1672,\n",
       " 'watch': 1901,\n",
       " 'offic': 1243,\n",
       " 'hey': 809,\n",
       " 'what': 1922,\n",
       " 'plan': 1348,\n",
       " 'sat': 1532,\n",
       " 'said': 1527,\n",
       " 'bad': 238,\n",
       " 'dat': 505,\n",
       " 'gal': 727,\n",
       " 'wat': 1900,\n",
       " 'mmm': 1116,\n",
       " 'that': 1759,\n",
       " 'better': 275,\n",
       " 'got': 763,\n",
       " 'roast': 1507,\n",
       " 'drink': 572,\n",
       " 'good': 757,\n",
       " 'indian': 886,\n",
       " 'ask': 204,\n",
       " 'abt': 84,\n",
       " 'movi': 1151,\n",
       " 'wan': 1892,\n",
       " 'oso': 1278,\n",
       " 'tmr': 1786,\n",
       " 'mayb': 1077,\n",
       " 'meet': 1085,\n",
       " 'hi': 810,\n",
       " 'wo': 1949,\n",
       " 'ard': 186,\n",
       " 'christma': 416,\n",
       " 'enjoy': 608,\n",
       " 'merri': 1093,\n",
       " 'think': 1764,\n",
       " 'go': 750,\n",
       " 'late': 976,\n",
       " 'school': 1540,\n",
       " 'especi': 617,\n",
       " 'class': 422,\n",
       " 'wednesday': 1910,\n",
       " 'probabl': 1413,\n",
       " 'fail': 649,\n",
       " 'test': 1751,\n",
       " 'friday': 714,\n",
       " 'day': 508,\n",
       " 'pain': 1293,\n",
       " 'stuff': 1692,\n",
       " 'basic': 247,\n",
       " 'drug': 576,\n",
       " 'sound': 1648,\n",
       " 'like': 1008,\n",
       " 'mani': 1063,\n",
       " 'talent': 1727,\n",
       " 'would': 1962,\n",
       " 'dinner': 545,\n",
       " 'date': 506,\n",
       " 'dad': 495,\n",
       " 'gon': 755,\n",
       " 'na': 1170,\n",
       " 'work': 1955,\n",
       " 'crazi': 478,\n",
       " 'question': 1432,\n",
       " 'lor': 1032,\n",
       " 'honey': 828,\n",
       " 'darl': 502,\n",
       " 'sexi': 1571,\n",
       " 'bun': 340,\n",
       " 'sugar': 1700,\n",
       " 'loverboy': 1041,\n",
       " 'boytoy': 316,\n",
       " 'ass': 208,\n",
       " 'gym': 778,\n",
       " 'network': 1196,\n",
       " 'technic': 1743,\n",
       " 'support': 1712,\n",
       " 'yep': 1985,\n",
       " 'pretti': 1404,\n",
       " 'longer': 1029,\n",
       " 'canada': 370,\n",
       " 'eh': 596,\n",
       " 'wow': 1963,\n",
       " 'must': 1167,\n",
       " 'done': 562,\n",
       " 'us': 1859,\n",
       " 'extra': 644,\n",
       " 'grin': 769,\n",
       " 'wil': 1934,\n",
       " 'lt': 1044,\n",
       " 'gt': 772,\n",
       " 'minut': 1110,\n",
       " 'space': 1651,\n",
       " 'luck': 1046,\n",
       " 'someon': 1635,\n",
       " 'new': 1199,\n",
       " 'hous': 839,\n",
       " 'easi': 587,\n",
       " 'ah': 113,\n",
       " 'select': 1555,\n",
       " 'mean': 1079,\n",
       " 'yeah': 1983,\n",
       " 'goe': 753,\n",
       " 'friend': 715,\n",
       " 'imma': 870,\n",
       " 'shit': 1579,\n",
       " 'anyth': 164,\n",
       " 'sleep': 1614,\n",
       " 'babe': 234,\n",
       " 'lost': 1035,\n",
       " 'aight': 118,\n",
       " 'someth': 1636,\n",
       " 'eat': 589,\n",
       " 'mu': 1160,\n",
       " 'kate': 947,\n",
       " 'even': 621,\n",
       " 'hope': 831,\n",
       " 'see': 1552,\n",
       " 'tomorrow': 1793,\n",
       " 'bit': 285,\n",
       " 'bloodi': 295,\n",
       " 'xxx': 1975,\n",
       " 'forgot': 702,\n",
       " 'leav': 986,\n",
       " 'put': 1428,\n",
       " 'meh': 1087,\n",
       " 'dun': 581,\n",
       " 'mind': 1106,\n",
       " 'seen': 1554,\n",
       " 'weight': 1915,\n",
       " 'gee': 737,\n",
       " 'pic': 1332,\n",
       " 'friendship': 716,\n",
       " 'dear': 512,\n",
       " 'near': 1187,\n",
       " 'hear': 799,\n",
       " 'live': 1016,\n",
       " 'cheer': 406,\n",
       " 'tear': 1741,\n",
       " 'alway': 141,\n",
       " 'gud': 774,\n",
       " 'ni8': 1203,\n",
       " 'plu': 1355,\n",
       " 'greet': 768,\n",
       " 'whenev': 1924,\n",
       " 'speak': 1654,\n",
       " 'still': 1677,\n",
       " 'exhaust': 637,\n",
       " 'went': 1919,\n",
       " 'everyth': 627,\n",
       " 'close': 427,\n",
       " 'fact': 648,\n",
       " 'hee': 802,\n",
       " 'lectur': 988,\n",
       " 'ìï': 1999,\n",
       " 'sch': 1539,\n",
       " 'finish': 681,\n",
       " 'cool': 463,\n",
       " 'lem': 993,\n",
       " 'around': 195,\n",
       " 'italian': 917,\n",
       " 'food': 696,\n",
       " 'invit': 906,\n",
       " 'xchat': 1971,\n",
       " 'final': 677,\n",
       " 'attempt': 214,\n",
       " 'contact': 458,\n",
       " 'chat': 401,\n",
       " '86688': 71,\n",
       " 'ldn': 981,\n",
       " '18': 23,\n",
       " 'yr': 1995,\n",
       " 'came': 367,\n",
       " 'look': 1030,\n",
       " 'flat': 690,\n",
       " 'seem': 1553,\n",
       " '50': 49,\n",
       " 'away': 231,\n",
       " 'wiv': 1945,\n",
       " 'woman': 1951,\n",
       " 'guy': 777,\n",
       " 'sinc': 1597,\n",
       " 'scream': 1543,\n",
       " 'princess': 1408,\n",
       " 'knock': 962,\n",
       " '80082': 64,\n",
       " 'gift': 744,\n",
       " 'store': 1682,\n",
       " 'choic': 414,\n",
       " 'cs': 483,\n",
       " 'age16': 111,\n",
       " 'happi': 791,\n",
       " 'noon': 1219,\n",
       " 'thank': 1756,\n",
       " 'saturday': 1533,\n",
       " 'brother': 325,\n",
       " 'role': 1511,\n",
       " 'model': 1128,\n",
       " 'cost': 466,\n",
       " 'guess': 775,\n",
       " 'isnt': 915,\n",
       " 'ya': 1977,\n",
       " 'wish': 1941,\n",
       " 'life': 1003,\n",
       " 'sri': 1666,\n",
       " 'da': 494,\n",
       " 'jst': 940,\n",
       " 'nw': 1233,\n",
       " 'tonight': 1795,\n",
       " 'open': 1264,\n",
       " 'room': 1513,\n",
       " 'thanksgiv': 1757,\n",
       " 'fone': 695,\n",
       " 'thru': 1774,\n",
       " 'differ': 539,\n",
       " 'decis': 515,\n",
       " 'believ': 269,\n",
       " 'workin': 1956,\n",
       " 'photo': 1330,\n",
       " 'shoot': 1580,\n",
       " 'well': 1917,\n",
       " 'great': 767,\n",
       " 'brotha': 324,\n",
       " 'ye': 1982,\n",
       " 'penc': 1315,\n",
       " 'fix': 686,\n",
       " 'via': 1871,\n",
       " 'access': 88,\n",
       " 'number': 1230,\n",
       " 'prepay': 1400,\n",
       " 'direct': 547,\n",
       " 'may': 1076,\n",
       " 'later': 977,\n",
       " 'lect': 987,\n",
       " 'havent': 797,\n",
       " 'cash': 383,\n",
       " 'custom': 489,\n",
       " 'servic': 1567,\n",
       " 'repres': 1487,\n",
       " '10p': 13,\n",
       " 'settl': 1569,\n",
       " 'year': 1984,\n",
       " 'sunday': 1705,\n",
       " 'holiday': 825,\n",
       " 'difficult': 540,\n",
       " 'eatin': 590,\n",
       " 'lunch': 1048,\n",
       " 'monday': 1134,\n",
       " 'dentist': 525,\n",
       " 'left': 989,\n",
       " 'saw': 1536,\n",
       " 'messag': 1095,\n",
       " 'weekend': 1913,\n",
       " 'yesterday': 1987,\n",
       " 'true': 1816,\n",
       " 'forget': 701,\n",
       " 'studi': 1691,\n",
       " 'sweet': 1718,\n",
       " 'heart': 801,\n",
       " 'chennai': 407,\n",
       " 'exam': 631,\n",
       " 'anytim': 166,\n",
       " 'sm': 1620,\n",
       " 'de': 509,\n",
       " 'ladi': 968,\n",
       " 'first': 682,\n",
       " 'second': 1549,\n",
       " 'take': 1725,\n",
       " 'place': 1346,\n",
       " 'feb': 663,\n",
       " 'remov': 1480,\n",
       " 'send': 1559,\n",
       " '87239': 74,\n",
       " 'knw': 964,\n",
       " 'tat': 1733,\n",
       " 'askd': 205,\n",
       " 'howev': 841,\n",
       " 'suggest': 1701,\n",
       " 'stay': 1675,\n",
       " 'abl': 83,\n",
       " 'everi': 624,\n",
       " '10': 10,\n",
       " 'releas': 1475,\n",
       " 'juli': 942,\n",
       " 'name': 1175,\n",
       " 'film': 676,\n",
       " 'bare': 246,\n",
       " 'wonder': 1952,\n",
       " 'lei': 992,\n",
       " 'readi': 1453,\n",
       " 'dude': 579,\n",
       " 'also': 140,\n",
       " 'tht': 1775,\n",
       " 'nyt': 1236,\n",
       " 'mmmmmm': 1118,\n",
       " 'ahmad': 116,\n",
       " 'begin': 267,\n",
       " 'closer': 428,\n",
       " 'side': 1591,\n",
       " 'cbe': 388,\n",
       " 'inform': 890,\n",
       " 'account': 90,\n",
       " 'delay': 520,\n",
       " 'rent': 1482,\n",
       " 'discuss': 551,\n",
       " 'anoth': 157,\n",
       " 'check': 405,\n",
       " 'hmmm': 819,\n",
       " 'kb': 948,\n",
       " 'power': 1387,\n",
       " 'yoga': 1993,\n",
       " 'dunno': 582,\n",
       " 'anot': 156,\n",
       " 'thk': 1766,\n",
       " 'liao': 999,\n",
       " 'didnt': 535,\n",
       " 'full': 721,\n",
       " 'escap': 616,\n",
       " 'fanci': 652,\n",
       " 'tomo': 1792,\n",
       " 'xma': 1972,\n",
       " 'ever': 623,\n",
       " 'reason': 1460,\n",
       " 'euro2004': 619,\n",
       " 'kept': 950,\n",
       " 'latest': 978,\n",
       " 'result': 1495,\n",
       " 'daili': 497,\n",
       " 'unsubscrib': 1848,\n",
       " 'forward': 704,\n",
       " 'comp': 445,\n",
       " 'word': 1954,\n",
       " 'polic': 1367,\n",
       " 'station': 1674,\n",
       " 'internet': 903,\n",
       " 'connect': 456,\n",
       " 'slow': 1618,\n",
       " 'thing': 1763,\n",
       " 'loan': 1019,\n",
       " 'credit': 480,\n",
       " '0800': 0,\n",
       " 'swim': 1719,\n",
       " 'pool': 1371,\n",
       " '40gb': 44,\n",
       " 'ipod': 910,\n",
       " 'mp3': 1152,\n",
       " 'player': 1351,\n",
       " 'collect': 437,\n",
       " '83355': 66,\n",
       " 'ibhltd': 855,\n",
       " 'ldnw15h': 982,\n",
       " 'bugi': 338,\n",
       " 'let': 997,\n",
       " 'floor': 693,\n",
       " 'window': 1938,\n",
       " 'shirt': 1578,\n",
       " 'sometim': 1638,\n",
       " 'mouth': 1149,\n",
       " 'dream': 569,\n",
       " 'world': 1957,\n",
       " 'joy': 939,\n",
       " 'lot': 1036,\n",
       " 'tv': 1827,\n",
       " 'show': 1586,\n",
       " 'rain': 1440,\n",
       " 'becom': 260,\n",
       " '2nd': 35,\n",
       " 'b4': 233,\n",
       " 'suck': 1699,\n",
       " 'download': 567,\n",
       " 'exe': 636,\n",
       " 'yun': 1996,\n",
       " 'btw': 332,\n",
       " 'wana': 1893,\n",
       " 'handl': 787,\n",
       " 'island': 914,\n",
       " 'album': 126,\n",
       " 'due': 580,\n",
       " 'frnd': 718,\n",
       " '62468': 55,\n",
       " 'hug': 845,\n",
       " 'lik': 1007,\n",
       " 'den': 524,\n",
       " 'cute': 491,\n",
       " 'none': 1218,\n",
       " 'hate': 795,\n",
       " 'slipper': 1617,\n",
       " 'gone': 756,\n",
       " '530': 52,\n",
       " 'ticket': 1779,\n",
       " 'straight': 1685,\n",
       " 'wet': 1921,\n",
       " 'xx': 1974,\n",
       " 'textoper': 1753,\n",
       " 'child': 410,\n",
       " 'caught': 386,\n",
       " 'till': 1781,\n",
       " 'appar': 175,\n",
       " 'tel': 1744,\n",
       " 'famili': 651,\n",
       " 'advanc': 102,\n",
       " 'jamster': 921,\n",
       " '2day': 33,\n",
       " 'normal': 1222,\n",
       " 'uniqu': 1842,\n",
       " 'rest': 1494,\n",
       " 'wot': 1961,\n",
       " 'outsid': 1283,\n",
       " 'alreadi': 137,\n",
       " 'ta': 1724,\n",
       " 'job': 931,\n",
       " 'avail': 223,\n",
       " 'co': 430,\n",
       " 'rock': 1508,\n",
       " 'rpli': 1518,\n",
       " 'ltd': 1045,\n",
       " 'cancel': 372,\n",
       " 'fish': 683,\n",
       " 'water': 1902,\n",
       " 'decim': 514,\n",
       " 'bill': 279,\n",
       " 'lol': 1026,\n",
       " 'hot': 836,\n",
       " 'air': 120,\n",
       " 'plenti': 1354,\n",
       " 'save': 1535,\n",
       " 'stress': 1688,\n",
       " 'money': 1135,\n",
       " 'sent': 1561,\n",
       " 'garag': 730,\n",
       " 'key': 952,\n",
       " 'address': 98,\n",
       " 'receiv': 1463,\n",
       " 'post': 1379,\n",
       " 'complet': 448,\n",
       " 'variou': 1868,\n",
       " 'locat': 1021,\n",
       " 'armand': 194,\n",
       " 'build': 339,\n",
       " 'consid': 457,\n",
       " 'smoke': 1624,\n",
       " 'issu': 916,\n",
       " 'california': 359,\n",
       " 'okay': 1251,\n",
       " 'snow': 1628,\n",
       " 'manag': 1062,\n",
       " 'hand': 786,\n",
       " 'two': 1828,\n",
       " 'librari': 1000,\n",
       " 'sick': 1590,\n",
       " 'request': 1488,\n",
       " 'congrat': 454,\n",
       " 'id': 859,\n",
       " 'luv': 1049,\n",
       " 'ive': 918,\n",
       " 'farm': 656,\n",
       " 'nope': 1220,\n",
       " 'daddi': 496,\n",
       " 'felt': 668,\n",
       " 'convey': 461,\n",
       " '750': 59,\n",
       " 'unlimit': 1845,\n",
       " 'camcord': 366,\n",
       " '08000930705': 2,\n",
       " 'troubl': 1815,\n",
       " 'big': 278,\n",
       " 'small': 1621,\n",
       " 'sun': 1704,\n",
       " 'thanx': 1758,\n",
       " 'gentleman': 741,\n",
       " 'treat': 1811,\n",
       " 'digniti': 542,\n",
       " 'respect': 1491,\n",
       " 'pa': 1288,\n",
       " 'appli': 177,\n",
       " 'oil': 1249,\n",
       " 'natur': 1182,\n",
       " 'notic': 1225,\n",
       " 'anyway': 167,\n",
       " 'nt': 1226,\n",
       " 'drive': 573,\n",
       " 'chikku': 409,\n",
       " 'hw': 853,\n",
       " 'cold': 435,\n",
       " 'heard': 800,\n",
       " 'violenc': 1875,\n",
       " 'condit': 452,\n",
       " 'ru': 1520,\n",
       " 'problem': 1414,\n",
       " 'far': 655,\n",
       " 'semest': 1558,\n",
       " 'best': 273,\n",
       " 'made': 1054,\n",
       " 'add': 97,\n",
       " 'least': 985,\n",
       " 'arrang': 196,\n",
       " 'hospit': 834,\n",
       " 'girl': 746,\n",
       " 'bed': 262,\n",
       " 'rather': 1447,\n",
       " 'pass': 1305,\n",
       " 'laptop': 972,\n",
       " 'stock': 1678,\n",
       " 'mon': 1133,\n",
       " 'muz': 1168,\n",
       " 'mob': 1122,\n",
       " '16': 22,\n",
       " 'enough': 609,\n",
       " 'august': 219,\n",
       " 'danc': 499,\n",
       " 'bf': 276,\n",
       " 'either': 598,\n",
       " 'idea': 860,\n",
       " 'content': 459,\n",
       " 'six': 1606,\n",
       " 'reach': 1451,\n",
       " 'pig': 1338,\n",
       " 'leh': 991,\n",
       " 'wake': 1888,\n",
       " 'smth': 1625,\n",
       " 'wif': 1932,\n",
       " 'scare': 1538,\n",
       " 'mah': 1055,\n",
       " 'loud': 1038,\n",
       " 'understand': 1840,\n",
       " 'sura': 1714,\n",
       " '2004': 27,\n",
       " '2optout': 37,\n",
       " 'ipad': 909,\n",
       " 'kind': 956,\n",
       " 'load': 1018,\n",
       " 'smile': 1623,\n",
       " 'gay': 734,\n",
       " 'upload': 1851,\n",
       " '08712460324': 7,\n",
       " 'aiyo': 124,\n",
       " 'poor': 1372,\n",
       " 'bath': 250,\n",
       " 'quit': 1435,\n",
       " 'fuckin': 720,\n",
       " 'compani': 446,\n",
       " 'usual': 1863,\n",
       " 'found': 705,\n",
       " 'cheap': 402,\n",
       " 'apologis': 173,\n",
       " 'somewher': 1639,\n",
       " 'pick': 1333,\n",
       " 'buck': 334,\n",
       " 'xy': 1976,\n",
       " 'batteri': 251,\n",
       " 'mr': 1153,\n",
       " 'uncl': 1839,\n",
       " 'term': 1749,\n",
       " 'mid': 1099,\n",
       " 'chanc': 396,\n",
       " 'win': 1936,\n",
       " 'custcar': 488,\n",
       " 'macho': 1051,\n",
       " 'budget': 336,\n",
       " 'bb': 253,\n",
       " 'bold': 301,\n",
       " 'dollar': 561,\n",
       " 'monkey': 1136,\n",
       " 'face': 646,\n",
       " 'bitch': 286,\n",
       " 'rite': 1505,\n",
       " 'confirm': 453,\n",
       " 'mum': 1162,\n",
       " 'advic': 104,\n",
       " 'cardiff': 378,\n",
       " 'entitl': 611,\n",
       " 'updat': 1849,\n",
       " 'colour': 439,\n",
       " 'winner': 1940,\n",
       " 'excit': 634,\n",
       " 'soon': 1643,\n",
       " 'eye': 645,\n",
       " 'visit': 1877,\n",
       " 'futur': 724,\n",
       " 'present': 1402,\n",
       " 'fine': 679,\n",
       " 'juz': 944,\n",
       " 'anythin': 165,\n",
       " 'special': 1655,\n",
       " 'happen': 790,\n",
       " '3g': 42,\n",
       " 'wid': 1931,\n",
       " 'mate': 1071,\n",
       " 'game': 728,\n",
       " 'nolin': 1217,\n",
       " 'flirt': 692,\n",
       " 'paperwork': 1297,\n",
       " 'cours': 471,\n",
       " 'boss': 307,\n",
       " 'afraid': 106,\n",
       " 'bear': 257,\n",
       " 'nat': 1180,\n",
       " 'rate': 1446,\n",
       " 'rofl': 1509,\n",
       " 'yet': 1988,\n",
       " 'im': 868,\n",
       " 'pete': 1325,\n",
       " 'holla': 826,\n",
       " 'excus': 635,\n",
       " 'wrong': 1966,\n",
       " 'rd': 1450,\n",
       " 'though': 1769,\n",
       " 'hell': 804,\n",
       " 'book': 303,\n",
       " 'return': 1496,\n",
       " 'immedi': 871,\n",
       " 'nice': 1204,\n",
       " 'catch': 385,\n",
       " 'ave': 225,\n",
       " 'often': 1246,\n",
       " 'express': 643,\n",
       " 'hurt': 851,\n",
       " 'boston': 308,\n",
       " 'statement': 1673,\n",
       " 'chang': 397,\n",
       " 'realiz': 1458,\n",
       " 'nyc': 1235,\n",
       " 'birthday': 283,\n",
       " 'offer': 1242,\n",
       " 'packag': 1290,\n",
       " 'themob': 1762,\n",
       " 'info': 889,\n",
       " '8th': 76,\n",
       " 'campu': 369,\n",
       " 'atm': 212,\n",
       " 'sea': 1544,\n",
       " 'coupl': 470,\n",
       " 'bell': 270,\n",
       " 'never': 1198,\n",
       " 'choos': 415,\n",
       " 'clean': 423,\n",
       " 'embarass': 603,\n",
       " 'drop': 575,\n",
       " 'morn': 1143,\n",
       " 'old': 1254,\n",
       " 'man': 1061,\n",
       " 'fight': 672,\n",
       " 'ai': 117,\n",
       " 'thinkin': 1765,\n",
       " 'goin': 754,\n",
       " 'mail': 1056,\n",
       " 'god': 752,\n",
       " 'pray': 1393,\n",
       " 'ill': 867,\n",
       " 'review': 1498,\n",
       " 'cover': 472,\n",
       " 'kick': 953,\n",
       " 'perform': 1320,\n",
       " 'iscom': 912,\n",
       " 'tnc': 1787,\n",
       " 'india': 885,\n",
       " 'lead': 983,\n",
       " 'lar': 973,\n",
       " 'tot': 1800,\n",
       " 'interest': 902,\n",
       " 'pleasur': 1353,\n",
       " 'pour': 1385,\n",
       " 'becoz': 261,\n",
       " 'glad': 748,\n",
       " 'total': 1801,\n",
       " 'handset': 788,\n",
       " 'del': 519,\n",
       " 'walk': 1890,\n",
       " 'faster': 658,\n",
       " 'hungri': 848,\n",
       " 'actual': 95,\n",
       " 'right': 1502,\n",
       " 'orang': 1271,\n",
       " '08000839402': 1,\n",
       " 'symbol': 1722,\n",
       " 'mumtaz': 1164,\n",
       " '4th': 47,\n",
       " 'wife': 1933,\n",
       " 'kill': 955,\n",
       " 'husband': 852,\n",
       " 'marri': 1068,\n",
       " 'th': 1754,\n",
       " 'deliveri': 523,\n",
       " 'sister': 1602,\n",
       " 'aris': 192,\n",
       " 'email': 602,\n",
       " 'wont': 1953,\n",
       " 'touch': 1802,\n",
       " 'permiss': 1323,\n",
       " 'matter': 1073,\n",
       " 'lmao': 1017,\n",
       " 'oni': 1258,\n",
       " 'busi': 343,\n",
       " 'wun': 1968,\n",
       " 'disturb': 554,\n",
       " 'beauti': 258,\n",
       " 'countri': 469,\n",
       " 'bank': 244,\n",
       " 'villag': 1874,\n",
       " 'road': 1506,\n",
       " 'forev': 700,\n",
       " 'social': 1629,\n",
       " 'project': 1418,\n",
       " 'citizen': 420,\n",
       " 'monthli': 1139,\n",
       " 'rich': 1500,\n",
       " 'politician': 1368,\n",
       " 'haiz': 783,\n",
       " 'yup': 1997,\n",
       " 'teacher': 1739,\n",
       " 'darren': 504,\n",
       " 'cut': 490,\n",
       " 'short': 1582,\n",
       " 'sir': 1601,\n",
       " 'huh': 846,\n",
       " 'cup': 486,\n",
       " 'end': 604,\n",
       " 'secret': 1550,\n",
       " 'admir': 99,\n",
       " 'reveal': 1497,\n",
       " 'write': 1964,\n",
       " 'voda': 1879,\n",
       " 'match': 1070,\n",
       " '08712300220': 5,\n",
       " 'quot': 1437,\n",
       " 'code': 432,\n",
       " 'standard': 1670,\n",
       " 'app': 174,\n",
       " '7250i': 58,\n",
       " 'auction': 216,\n",
       " 'part': 1302,\n",
       " '86021': 70,\n",
       " 'welcom': 1916,\n",
       " 'english': 607,\n",
       " 'shall': 1572,\n",
       " 'hello': 806,\n",
       " 'bore': 305,\n",
       " 'thought': 1770,\n",
       " 'roommat': 1514,\n",
       " 'dubsack': 578,\n",
       " 'bring': 321,\n",
       " 'in2': 877,\n",
       " 'an': 149,\n",
       " 'current': 487,\n",
       " 'pound': 1384,\n",
       " 'maxim': 1075,\n",
       " 'cc': 389,\n",
       " '08718720201': 8,\n",
       " 'po': 1358,\n",
       " 'box': 314,\n",
       " 'abiola': 82,\n",
       " 'pop': 1373,\n",
       " 'instead': 897,\n",
       " 'tuition': 1825,\n",
       " 'surf': 1716,\n",
       " 'almost': 134,\n",
       " 'print': 1409,\n",
       " 'forum': 703,\n",
       " 'prob': 1412,\n",
       " 'replac': 1484,\n",
       " 'ignor': 864,\n",
       " 'wyli': 1969,\n",
       " 'weed': 1911,\n",
       " 'carlo': 381,\n",
       " 'jan': 922,\n",
       " 'al': 125,\n",
       " 'ofic': 1245,\n",
       " 'fr': 706,\n",
       " 'market': 1067,\n",
       " 'occur': 1240,\n",
       " 'pobox': 1359,\n",
       " 'sooner': 1644,\n",
       " 'gari': 731,\n",
       " 'ptbo': 1423,\n",
       " 'cheaper': 403,\n",
       " 'hiya': 817,\n",
       " 'travel': 1810,\n",
       " 'tough': 1803,\n",
       " 'improv': 876,\n",
       " 'month': 1138,\n",
       " 'angri': 152,\n",
       " 'aft': 107,\n",
       " 'supervisor': 1710,\n",
       " 'diff': 538,\n",
       " 'wen': 1918,\n",
       " 'bcum': 255,\n",
       " 'dnt': 555,\n",
       " 'serious': 1565,\n",
       " 'coz': 473,\n",
       " 'deep': 516,\n",
       " 'ha': 779,\n",
       " 'ju': 941,\n",
       " 'ate': 210,\n",
       " 'ar': 184,\n",
       " 'discount': 550,\n",
       " 'charg': 398,\n",
       " 'quick': 1433,\n",
       " 'legal': 990,\n",
       " 'trip': 1814,\n",
       " 'yiju': 1990,\n",
       " 'group': 770,\n",
       " 'askin': 206,\n",
       " 'earlier': 585,\n",
       " 'haf': 780,\n",
       " 'infern': 888,\n",
       " 'affair': 105,\n",
       " 'meant': 1080,\n",
       " 'calcul': 356,\n",
       " 'expens': 639,\n",
       " 'practic': 1392,\n",
       " 'decid': 513,\n",
       " 'oop': 1263,\n",
       " 'center': 393,\n",
       " 'knew': 961,\n",
       " 'break': 318,\n",
       " 'midnight': 1101,\n",
       " 'til': 1780,\n",
       " 'march': 1065,\n",
       " 'fran': 707,\n",
       " 'math': 1072,\n",
       " 'urgent': 1855,\n",
       " 'bonu': 302,\n",
       " 'caller': 362,\n",
       " 'asap': 202,\n",
       " 'polyphon': 1369,\n",
       " 'super': 1709,\n",
       " 'sn': 1626,\n",
       " 'subscript': 1697,\n",
       " 'slice': 1616,\n",
       " 'tho': 1768,\n",
       " 'mark': 1066,\n",
       " 'turn': 1826,\n",
       " ...}"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "tfidf = TfidfVectorizer(max_features=2000)\n",
    "tfidf.fit(X_train)\n",
    "vocab = tfidf.vocabulary_\n",
    "\n",
    "display(vocab)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train = tfidf.transform(X_train).toarray()   # type: ignore\n",
    "X_test = tfidf.transform(X_test).toarray()  # type: ignore"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Scaling\n",
    "\n",
    "With TF-IDF vectorization, scaling isn't typically needed because TF-IDF (Term Frequency-Inverse Document Frequency) already normalizes the importance of each term based on its frequency across documents, ensuring that all features are on a comparable scale for machine learning models.\n",
    "\n",
    "<https://stackoverflow.com/a/36678046/12197778>\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [],
   "source": [
    "# scaler = MinMaxScaler()\n",
    "# scaler.fit(X_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [],
   "source": [
    "# X_train = scaler.transform(X_train)\n",
    "# X_test = scaler.transform(X_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Scoring\n",
    "\n",
    "#### Scores\n",
    "\n",
    "- $TP$ - True positive (spam classified as spam)\n",
    "- $TN$ - True negative (ham classified as ham)\n",
    "- $FP$ - False positive (ham classified as spam)\n",
    "- $FN$ - False negative (spam classified as ham)\n",
    "\n",
    "#### Metrics\n",
    "\n",
    "- Accuracy = $(TP + TN) / (TP + TN + FP + FN)$\n",
    "- Precision = $TP / (TP + FP)$\n",
    "  - How many predicted positives are actually positive\n",
    "- Recall = $TP / (TP + FN)$\n",
    "  - How many actual positives were detected\n",
    "\n",
    "#### Confusion matrix\n",
    "\n",
    "|                     | Predicted Negative | Predicted Positive |\n",
    "| ------------------- | ------------------ | ------------------ |\n",
    "| **Actual Negative** | TN                 | FP                 |\n",
    "| **Actual Positive** | FN                 | TP                 |\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [],
   "source": [
    "def calc_scores(clf, X_test, y_test):\n",
    "    y_pred = clf.predict(X_test)\n",
    "\n",
    "    acc = accuracy_score(y_test, y_pred)\n",
    "    prec = precision_score(y_test, y_pred)\n",
    "    rec = recall_score(y_test, y_pred)\n",
    "    cm = confusion_matrix(y_test, y_pred)\n",
    "\n",
    "    return acc, prec, rec, cm\n",
    "\n",
    "\n",
    "def print_scores(acc, prec, rec, cm):\n",
    "    output = f\"Accuracy: {acc}\\n\"\n",
    "    output += f\"Precision: {prec}\\n\"\n",
    "    output += f\"Recall: {rec}\\n\"\n",
    "    output += \"Confusion Matrix:\\n\"\n",
    "    output += f\"{cm}\\n\"\n",
    "\n",
    "    print(output)\n",
    "\n",
    "\n",
    "def create_perf_df(scores_dict):\n",
    "    perf_df = pd.DataFrame(\n",
    "        scores_dict,\n",
    "        index=[\n",
    "            'accuracy',\n",
    "            'precision',\n",
    "            'recall',\n",
    "            'confusion_matrix'\n",
    "        ]\n",
    "    )\n",
    "    perf_df = perf_df.drop('confusion_matrix')\n",
    "    perf_df = perf_df.sort_values(by='precision', axis=1, ascending=False)\n",
    "    perf_df = perf_df.T\n",
    "    return perf_df"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Model building"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Test Naive Bayes models"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.naive_bayes import GaussianNB, MultinomialNB, BernoulliNB"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [],
   "source": [
    "nb_clfs = {\n",
    "    'Gaussian Naive Bayes': GaussianNB(),\n",
    "    'Multinomial Naive Bayes': MultinomialNB(),\n",
    "    'Bernoulli Naive Bayes': BernoulliNB()\n",
    "}\n",
    "\n",
    "\n",
    "nb_scores_dict = {}\n",
    "for name, clf in nb_clfs.items():\n",
    "    clf.fit(X_train, y_train)\n",
    "    scores = calc_scores(clf, X_test, y_test)\n",
    "    nb_scores_dict[name] = scores"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "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>accuracy</th>\n",
       "      <th>precision</th>\n",
       "      <th>recall</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Multinomial Naive Bayes</th>\n",
       "      <td>0.973888</td>\n",
       "      <td>0.983607</td>\n",
       "      <td>0.827586</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Bernoulli Naive Bayes</th>\n",
       "      <td>0.974855</td>\n",
       "      <td>0.954198</td>\n",
       "      <td>0.862069</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Gaussian Naive Bayes</th>\n",
       "      <td>0.823985</td>\n",
       "      <td>0.436426</td>\n",
       "      <td>0.875862</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                         accuracy precision    recall\n",
       "Multinomial Naive Bayes  0.973888  0.983607  0.827586\n",
       "Bernoulli Naive Bayes    0.974855  0.954198  0.862069\n",
       "Gaussian Naive Bayes     0.823985  0.436426  0.875862"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "nb_perf_df = create_perf_df(nb_scores_dict)\n",
    "display(nb_perf_df)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Find best performing models"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.svm import SVC\n",
    "from sklearn.naive_bayes import MultinomialNB\n",
    "from sklearn.tree import DecisionTreeClassifier\n",
    "from sklearn.neighbors import KNeighborsClassifier\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from sklearn.ensemble import AdaBoostClassifier\n",
    "from sklearn.ensemble import BaggingClassifier\n",
    "from sklearn.ensemble import ExtraTreesClassifier\n",
    "from sklearn.ensemble import GradientBoostingClassifier\n",
    "from xgboost import XGBClassifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [],
   "source": [
    "clfs = {\n",
    "    'LR': LogisticRegression(solver='liblinear', penalty='l1', random_state=RS),\n",
    "    'SVC': SVC(kernel='sigmoid', gamma=1.0, random_state=RS),\n",
    "    'MNB': MultinomialNB(),\n",
    "    'DTC': DecisionTreeClassifier(max_depth=5, random_state=RS),\n",
    "    'KNC': KNeighborsClassifier(),\n",
    "    'RFC': RandomForestClassifier(n_estimators=50, random_state=RS),\n",
    "    'ABC': AdaBoostClassifier(n_estimators=50, random_state=RS),\n",
    "    'BC': BaggingClassifier(n_estimators=50, random_state=RS),\n",
    "    'ETC': ExtraTreesClassifier(n_estimators=50, random_state=RS),\n",
    "    'GBC': GradientBoostingClassifier(n_estimators=50, random_state=RS),\n",
    "    'XGBC': XGBClassifier(n_estimators=50, random_state=RS)\n",
    "}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [],
   "source": [
    "clfs_scores_dict = {}\n",
    "\n",
    "for name, clf in clfs.items():\n",
    "    clf.fit(X_train, y_train)\n",
    "    scores = calc_scores(clf, X_test, y_test)\n",
    "    clfs_scores_dict[name] = scores"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "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>accuracy</th>\n",
       "      <th>precision</th>\n",
       "      <th>recall</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>KNC</th>\n",
       "      <td>0.910058</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.358621</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>RFC</th>\n",
       "      <td>0.970986</td>\n",
       "      <td>0.991453</td>\n",
       "      <td>0.8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MNB</th>\n",
       "      <td>0.973888</td>\n",
       "      <td>0.983607</td>\n",
       "      <td>0.827586</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ETC</th>\n",
       "      <td>0.976789</td>\n",
       "      <td>0.961832</td>\n",
       "      <td>0.868966</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SVC</th>\n",
       "      <td>0.978723</td>\n",
       "      <td>0.955556</td>\n",
       "      <td>0.889655</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>GBC</th>\n",
       "      <td>0.945841</td>\n",
       "      <td>0.949495</td>\n",
       "      <td>0.648276</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>XGBC</th>\n",
       "      <td>0.965184</td>\n",
       "      <td>0.936</td>\n",
       "      <td>0.806897</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BC</th>\n",
       "      <td>0.964217</td>\n",
       "      <td>0.897059</td>\n",
       "      <td>0.841379</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>LR</th>\n",
       "      <td>0.954545</td>\n",
       "      <td>0.876923</td>\n",
       "      <td>0.786207</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ABC</th>\n",
       "      <td>0.951644</td>\n",
       "      <td>0.868217</td>\n",
       "      <td>0.772414</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>DTC</th>\n",
       "      <td>0.928433</td>\n",
       "      <td>0.803419</td>\n",
       "      <td>0.648276</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      accuracy precision    recall\n",
       "KNC   0.910058       1.0  0.358621\n",
       "RFC   0.970986  0.991453       0.8\n",
       "MNB   0.973888  0.983607  0.827586\n",
       "ETC   0.976789  0.961832  0.868966\n",
       "SVC   0.978723  0.955556  0.889655\n",
       "GBC   0.945841  0.949495  0.648276\n",
       "XGBC  0.965184     0.936  0.806897\n",
       "BC    0.964217  0.897059  0.841379\n",
       "LR    0.954545  0.876923  0.786207\n",
       "ABC   0.951644  0.868217  0.772414\n",
       "DTC   0.928433  0.803419  0.648276"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "clfs_perf_df = create_perf_df(clfs_scores_dict)\n",
    "display(clfs_perf_df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1000x500 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "bar_width = 0.25\n",
    "\n",
    "r1 = np.arange(len(clfs_perf_df)) - bar_width\n",
    "r2 = [x + bar_width for x in r1]\n",
    "r3 = [x + bar_width for x in r2]\n",
    "\n",
    "plt.figure(figsize=(10, 5))\n",
    "plt.bar(\n",
    "    r1,\n",
    "    clfs_perf_df['accuracy'],\n",
    "    label='Accuracy',\n",
    "    width=bar_width,\n",
    "    color='orange'\n",
    ")\n",
    "plt.bar(\n",
    "    r2,\n",
    "    clfs_perf_df['precision'],\n",
    "    label='Precision',\n",
    "    width=bar_width,\n",
    "    color='green'\n",
    ")\n",
    "plt.bar(\n",
    "    r3,\n",
    "    clfs_perf_df['recall'],\n",
    "    label='Recall',\n",
    "    width=bar_width,\n",
    "    color='red'\n",
    ")\n",
    "plt.xticks(\n",
    "    rotation=90,\n",
    "    ticks=range(len(clfs_perf_df)),\n",
    "    labels=clfs_perf_df.index  # type: ignore\n",
    ")\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### **Best performing models**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [],
   "source": [
    "clfs_final = {\n",
    "    'RFC': clfs['RFC'],\n",
    "    'MNB': clfs['MNB'],\n",
    "    'ETC': clfs['ETC'],\n",
    "    'SVC': clfs['SVC'],\n",
    "}"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Models optimization"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Find best vectorizer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import GridSearchCV\n",
    "from sklearn.pipeline import Pipeline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [],
   "source": [
    "# from sklearn.base import TransformerMixin\n",
    "\n",
    "# class ToArrayTransformer(TransformerMixin):\n",
    "#     def fit(self, X, y=None):\n",
    "#         return self\n",
    "\n",
    "#     def transform(self, X):\n",
    "#         return X.toarray()\n",
    "\n",
    "# scalers_2 = [\n",
    "#     None,\n",
    "#     MinMaxScaler()\n",
    "# ]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<style>#sk-container-id-1 {color: black;background-color: white;}#sk-container-id-1 pre{padding: 0;}#sk-container-id-1 div.sk-toggleable {background-color: white;}#sk-container-id-1 label.sk-toggleable__label {cursor: pointer;display: block;width: 100%;margin-bottom: 0;padding: 0.3em;box-sizing: border-box;text-align: center;}#sk-container-id-1 label.sk-toggleable__label-arrow:before {content: \"▸\";float: left;margin-right: 0.25em;color: #696969;}#sk-container-id-1 label.sk-toggleable__label-arrow:hover:before {color: black;}#sk-container-id-1 div.sk-estimator:hover label.sk-toggleable__label-arrow:before {color: black;}#sk-container-id-1 div.sk-toggleable__content {max-height: 0;max-width: 0;overflow: hidden;text-align: left;background-color: #f0f8ff;}#sk-container-id-1 div.sk-toggleable__content pre {margin: 0.2em;color: black;border-radius: 0.25em;background-color: #f0f8ff;}#sk-container-id-1 input.sk-toggleable__control:checked~div.sk-toggleable__content {max-height: 200px;max-width: 100%;overflow: auto;}#sk-container-id-1 input.sk-toggleable__control:checked~label.sk-toggleable__label-arrow:before {content: \"▾\";}#sk-container-id-1 div.sk-estimator input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-1 div.sk-label input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-1 input.sk-hidden--visually {border: 0;clip: rect(1px 1px 1px 1px);clip: rect(1px, 1px, 1px, 1px);height: 1px;margin: -1px;overflow: hidden;padding: 0;position: absolute;width: 1px;}#sk-container-id-1 div.sk-estimator {font-family: monospace;background-color: #f0f8ff;border: 1px dotted black;border-radius: 0.25em;box-sizing: border-box;margin-bottom: 0.5em;}#sk-container-id-1 div.sk-estimator:hover {background-color: #d4ebff;}#sk-container-id-1 div.sk-parallel-item::after {content: \"\";width: 100%;border-bottom: 1px solid gray;flex-grow: 1;}#sk-container-id-1 div.sk-label:hover label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-1 div.sk-serial::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 0;bottom: 0;left: 50%;z-index: 0;}#sk-container-id-1 div.sk-serial {display: flex;flex-direction: column;align-items: center;background-color: white;padding-right: 0.2em;padding-left: 0.2em;position: relative;}#sk-container-id-1 div.sk-item {position: relative;z-index: 1;}#sk-container-id-1 div.sk-parallel {display: flex;align-items: stretch;justify-content: center;background-color: white;position: relative;}#sk-container-id-1 div.sk-item::before, #sk-container-id-1 div.sk-parallel-item::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 0;bottom: 0;left: 50%;z-index: -1;}#sk-container-id-1 div.sk-parallel-item {display: flex;flex-direction: column;z-index: 1;position: relative;background-color: white;}#sk-container-id-1 div.sk-parallel-item:first-child::after {align-self: flex-end;width: 50%;}#sk-container-id-1 div.sk-parallel-item:last-child::after {align-self: flex-start;width: 50%;}#sk-container-id-1 div.sk-parallel-item:only-child::after {width: 0;}#sk-container-id-1 div.sk-dashed-wrapped {border: 1px dashed gray;margin: 0 0.4em 0.5em 0.4em;box-sizing: border-box;padding-bottom: 0.4em;background-color: white;}#sk-container-id-1 div.sk-label label {font-family: monospace;font-weight: bold;display: inline-block;line-height: 1.2em;}#sk-container-id-1 div.sk-label-container {text-align: center;}#sk-container-id-1 div.sk-container {/* jupyter's `normalize.less` sets `[hidden] { display: none; }` but bootstrap.min.css set `[hidden] { display: none !important; }` so we also need the `!important` here to be able to override the default hidden behavior on the sphinx rendered scikit-learn.org. See: https://github.com/scikit-learn/scikit-learn/issues/21755 */display: inline-block !important;position: relative;}#sk-container-id-1 div.sk-text-repr-fallback {display: none;}</style><div id=\"sk-container-id-1\" class=\"sk-top-container\"><div class=\"sk-text-repr-fallback\"><pre>GridSearchCV(cv=5,\n",
       "             estimator=Pipeline(memory=&#x27;cachedir&#x27;,\n",
       "                                steps=[(&#x27;vectorizer&#x27;, None),\n",
       "                                       (&#x27;classifier&#x27;, None)]),\n",
       "             param_grid=[{&#x27;classifier&#x27;: [RandomForestClassifier(n_estimators=50,\n",
       "                                                                random_state=42),\n",
       "                                         MultinomialNB(),\n",
       "                                         ExtraTreesClassifier(n_estimators=50,\n",
       "                                                              random_state=42),\n",
       "                                         SVC(gamma=1.0, kernel=&#x27;sigmoid&#x27;,\n",
       "                                             random_state=42)],\n",
       "                          &#x27;vectorizer&#x27;: [TfidfVectorizer(max_features=1...\n",
       "                                         SVC(gamma=1.0, kernel=&#x27;sigmoid&#x27;,\n",
       "                                             random_state=42)],\n",
       "                          &#x27;vectorizer&#x27;: [TfidfVectorizer(max_features=2000)]},\n",
       "                         {&#x27;classifier&#x27;: [RandomForestClassifier(n_estimators=50,\n",
       "                                                                random_state=42),\n",
       "                                         MultinomialNB(),\n",
       "                                         ExtraTreesClassifier(n_estimators=50,\n",
       "                                                              random_state=42),\n",
       "                                         SVC(gamma=1.0, kernel=&#x27;sigmoid&#x27;,\n",
       "                                             random_state=42)],\n",
       "                          &#x27;vectorizer&#x27;: [TfidfVectorizer(max_features=3000)]}],\n",
       "             scoring=&#x27;precision&#x27;)</pre><b>In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook. <br />On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.</b></div><div class=\"sk-container\" hidden><div class=\"sk-item sk-dashed-wrapped\"><div class=\"sk-label-container\"><div class=\"sk-label sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-1\" type=\"checkbox\" ><label for=\"sk-estimator-id-1\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">GridSearchCV</label><div class=\"sk-toggleable__content\"><pre>GridSearchCV(cv=5,\n",
       "             estimator=Pipeline(memory=&#x27;cachedir&#x27;,\n",
       "                                steps=[(&#x27;vectorizer&#x27;, None),\n",
       "                                       (&#x27;classifier&#x27;, None)]),\n",
       "             param_grid=[{&#x27;classifier&#x27;: [RandomForestClassifier(n_estimators=50,\n",
       "                                                                random_state=42),\n",
       "                                         MultinomialNB(),\n",
       "                                         ExtraTreesClassifier(n_estimators=50,\n",
       "                                                              random_state=42),\n",
       "                                         SVC(gamma=1.0, kernel=&#x27;sigmoid&#x27;,\n",
       "                                             random_state=42)],\n",
       "                          &#x27;vectorizer&#x27;: [TfidfVectorizer(max_features=1...\n",
       "                                         SVC(gamma=1.0, kernel=&#x27;sigmoid&#x27;,\n",
       "                                             random_state=42)],\n",
       "                          &#x27;vectorizer&#x27;: [TfidfVectorizer(max_features=2000)]},\n",
       "                         {&#x27;classifier&#x27;: [RandomForestClassifier(n_estimators=50,\n",
       "                                                                random_state=42),\n",
       "                                         MultinomialNB(),\n",
       "                                         ExtraTreesClassifier(n_estimators=50,\n",
       "                                                              random_state=42),\n",
       "                                         SVC(gamma=1.0, kernel=&#x27;sigmoid&#x27;,\n",
       "                                             random_state=42)],\n",
       "                          &#x27;vectorizer&#x27;: [TfidfVectorizer(max_features=3000)]}],\n",
       "             scoring=&#x27;precision&#x27;)</pre></div></div></div><div class=\"sk-parallel\"><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-2\" type=\"checkbox\" ><label for=\"sk-estimator-id-2\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">estimator: Pipeline</label><div class=\"sk-toggleable__content\"><pre>Pipeline(memory=&#x27;cachedir&#x27;, steps=[(&#x27;vectorizer&#x27;, None), (&#x27;classifier&#x27;, None)])</pre></div></div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-3\" type=\"checkbox\" ><label for=\"sk-estimator-id-3\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">None</label><div class=\"sk-toggleable__content\"><pre>None</pre></div></div></div><div class=\"sk-item\"><div class=\"sk-estimator sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-4\" type=\"checkbox\" ><label for=\"sk-estimator-id-4\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">None</label><div class=\"sk-toggleable__content\"><pre>None</pre></div></div></div></div></div></div></div></div></div></div></div></div>"
      ],
      "text/plain": [
       "GridSearchCV(cv=5,\n",
       "             estimator=Pipeline(memory='cachedir',\n",
       "                                steps=[('vectorizer', None),\n",
       "                                       ('classifier', None)]),\n",
       "             param_grid=[{'classifier': [RandomForestClassifier(n_estimators=50,\n",
       "                                                                random_state=42),\n",
       "                                         MultinomialNB(),\n",
       "                                         ExtraTreesClassifier(n_estimators=50,\n",
       "                                                              random_state=42),\n",
       "                                         SVC(gamma=1.0, kernel='sigmoid',\n",
       "                                             random_state=42)],\n",
       "                          'vectorizer': [TfidfVectorizer(max_features=1...\n",
       "                                         SVC(gamma=1.0, kernel='sigmoid',\n",
       "                                             random_state=42)],\n",
       "                          'vectorizer': [TfidfVectorizer(max_features=2000)]},\n",
       "                         {'classifier': [RandomForestClassifier(n_estimators=50,\n",
       "                                                                random_state=42),\n",
       "                                         MultinomialNB(),\n",
       "                                         ExtraTreesClassifier(n_estimators=50,\n",
       "                                                              random_state=42),\n",
       "                                         SVC(gamma=1.0, kernel='sigmoid',\n",
       "                                             random_state=42)],\n",
       "                          'vectorizer': [TfidfVectorizer(max_features=3000)]}],\n",
       "             scoring='precision')"
      ]
     },
     "execution_count": 54,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train_vectorizer, X_test_vectorizer, y_train_vectorizer, y_test_vectorizer = train_test_split(\n",
    "    df['text_transformed'],\n",
    "    df['target'],\n",
    "    test_size=0.2,\n",
    "    random_state=RS\n",
    ")\n",
    "\n",
    "param_grid_vectorizer = [\n",
    "    {\n",
    "        'vectorizer': [TfidfVectorizer(max_features=1000)],\n",
    "        'classifier': list(clfs_final.values())\n",
    "    },\n",
    "    {\n",
    "        'vectorizer': [TfidfVectorizer(max_features=2000)],\n",
    "        'classifier':  list(clfs_final.values())\n",
    "    },\n",
    "    {\n",
    "        'vectorizer': [TfidfVectorizer(max_features=3000)],\n",
    "        'classifier':   list(clfs_final.values())\n",
    "    }\n",
    "]\n",
    "\n",
    "pipeline_vectorizer = Pipeline(\n",
    "    [\n",
    "        ('vectorizer', None),\n",
    "        ('classifier', None)\n",
    "    ],\n",
    "    memory='cachedir'\n",
    ")\n",
    "\n",
    "grid_search_vectorizer = GridSearchCV(\n",
    "    pipeline_vectorizer,\n",
    "    param_grid_vectorizer,\n",
    "    cv=5,\n",
    "    scoring='precision'\n",
    ")\n",
    "grid_search_vectorizer.fit(X_train_vectorizer, y_train_vectorizer)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "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>param_classifier</th>\n",
       "      <th>param_vectorizer</th>\n",
       "      <th>split0_test_score</th>\n",
       "      <th>split1_test_score</th>\n",
       "      <th>split2_test_score</th>\n",
       "      <th>split3_test_score</th>\n",
       "      <th>split4_test_score</th>\n",
       "      <th>mean_test_score</th>\n",
       "      <th>std_test_score</th>\n",
       "      <th>rank_test_score</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>MultinomialNB()</td>\n",
       "      <td>TfidfVectorizer(max_features=2000)</td>\n",
       "      <td>0.988235</td>\n",
       "      <td>0.986486</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.994944</td>\n",
       "      <td>0.006217</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>MultinomialNB()</td>\n",
       "      <td>TfidfVectorizer(max_features=3000)</td>\n",
       "      <td>0.987013</td>\n",
       "      <td>0.985507</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.994504</td>\n",
       "      <td>0.006748</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>(DecisionTreeClassifier(max_features='sqrt', r...</td>\n",
       "      <td>TfidfVectorizer(max_features=3000)</td>\n",
       "      <td>0.988506</td>\n",
       "      <td>0.974026</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.986111</td>\n",
       "      <td>0.989729</td>\n",
       "      <td>0.009718</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>(DecisionTreeClassifier(max_features='sqrt', r...</td>\n",
       "      <td>TfidfVectorizer(max_features=2000)</td>\n",
       "      <td>0.988235</td>\n",
       "      <td>0.987805</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.987500</td>\n",
       "      <td>0.972603</td>\n",
       "      <td>0.987229</td>\n",
       "      <td>0.008700</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>MultinomialNB()</td>\n",
       "      <td>TfidfVectorizer(max_features=1000)</td>\n",
       "      <td>0.988506</td>\n",
       "      <td>0.961538</td>\n",
       "      <td>0.978723</td>\n",
       "      <td>0.988372</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.983428</td>\n",
       "      <td>0.012854</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>(ExtraTreeClassifier(random_state=1608637542),...</td>\n",
       "      <td>TfidfVectorizer(max_features=2000)</td>\n",
       "      <td>0.977528</td>\n",
       "      <td>0.963415</td>\n",
       "      <td>0.977273</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.946667</td>\n",
       "      <td>0.972976</td>\n",
       "      <td>0.017624</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>SVC(gamma=1.0, kernel='sigmoid', random_state=42)</td>\n",
       "      <td>TfidfVectorizer(max_features=2000)</td>\n",
       "      <td>0.968421</td>\n",
       "      <td>0.975610</td>\n",
       "      <td>0.978495</td>\n",
       "      <td>0.955556</td>\n",
       "      <td>0.985714</td>\n",
       "      <td>0.972759</td>\n",
       "      <td>0.010234</td>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>(DecisionTreeClassifier(max_features='sqrt', r...</td>\n",
       "      <td>TfidfVectorizer(max_features=1000)</td>\n",
       "      <td>0.943820</td>\n",
       "      <td>0.987952</td>\n",
       "      <td>0.988636</td>\n",
       "      <td>0.975610</td>\n",
       "      <td>0.960000</td>\n",
       "      <td>0.971204</td>\n",
       "      <td>0.017200</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>(ExtraTreeClassifier(random_state=1608637542),...</td>\n",
       "      <td>TfidfVectorizer(max_features=3000)</td>\n",
       "      <td>0.988636</td>\n",
       "      <td>0.952941</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.976744</td>\n",
       "      <td>0.936709</td>\n",
       "      <td>0.971006</td>\n",
       "      <td>0.023182</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>SVC(gamma=1.0, kernel='sigmoid', random_state=42)</td>\n",
       "      <td>TfidfVectorizer(max_features=3000)</td>\n",
       "      <td>0.958333</td>\n",
       "      <td>0.976190</td>\n",
       "      <td>0.989130</td>\n",
       "      <td>0.956522</td>\n",
       "      <td>0.973333</td>\n",
       "      <td>0.970702</td>\n",
       "      <td>0.012089</td>\n",
       "      <td>10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>SVC(gamma=1.0, kernel='sigmoid', random_state=42)</td>\n",
       "      <td>TfidfVectorizer(max_features=1000)</td>\n",
       "      <td>0.977778</td>\n",
       "      <td>0.951807</td>\n",
       "      <td>0.967742</td>\n",
       "      <td>0.955556</td>\n",
       "      <td>0.972222</td>\n",
       "      <td>0.965021</td>\n",
       "      <td>0.009861</td>\n",
       "      <td>11</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>(ExtraTreeClassifier(random_state=1608637542),...</td>\n",
       "      <td>TfidfVectorizer(max_features=1000)</td>\n",
       "      <td>0.956044</td>\n",
       "      <td>0.943182</td>\n",
       "      <td>0.978261</td>\n",
       "      <td>0.977778</td>\n",
       "      <td>0.936709</td>\n",
       "      <td>0.958395</td>\n",
       "      <td>0.017191</td>\n",
       "      <td>12</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                     param_classifier  \\\n",
       "5                                     MultinomialNB()   \n",
       "9                                     MultinomialNB()   \n",
       "8   (DecisionTreeClassifier(max_features='sqrt', r...   \n",
       "4   (DecisionTreeClassifier(max_features='sqrt', r...   \n",
       "1                                     MultinomialNB()   \n",
       "6   (ExtraTreeClassifier(random_state=1608637542),...   \n",
       "7   SVC(gamma=1.0, kernel='sigmoid', random_state=42)   \n",
       "0   (DecisionTreeClassifier(max_features='sqrt', r...   \n",
       "10  (ExtraTreeClassifier(random_state=1608637542),...   \n",
       "11  SVC(gamma=1.0, kernel='sigmoid', random_state=42)   \n",
       "3   SVC(gamma=1.0, kernel='sigmoid', random_state=42)   \n",
       "2   (ExtraTreeClassifier(random_state=1608637542),...   \n",
       "\n",
       "                      param_vectorizer  split0_test_score  split1_test_score  \\\n",
       "5   TfidfVectorizer(max_features=2000)           0.988235           0.986486   \n",
       "9   TfidfVectorizer(max_features=3000)           0.987013           0.985507   \n",
       "8   TfidfVectorizer(max_features=3000)           0.988506           0.974026   \n",
       "4   TfidfVectorizer(max_features=2000)           0.988235           0.987805   \n",
       "1   TfidfVectorizer(max_features=1000)           0.988506           0.961538   \n",
       "6   TfidfVectorizer(max_features=2000)           0.977528           0.963415   \n",
       "7   TfidfVectorizer(max_features=2000)           0.968421           0.975610   \n",
       "0   TfidfVectorizer(max_features=1000)           0.943820           0.987952   \n",
       "10  TfidfVectorizer(max_features=3000)           0.988636           0.952941   \n",
       "11  TfidfVectorizer(max_features=3000)           0.958333           0.976190   \n",
       "3   TfidfVectorizer(max_features=1000)           0.977778           0.951807   \n",
       "2   TfidfVectorizer(max_features=1000)           0.956044           0.943182   \n",
       "\n",
       "    split2_test_score  split3_test_score  split4_test_score  mean_test_score  \\\n",
       "5            1.000000           1.000000           1.000000         0.994944   \n",
       "9            1.000000           1.000000           1.000000         0.994504   \n",
       "8            1.000000           1.000000           0.986111         0.989729   \n",
       "4            1.000000           0.987500           0.972603         0.987229   \n",
       "1            0.978723           0.988372           1.000000         0.983428   \n",
       "6            0.977273           1.000000           0.946667         0.972976   \n",
       "7            0.978495           0.955556           0.985714         0.972759   \n",
       "0            0.988636           0.975610           0.960000         0.971204   \n",
       "10           1.000000           0.976744           0.936709         0.971006   \n",
       "11           0.989130           0.956522           0.973333         0.970702   \n",
       "3            0.967742           0.955556           0.972222         0.965021   \n",
       "2            0.978261           0.977778           0.936709         0.958395   \n",
       "\n",
       "    std_test_score  rank_test_score  \n",
       "5         0.006217                1  \n",
       "9         0.006748                2  \n",
       "8         0.009718                3  \n",
       "4         0.008700                4  \n",
       "1         0.012854                5  \n",
       "6         0.017624                6  \n",
       "7         0.010234                7  \n",
       "0         0.017200                8  \n",
       "10        0.023182                9  \n",
       "11        0.012089               10  \n",
       "3         0.009861               11  \n",
       "2         0.017191               12  "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "results_vectorizer = pd.DataFrame(grid_search_vectorizer.cv_results_)\n",
    "results_vectorizer = results_vectorizer.sort_values('rank_test_score')\n",
    "results_short_vectorizer = results_vectorizer.drop(\n",
    "    columns=[\n",
    "        'mean_fit_time', 'std_fit_time', 'mean_score_time', 'std_score_time', 'params'\n",
    "    ]\n",
    ")\n",
    "display(results_short_vectorizer)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [],
   "source": [
    "vectorizer_final = TfidfVectorizer(max_features=2000)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Additional features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_standard_scores():\n",
    "    X_train, X_test, y_train, y_test = train_test_split(\n",
    "        df['text_transformed'],\n",
    "        df['target'],\n",
    "        test_size=0.2,\n",
    "        random_state=RS\n",
    "    )\n",
    "\n",
    "    vectorizer_final.fit(X_train)\n",
    "\n",
    "    X_train = vectorizer_final.transform(\n",
    "        X_train\n",
    "    ).toarray()  # type: ignore\n",
    "    X_test = vectorizer_final.transform(\n",
    "        X_test\n",
    "    ).toarray()  # type: ignore\n",
    "\n",
    "    scores = {}\n",
    "    for name, clf in clfs_final.items():\n",
    "        clf.fit(X_train, y_train)\n",
    "        scores[name] = calc_scores(clf, X_test, y_test)\n",
    "\n",
    "    scores_df = create_perf_df(scores)\n",
    "    return scores_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [],
   "source": [
    "standard_perf_df = get_standard_scores()\n",
    "standard_perf_df = standard_perf_df.sort_index()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Check text augmentation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_text_augmented_scores():\n",
    "    X_train, X_test, y_train, y_test = train_test_split(\n",
    "        df['text_augmented'],\n",
    "        df['target'],\n",
    "        test_size=0.2,\n",
    "        random_state=RS\n",
    "    )\n",
    "\n",
    "    vectorizer_final.fit(X_train)\n",
    "\n",
    "    X_train = vectorizer_final.transform(\n",
    "        X_train\n",
    "    ).toarray()  # type: ignore\n",
    "    X_test = vectorizer_final.transform(\n",
    "        X_test\n",
    "    ).toarray()  # type: ignore\n",
    "\n",
    "    scores = {}\n",
    "    for name, clf in clfs_final.items():\n",
    "        clf.fit(X_train, y_train)\n",
    "        scores[name] = calc_scores(clf, X_test, y_test)\n",
    "\n",
    "    scores_df = create_perf_df(scores)\n",
    "    return scores_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "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>accuracy</th>\n",
       "      <th>precision</th>\n",
       "      <th>recall</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>ETC</th>\n",
       "      <td>0.976789</td>\n",
       "      <td>0.961832</td>\n",
       "      <td>0.868966</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MNB</th>\n",
       "      <td>0.973888</td>\n",
       "      <td>0.983607</td>\n",
       "      <td>0.827586</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>RFC</th>\n",
       "      <td>0.970986</td>\n",
       "      <td>0.991453</td>\n",
       "      <td>0.8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SVC</th>\n",
       "      <td>0.978723</td>\n",
       "      <td>0.955556</td>\n",
       "      <td>0.889655</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     accuracy precision    recall\n",
       "ETC  0.976789  0.961832  0.868966\n",
       "MNB  0.973888  0.983607  0.827586\n",
       "RFC  0.970986  0.991453       0.8\n",
       "SVC  0.978723  0.955556  0.889655"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "display(standard_perf_df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "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>accuracy</th>\n",
       "      <th>precision</th>\n",
       "      <th>recall</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>ETC</th>\n",
       "      <td>0.981625</td>\n",
       "      <td>0.977273</td>\n",
       "      <td>0.889655</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MNB</th>\n",
       "      <td>0.972921</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.806897</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>RFC</th>\n",
       "      <td>0.976789</td>\n",
       "      <td>0.984</td>\n",
       "      <td>0.848276</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SVC</th>\n",
       "      <td>0.98646</td>\n",
       "      <td>0.985185</td>\n",
       "      <td>0.917241</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     accuracy precision    recall\n",
       "ETC  0.981625  0.977273  0.889655\n",
       "MNB  0.972921       1.0  0.806897\n",
       "RFC  0.976789     0.984  0.848276\n",
       "SVC   0.98646  0.985185  0.917241"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "text_augmented_perf_df = get_text_augmented_scores()\n",
    "text_augmented_perf_df = text_augmented_perf_df.sort_index()\n",
    "\n",
    "display(text_augmented_perf_df)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Check characters count"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_character_count_scores():\n",
    "    X_train, X_test, y_train, y_test = train_test_split(\n",
    "        np.column_stack((\n",
    "            df['text_transformed'],\n",
    "            df['characters_count_scaled'],\n",
    "        )),\n",
    "        df['target'],\n",
    "        test_size=0.2,\n",
    "        random_state=RS\n",
    "    )\n",
    "\n",
    "    vectorizer_final.fit(X_train[:, 0])\n",
    "\n",
    "    X_train_a = vectorizer_final.transform(\n",
    "        X_train[:, 0]).toarray()  # type: ignore\n",
    "    X_test_a = vectorizer_final.transform(\n",
    "        X_test[:, 0]).toarray()  # type: ignore\n",
    "    X_train_b = X_train[:, 1:2].reshape(-1, 1)\n",
    "    X_test_b = X_test[:, 1:2].reshape(-1, 1)\n",
    "\n",
    "    X_train = np.concatenate((X_train_a, X_train_b), axis=1)\n",
    "    X_test = np.concatenate((X_test_a, X_test_b), axis=1)\n",
    "\n",
    "    scores = {}\n",
    "    for name, clf in clfs_final.items():\n",
    "        clf.fit(X_train, y_train)\n",
    "        scores[name] = calc_scores(clf, X_test, y_test)\n",
    "\n",
    "    scores_df = create_perf_df(scores)\n",
    "    return scores_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "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>accuracy</th>\n",
       "      <th>precision</th>\n",
       "      <th>recall</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>ETC</th>\n",
       "      <td>0.976789</td>\n",
       "      <td>0.961832</td>\n",
       "      <td>0.868966</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MNB</th>\n",
       "      <td>0.973888</td>\n",
       "      <td>0.983607</td>\n",
       "      <td>0.827586</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>RFC</th>\n",
       "      <td>0.970986</td>\n",
       "      <td>0.991453</td>\n",
       "      <td>0.8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SVC</th>\n",
       "      <td>0.978723</td>\n",
       "      <td>0.955556</td>\n",
       "      <td>0.889655</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     accuracy precision    recall\n",
       "ETC  0.976789  0.961832  0.868966\n",
       "MNB  0.973888  0.983607  0.827586\n",
       "RFC  0.970986  0.991453       0.8\n",
       "SVC  0.978723  0.955556  0.889655"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "display(standard_perf_df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "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>accuracy</th>\n",
       "      <th>precision</th>\n",
       "      <th>recall</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>ETC</th>\n",
       "      <td>0.979691</td>\n",
       "      <td>0.962687</td>\n",
       "      <td>0.889655</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MNB</th>\n",
       "      <td>0.973888</td>\n",
       "      <td>0.983607</td>\n",
       "      <td>0.827586</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>RFC</th>\n",
       "      <td>0.972921</td>\n",
       "      <td>0.983471</td>\n",
       "      <td>0.82069</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SVC</th>\n",
       "      <td>0.979691</td>\n",
       "      <td>0.955882</td>\n",
       "      <td>0.896552</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     accuracy precision    recall\n",
       "ETC  0.979691  0.962687  0.889655\n",
       "MNB  0.973888  0.983607  0.827586\n",
       "RFC  0.972921  0.983471   0.82069\n",
       "SVC  0.979691  0.955882  0.896552"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "character_count_perf_df = get_character_count_scores()\n",
    "character_count_perf_df = character_count_perf_df.sort_index()\n",
    "\n",
    "display(character_count_perf_df)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Models hyperparameters tuning\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import RandomizedSearchCV"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 112,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train_final, X_test_final, y_train_final, y_test_final = train_test_split(\n",
    "    df['text_augmented'],\n",
    "    df['target'],\n",
    "    test_size=0.2,\n",
    "    random_state=RS\n",
    ")\n",
    "\n",
    "vectorizer_final.fit(X_train_final)\n",
    "\n",
    "X_train_final = vectorizer_final.transform(X_train_final)\n",
    "X_test_final = vectorizer_final.transform(X_test_final)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 99,
   "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>mean_fit_time</th>\n",
       "      <th>std_fit_time</th>\n",
       "      <th>mean_score_time</th>\n",
       "      <th>std_score_time</th>\n",
       "      <th>param_alpha</th>\n",
       "      <th>param_fit_prior</th>\n",
       "      <th>params</th>\n",
       "      <th>split0_test_score</th>\n",
       "      <th>split1_test_score</th>\n",
       "      <th>split2_test_score</th>\n",
       "      <th>split3_test_score</th>\n",
       "      <th>split4_test_score</th>\n",
       "      <th>mean_test_score</th>\n",
       "      <th>std_test_score</th>\n",
       "      <th>rank_test_score</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>0.001090</td>\n",
       "      <td>0.000024</td>\n",
       "      <td>0.000824</td>\n",
       "      <td>0.000005</td>\n",
       "      <td>0.12</td>\n",
       "      <td>True</td>\n",
       "      <td>{'alpha': 0.12, 'fit_prior': True}</td>\n",
       "      <td>0.927083</td>\n",
       "      <td>0.869565</td>\n",
       "      <td>0.949495</td>\n",
       "      <td>0.938144</td>\n",
       "      <td>0.892473</td>\n",
       "      <td>0.915352</td>\n",
       "      <td>0.029808</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28</th>\n",
       "      <td>0.001068</td>\n",
       "      <td>0.000016</td>\n",
       "      <td>0.000828</td>\n",
       "      <td>0.000004</td>\n",
       "      <td>0.15</td>\n",
       "      <td>True</td>\n",
       "      <td>{'alpha': 0.15000000000000002, 'fit_prior': True}</td>\n",
       "      <td>0.927083</td>\n",
       "      <td>0.868132</td>\n",
       "      <td>0.944162</td>\n",
       "      <td>0.943005</td>\n",
       "      <td>0.892473</td>\n",
       "      <td>0.914971</td>\n",
       "      <td>0.029949</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>34</th>\n",
       "      <td>0.001071</td>\n",
       "      <td>0.000012</td>\n",
       "      <td>0.000843</td>\n",
       "      <td>0.000037</td>\n",
       "      <td>0.18</td>\n",
       "      <td>True</td>\n",
       "      <td>{'alpha': 0.18000000000000002, 'fit_prior': True}</td>\n",
       "      <td>0.927083</td>\n",
       "      <td>0.868132</td>\n",
       "      <td>0.944162</td>\n",
       "      <td>0.943005</td>\n",
       "      <td>0.892473</td>\n",
       "      <td>0.914971</td>\n",
       "      <td>0.029949</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>36</th>\n",
       "      <td>0.001078</td>\n",
       "      <td>0.000016</td>\n",
       "      <td>0.000835</td>\n",
       "      <td>0.000018</td>\n",
       "      <td>0.19</td>\n",
       "      <td>True</td>\n",
       "      <td>{'alpha': 0.19, 'fit_prior': True}</td>\n",
       "      <td>0.927083</td>\n",
       "      <td>0.868132</td>\n",
       "      <td>0.944162</td>\n",
       "      <td>0.943005</td>\n",
       "      <td>0.892473</td>\n",
       "      <td>0.914971</td>\n",
       "      <td>0.029949</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26</th>\n",
       "      <td>0.001081</td>\n",
       "      <td>0.000020</td>\n",
       "      <td>0.000826</td>\n",
       "      <td>0.000016</td>\n",
       "      <td>0.14</td>\n",
       "      <td>True</td>\n",
       "      <td>{'alpha': 0.14, 'fit_prior': True}</td>\n",
       "      <td>0.927083</td>\n",
       "      <td>0.868132</td>\n",
       "      <td>0.949495</td>\n",
       "      <td>0.937500</td>\n",
       "      <td>0.892473</td>\n",
       "      <td>0.914937</td>\n",
       "      <td>0.030154</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    mean_fit_time  std_fit_time  mean_score_time  std_score_time param_alpha  \\\n",
       "22       0.001090      0.000024         0.000824        0.000005        0.12   \n",
       "28       0.001068      0.000016         0.000828        0.000004        0.15   \n",
       "34       0.001071      0.000012         0.000843        0.000037        0.18   \n",
       "36       0.001078      0.000016         0.000835        0.000018        0.19   \n",
       "26       0.001081      0.000020         0.000826        0.000016        0.14   \n",
       "\n",
       "   param_fit_prior                                             params  \\\n",
       "22            True                 {'alpha': 0.12, 'fit_prior': True}   \n",
       "28            True  {'alpha': 0.15000000000000002, 'fit_prior': True}   \n",
       "34            True  {'alpha': 0.18000000000000002, 'fit_prior': True}   \n",
       "36            True                 {'alpha': 0.19, 'fit_prior': True}   \n",
       "26            True                 {'alpha': 0.14, 'fit_prior': True}   \n",
       "\n",
       "    split0_test_score  split1_test_score  split2_test_score  \\\n",
       "22           0.927083           0.869565           0.949495   \n",
       "28           0.927083           0.868132           0.944162   \n",
       "34           0.927083           0.868132           0.944162   \n",
       "36           0.927083           0.868132           0.944162   \n",
       "26           0.927083           0.868132           0.949495   \n",
       "\n",
       "    split3_test_score  split4_test_score  mean_test_score  std_test_score  \\\n",
       "22           0.938144           0.892473         0.915352        0.029808   \n",
       "28           0.943005           0.892473         0.914971        0.029949   \n",
       "34           0.943005           0.892473         0.914971        0.029949   \n",
       "36           0.943005           0.892473         0.914971        0.029949   \n",
       "26           0.937500           0.892473         0.914937        0.030154   \n",
       "\n",
       "    rank_test_score  \n",
       "22                1  \n",
       "28                2  \n",
       "34                2  \n",
       "36                2  \n",
       "26                5  "
      ]
     },
     "execution_count": 99,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "param_grid_mnb = {\n",
    "    'alpha': np.linspace(0.01, 1.0, 100),\n",
    "    'fit_prior': [True, False],\n",
    "}\n",
    "\n",
    "param_search_mnb = GridSearchCV(\n",
    "    MultinomialNB(),\n",
    "    param_grid_mnb,\n",
    "    cv=5,\n",
    "    scoring='f1',\n",
    ")\n",
    "param_search_mnb.fit(X_train_final, y_train_final)\n",
    "\n",
    "results_mnb = pd.DataFrame(param_search_mnb.cv_results_)\n",
    "results_mnb = results_mnb.sort_values('rank_test_score')\n",
    "results_mnb.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 101,
   "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>accuracy</th>\n",
       "      <th>precision</th>\n",
       "      <th>recall</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>ETC</th>\n",
       "      <td>0.981625</td>\n",
       "      <td>0.977273</td>\n",
       "      <td>0.889655</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MNB</th>\n",
       "      <td>0.972921</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.806897</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>RFC</th>\n",
       "      <td>0.976789</td>\n",
       "      <td>0.984</td>\n",
       "      <td>0.848276</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SVC</th>\n",
       "      <td>0.98646</td>\n",
       "      <td>0.985185</td>\n",
       "      <td>0.917241</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     accuracy precision    recall\n",
       "ETC  0.981625  0.977273  0.889655\n",
       "MNB  0.972921       1.0  0.806897\n",
       "RFC  0.976789     0.984  0.848276\n",
       "SVC   0.98646  0.985185  0.917241"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "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>accuracy</th>\n",
       "      <th>precision</th>\n",
       "      <th>recall</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>MNB</th>\n",
       "      <td>0.982592</td>\n",
       "      <td>0.992248</td>\n",
       "      <td>0.882759</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     accuracy precision    recall\n",
       "MNB  0.982592  0.992248  0.882759"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "mnb_final = MultinomialNB(**param_search_mnb.best_params_)\n",
    "mnb_final.fit(X_train_final, y_train_final)\n",
    "\n",
    "mnb_scores = calc_scores(mnb_final, X_test_final, y_test_final)\n",
    "mnb_perf_df = create_perf_df({'MNB': mnb_scores})\n",
    "\n",
    "display(text_augmented_perf_df)\n",
    "display(mnb_perf_df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "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>mean_fit_time</th>\n",
       "      <th>std_fit_time</th>\n",
       "      <th>mean_score_time</th>\n",
       "      <th>std_score_time</th>\n",
       "      <th>param_n_estimators</th>\n",
       "      <th>param_min_samples_split</th>\n",
       "      <th>param_min_samples_leaf</th>\n",
       "      <th>param_max_depth</th>\n",
       "      <th>param_bootstrap</th>\n",
       "      <th>params</th>\n",
       "      <th>split0_test_score</th>\n",
       "      <th>split1_test_score</th>\n",
       "      <th>split2_test_score</th>\n",
       "      <th>split3_test_score</th>\n",
       "      <th>split4_test_score</th>\n",
       "      <th>mean_test_score</th>\n",
       "      <th>std_test_score</th>\n",
       "      <th>rank_test_score</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>1.015743</td>\n",
       "      <td>0.043097</td>\n",
       "      <td>0.050976</td>\n",
       "      <td>0.000974</td>\n",
       "      <td>200</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>None</td>\n",
       "      <td>False</td>\n",
       "      <td>{'n_estimators': 200, 'min_samples_split': 2, ...</td>\n",
       "      <td>0.905263</td>\n",
       "      <td>0.891304</td>\n",
       "      <td>0.903226</td>\n",
       "      <td>0.898396</td>\n",
       "      <td>0.860215</td>\n",
       "      <td>0.891681</td>\n",
       "      <td>0.016450</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>0.621583</td>\n",
       "      <td>0.028852</td>\n",
       "      <td>0.034171</td>\n",
       "      <td>0.000076</td>\n",
       "      <td>200</td>\n",
       "      <td>5</td>\n",
       "      <td>1</td>\n",
       "      <td>75</td>\n",
       "      <td>False</td>\n",
       "      <td>{'n_estimators': 200, 'min_samples_split': 5, ...</td>\n",
       "      <td>0.916667</td>\n",
       "      <td>0.872928</td>\n",
       "      <td>0.903226</td>\n",
       "      <td>0.892473</td>\n",
       "      <td>0.854054</td>\n",
       "      <td>0.887870</td>\n",
       "      <td>0.022150</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28</th>\n",
       "      <td>1.512301</td>\n",
       "      <td>0.103609</td>\n",
       "      <td>0.086143</td>\n",
       "      <td>0.002636</td>\n",
       "      <td>400</td>\n",
       "      <td>5</td>\n",
       "      <td>1</td>\n",
       "      <td>None</td>\n",
       "      <td>True</td>\n",
       "      <td>{'n_estimators': 400, 'min_samples_split': 5, ...</td>\n",
       "      <td>0.893617</td>\n",
       "      <td>0.890110</td>\n",
       "      <td>0.897297</td>\n",
       "      <td>0.885246</td>\n",
       "      <td>0.828729</td>\n",
       "      <td>0.879000</td>\n",
       "      <td>0.025448</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>0.789551</td>\n",
       "      <td>0.048218</td>\n",
       "      <td>0.043792</td>\n",
       "      <td>0.001031</td>\n",
       "      <td>200</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>None</td>\n",
       "      <td>True</td>\n",
       "      <td>{'n_estimators': 200, 'min_samples_split': 2, ...</td>\n",
       "      <td>0.887701</td>\n",
       "      <td>0.877778</td>\n",
       "      <td>0.897297</td>\n",
       "      <td>0.885246</td>\n",
       "      <td>0.828729</td>\n",
       "      <td>0.875350</td>\n",
       "      <td>0.024131</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>0.187136</td>\n",
       "      <td>0.012614</td>\n",
       "      <td>0.011891</td>\n",
       "      <td>0.000427</td>\n",
       "      <td>50</td>\n",
       "      <td>5</td>\n",
       "      <td>1</td>\n",
       "      <td>None</td>\n",
       "      <td>True</td>\n",
       "      <td>{'n_estimators': 50, 'min_samples_split': 5, '...</td>\n",
       "      <td>0.893617</td>\n",
       "      <td>0.860335</td>\n",
       "      <td>0.897297</td>\n",
       "      <td>0.879121</td>\n",
       "      <td>0.828729</td>\n",
       "      <td>0.871820</td>\n",
       "      <td>0.025164</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    mean_fit_time  std_fit_time  mean_score_time  std_score_time  \\\n",
       "14       1.015743      0.043097         0.050976        0.000974   \n",
       "16       0.621583      0.028852         0.034171        0.000076   \n",
       "28       1.512301      0.103609         0.086143        0.002636   \n",
       "25       0.789551      0.048218         0.043792        0.001031   \n",
       "10       0.187136      0.012614         0.011891        0.000427   \n",
       "\n",
       "   param_n_estimators param_min_samples_split param_min_samples_leaf  \\\n",
       "14                200                       2                      1   \n",
       "16                200                       5                      1   \n",
       "28                400                       5                      1   \n",
       "25                200                       2                      1   \n",
       "10                 50                       5                      1   \n",
       "\n",
       "   param_max_depth param_bootstrap  \\\n",
       "14            None           False   \n",
       "16              75           False   \n",
       "28            None            True   \n",
       "25            None            True   \n",
       "10            None            True   \n",
       "\n",
       "                                               params  split0_test_score  \\\n",
       "14  {'n_estimators': 200, 'min_samples_split': 2, ...           0.905263   \n",
       "16  {'n_estimators': 200, 'min_samples_split': 5, ...           0.916667   \n",
       "28  {'n_estimators': 400, 'min_samples_split': 5, ...           0.893617   \n",
       "25  {'n_estimators': 200, 'min_samples_split': 2, ...           0.887701   \n",
       "10  {'n_estimators': 50, 'min_samples_split': 5, '...           0.893617   \n",
       "\n",
       "    split1_test_score  split2_test_score  split3_test_score  \\\n",
       "14           0.891304           0.903226           0.898396   \n",
       "16           0.872928           0.903226           0.892473   \n",
       "28           0.890110           0.897297           0.885246   \n",
       "25           0.877778           0.897297           0.885246   \n",
       "10           0.860335           0.897297           0.879121   \n",
       "\n",
       "    split4_test_score  mean_test_score  std_test_score  rank_test_score  \n",
       "14           0.860215         0.891681        0.016450                1  \n",
       "16           0.854054         0.887870        0.022150                2  \n",
       "28           0.828729         0.879000        0.025448                3  \n",
       "25           0.828729         0.875350        0.024131                4  \n",
       "10           0.828729         0.871820        0.025164                5  "
      ]
     },
     "execution_count": 82,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "param_grid_etc = {\n",
    "    'n_estimators': [50, 100, 200, 300, 400, 500],\n",
    "    'max_depth': [None, 10, 25, 50, 75],\n",
    "    'min_samples_split': [2, 5, 10],\n",
    "    'min_samples_leaf': [1, 2, 4],\n",
    "    'bootstrap': [True, False]\n",
    "}\n",
    "\n",
    "param_search_etc = RandomizedSearchCV(\n",
    "    ExtraTreesClassifier(random_state=RS),\n",
    "    param_grid_etc,\n",
    "    cv=5,\n",
    "    scoring='f1',\n",
    "    n_iter=30,\n",
    "    random_state=RS\n",
    ")\n",
    "param_search_etc.fit(X_train_final, y_train_final)\n",
    "\n",
    "results_etc = pd.DataFrame(param_search_etc.cv_results_)\n",
    "results_etc = results_etc.sort_values('rank_test_score')\n",
    "results_etc.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "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>accuracy</th>\n",
       "      <th>precision</th>\n",
       "      <th>recall</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>ETC</th>\n",
       "      <td>0.981625</td>\n",
       "      <td>0.977273</td>\n",
       "      <td>0.889655</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MNB</th>\n",
       "      <td>0.972921</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.806897</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>RFC</th>\n",
       "      <td>0.976789</td>\n",
       "      <td>0.984</td>\n",
       "      <td>0.848276</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SVC</th>\n",
       "      <td>0.98646</td>\n",
       "      <td>0.985185</td>\n",
       "      <td>0.917241</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     accuracy precision    recall\n",
       "ETC  0.981625  0.977273  0.889655\n",
       "MNB  0.972921       1.0  0.806897\n",
       "RFC  0.976789     0.984  0.848276\n",
       "SVC   0.98646  0.985185  0.917241"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "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>accuracy</th>\n",
       "      <th>precision</th>\n",
       "      <th>recall</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>ETC</th>\n",
       "      <td>0.984526</td>\n",
       "      <td>0.984962</td>\n",
       "      <td>0.903448</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     accuracy precision    recall\n",
       "ETC  0.984526  0.984962  0.903448"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "etc_final = ExtraTreesClassifier(\n",
    "    **param_search_etc.best_params_,\n",
    "    random_state=RS\n",
    ")\n",
    "etc_final.fit(X_train_final, y_train_final)\n",
    "\n",
    "etc_scores = calc_scores(etc_final, X_test_final, y_test_final)\n",
    "etc_perf_df = create_perf_df({'ETC': etc_scores})\n",
    "\n",
    "display(text_augmented_perf_df)\n",
    "display(etc_perf_df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "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>mean_fit_time</th>\n",
       "      <th>std_fit_time</th>\n",
       "      <th>mean_score_time</th>\n",
       "      <th>std_score_time</th>\n",
       "      <th>param_n_estimators</th>\n",
       "      <th>param_min_samples_split</th>\n",
       "      <th>param_min_samples_leaf</th>\n",
       "      <th>param_max_depth</th>\n",
       "      <th>param_bootstrap</th>\n",
       "      <th>params</th>\n",
       "      <th>split0_test_score</th>\n",
       "      <th>split1_test_score</th>\n",
       "      <th>split2_test_score</th>\n",
       "      <th>split3_test_score</th>\n",
       "      <th>split4_test_score</th>\n",
       "      <th>mean_test_score</th>\n",
       "      <th>std_test_score</th>\n",
       "      <th>rank_test_score</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>1.022487</td>\n",
       "      <td>0.038973</td>\n",
       "      <td>0.034093</td>\n",
       "      <td>0.000537</td>\n",
       "      <td>200</td>\n",
       "      <td>5</td>\n",
       "      <td>1</td>\n",
       "      <td>75</td>\n",
       "      <td>False</td>\n",
       "      <td>{'n_estimators': 200, 'min_samples_split': 5, ...</td>\n",
       "      <td>0.910053</td>\n",
       "      <td>0.902174</td>\n",
       "      <td>0.903226</td>\n",
       "      <td>0.892473</td>\n",
       "      <td>0.836957</td>\n",
       "      <td>0.888976</td>\n",
       "      <td>0.026607</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>1.266701</td>\n",
       "      <td>0.038599</td>\n",
       "      <td>0.039853</td>\n",
       "      <td>0.000837</td>\n",
       "      <td>200</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>None</td>\n",
       "      <td>False</td>\n",
       "      <td>{'n_estimators': 200, 'min_samples_split': 2, ...</td>\n",
       "      <td>0.905263</td>\n",
       "      <td>0.890110</td>\n",
       "      <td>0.897297</td>\n",
       "      <td>0.886486</td>\n",
       "      <td>0.836957</td>\n",
       "      <td>0.883223</td>\n",
       "      <td>0.024011</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.370434</td>\n",
       "      <td>0.012593</td>\n",
       "      <td>0.016190</td>\n",
       "      <td>0.000145</td>\n",
       "      <td>100</td>\n",
       "      <td>10</td>\n",
       "      <td>1</td>\n",
       "      <td>75</td>\n",
       "      <td>True</td>\n",
       "      <td>{'n_estimators': 100, 'min_samples_split': 10,...</td>\n",
       "      <td>0.875676</td>\n",
       "      <td>0.877778</td>\n",
       "      <td>0.891304</td>\n",
       "      <td>0.874317</td>\n",
       "      <td>0.822222</td>\n",
       "      <td>0.868259</td>\n",
       "      <td>0.023802</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28</th>\n",
       "      <td>1.698513</td>\n",
       "      <td>0.067805</td>\n",
       "      <td>0.071079</td>\n",
       "      <td>0.002490</td>\n",
       "      <td>400</td>\n",
       "      <td>5</td>\n",
       "      <td>1</td>\n",
       "      <td>None</td>\n",
       "      <td>True</td>\n",
       "      <td>{'n_estimators': 400, 'min_samples_split': 5, ...</td>\n",
       "      <td>0.880435</td>\n",
       "      <td>0.871508</td>\n",
       "      <td>0.885246</td>\n",
       "      <td>0.866667</td>\n",
       "      <td>0.826816</td>\n",
       "      <td>0.866134</td>\n",
       "      <td>0.020712</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>0.210764</td>\n",
       "      <td>0.011200</td>\n",
       "      <td>0.009689</td>\n",
       "      <td>0.000227</td>\n",
       "      <td>50</td>\n",
       "      <td>5</td>\n",
       "      <td>1</td>\n",
       "      <td>None</td>\n",
       "      <td>True</td>\n",
       "      <td>{'n_estimators': 50, 'min_samples_split': 5, '...</td>\n",
       "      <td>0.877005</td>\n",
       "      <td>0.872928</td>\n",
       "      <td>0.886486</td>\n",
       "      <td>0.866667</td>\n",
       "      <td>0.820225</td>\n",
       "      <td>0.864662</td>\n",
       "      <td>0.023133</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    mean_fit_time  std_fit_time  mean_score_time  std_score_time  \\\n",
       "16       1.022487      0.038973         0.034093        0.000537   \n",
       "14       1.266701      0.038599         0.039853        0.000837   \n",
       "0        0.370434      0.012593         0.016190        0.000145   \n",
       "28       1.698513      0.067805         0.071079        0.002490   \n",
       "10       0.210764      0.011200         0.009689        0.000227   \n",
       "\n",
       "   param_n_estimators param_min_samples_split param_min_samples_leaf  \\\n",
       "16                200                       5                      1   \n",
       "14                200                       2                      1   \n",
       "0                 100                      10                      1   \n",
       "28                400                       5                      1   \n",
       "10                 50                       5                      1   \n",
       "\n",
       "   param_max_depth param_bootstrap  \\\n",
       "16              75           False   \n",
       "14            None           False   \n",
       "0               75            True   \n",
       "28            None            True   \n",
       "10            None            True   \n",
       "\n",
       "                                               params  split0_test_score  \\\n",
       "16  {'n_estimators': 200, 'min_samples_split': 5, ...           0.910053   \n",
       "14  {'n_estimators': 200, 'min_samples_split': 2, ...           0.905263   \n",
       "0   {'n_estimators': 100, 'min_samples_split': 10,...           0.875676   \n",
       "28  {'n_estimators': 400, 'min_samples_split': 5, ...           0.880435   \n",
       "10  {'n_estimators': 50, 'min_samples_split': 5, '...           0.877005   \n",
       "\n",
       "    split1_test_score  split2_test_score  split3_test_score  \\\n",
       "16           0.902174           0.903226           0.892473   \n",
       "14           0.890110           0.897297           0.886486   \n",
       "0            0.877778           0.891304           0.874317   \n",
       "28           0.871508           0.885246           0.866667   \n",
       "10           0.872928           0.886486           0.866667   \n",
       "\n",
       "    split4_test_score  mean_test_score  std_test_score  rank_test_score  \n",
       "16           0.836957         0.888976        0.026607                1  \n",
       "14           0.836957         0.883223        0.024011                2  \n",
       "0            0.822222         0.868259        0.023802                3  \n",
       "28           0.826816         0.866134        0.020712                4  \n",
       "10           0.820225         0.864662        0.023133                5  "
      ]
     },
     "execution_count": 80,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "param_grid_rfc = {\n",
    "    'n_estimators': [50, 100, 200, 300, 400, 500],\n",
    "    'max_depth': [None, 10, 25, 50, 75],\n",
    "    'min_samples_split': [2, 5, 10],\n",
    "    'min_samples_leaf': [1, 2, 4],\n",
    "    'bootstrap': [True, False]\n",
    "}\n",
    "\n",
    "param_search_rfc = RandomizedSearchCV(\n",
    "    RandomForestClassifier(random_state=RS),\n",
    "    param_grid_rfc,\n",
    "    cv=5,\n",
    "    scoring='f1',\n",
    "    n_iter=30,\n",
    "    random_state=RS\n",
    ")\n",
    "param_search_rfc.fit(X_train_final, y_train_final)\n",
    "\n",
    "results_rfc = pd.DataFrame(param_search_rfc.cv_results_)\n",
    "results_rfc = results_rfc.sort_values('rank_test_score')\n",
    "results_rfc.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "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>accuracy</th>\n",
       "      <th>precision</th>\n",
       "      <th>recall</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>ETC</th>\n",
       "      <td>0.981625</td>\n",
       "      <td>0.977273</td>\n",
       "      <td>0.889655</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MNB</th>\n",
       "      <td>0.972921</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.806897</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>RFC</th>\n",
       "      <td>0.976789</td>\n",
       "      <td>0.984</td>\n",
       "      <td>0.848276</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SVC</th>\n",
       "      <td>0.98646</td>\n",
       "      <td>0.985185</td>\n",
       "      <td>0.917241</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     accuracy precision    recall\n",
       "ETC  0.981625  0.977273  0.889655\n",
       "MNB  0.972921       1.0  0.806897\n",
       "RFC  0.976789     0.984  0.848276\n",
       "SVC   0.98646  0.985185  0.917241"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "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>accuracy</th>\n",
       "      <th>precision</th>\n",
       "      <th>recall</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>RFC</th>\n",
       "      <td>0.981625</td>\n",
       "      <td>0.992188</td>\n",
       "      <td>0.875862</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     accuracy precision    recall\n",
       "RFC  0.981625  0.992188  0.875862"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "rfc_final = ExtraTreesClassifier(\n",
    "    **param_search_rfc.best_params_,\n",
    "    random_state=RS\n",
    ")\n",
    "rfc_final.fit(X_train_final, y_train_final)\n",
    "\n",
    "rfc_scores = calc_scores(rfc_final, X_test_final, y_test_final)\n",
    "rfc_perf_df = create_perf_df({'RFC': rfc_scores})\n",
    "\n",
    "display(text_augmented_perf_df)\n",
    "display(rfc_perf_df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 108,
   "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>mean_fit_time</th>\n",
       "      <th>std_fit_time</th>\n",
       "      <th>mean_score_time</th>\n",
       "      <th>std_score_time</th>\n",
       "      <th>param_kernel</th>\n",
       "      <th>param_gamma</th>\n",
       "      <th>param_degree</th>\n",
       "      <th>param_class_weight</th>\n",
       "      <th>param_C</th>\n",
       "      <th>params</th>\n",
       "      <th>split0_test_score</th>\n",
       "      <th>split1_test_score</th>\n",
       "      <th>split2_test_score</th>\n",
       "      <th>split3_test_score</th>\n",
       "      <th>split4_test_score</th>\n",
       "      <th>mean_test_score</th>\n",
       "      <th>std_test_score</th>\n",
       "      <th>rank_test_score</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.208500</td>\n",
       "      <td>0.002050</td>\n",
       "      <td>0.047819</td>\n",
       "      <td>0.000487</td>\n",
       "      <td>linear</td>\n",
       "      <td>scale</td>\n",
       "      <td>3</td>\n",
       "      <td>balanced</td>\n",
       "      <td>0.5</td>\n",
       "      <td>{'kernel': 'linear', 'gamma': 'scale', 'degree...</td>\n",
       "      <td>0.902913</td>\n",
       "      <td>0.903226</td>\n",
       "      <td>0.920792</td>\n",
       "      <td>0.911765</td>\n",
       "      <td>0.908163</td>\n",
       "      <td>0.909372</td>\n",
       "      <td>0.006589</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>0.206866</td>\n",
       "      <td>0.003210</td>\n",
       "      <td>0.047958</td>\n",
       "      <td>0.000473</td>\n",
       "      <td>linear</td>\n",
       "      <td>auto</td>\n",
       "      <td>4</td>\n",
       "      <td>balanced</td>\n",
       "      <td>0.5</td>\n",
       "      <td>{'kernel': 'linear', 'gamma': 'auto', 'degree'...</td>\n",
       "      <td>0.902913</td>\n",
       "      <td>0.903226</td>\n",
       "      <td>0.920792</td>\n",
       "      <td>0.911765</td>\n",
       "      <td>0.908163</td>\n",
       "      <td>0.909372</td>\n",
       "      <td>0.006589</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>0.208077</td>\n",
       "      <td>0.003306</td>\n",
       "      <td>0.047730</td>\n",
       "      <td>0.000688</td>\n",
       "      <td>linear</td>\n",
       "      <td>auto</td>\n",
       "      <td>3</td>\n",
       "      <td>balanced</td>\n",
       "      <td>0.5</td>\n",
       "      <td>{'kernel': 'linear', 'gamma': 'auto', 'degree'...</td>\n",
       "      <td>0.902913</td>\n",
       "      <td>0.903226</td>\n",
       "      <td>0.920792</td>\n",
       "      <td>0.911765</td>\n",
       "      <td>0.908163</td>\n",
       "      <td>0.909372</td>\n",
       "      <td>0.006589</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.170360</td>\n",
       "      <td>0.003602</td>\n",
       "      <td>0.038231</td>\n",
       "      <td>0.000382</td>\n",
       "      <td>linear</td>\n",
       "      <td>scale</td>\n",
       "      <td>3</td>\n",
       "      <td>balanced</td>\n",
       "      <td>1</td>\n",
       "      <td>{'kernel': 'linear', 'gamma': 'scale', 'degree...</td>\n",
       "      <td>0.912621</td>\n",
       "      <td>0.898396</td>\n",
       "      <td>0.925373</td>\n",
       "      <td>0.897561</td>\n",
       "      <td>0.908163</td>\n",
       "      <td>0.908423</td>\n",
       "      <td>0.010233</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>36</th>\n",
       "      <td>0.147350</td>\n",
       "      <td>0.006508</td>\n",
       "      <td>0.030201</td>\n",
       "      <td>0.001086</td>\n",
       "      <td>linear</td>\n",
       "      <td>auto</td>\n",
       "      <td>3</td>\n",
       "      <td>balanced</td>\n",
       "      <td>10</td>\n",
       "      <td>{'kernel': 'linear', 'gamma': 'auto', 'degree'...</td>\n",
       "      <td>0.906404</td>\n",
       "      <td>0.881720</td>\n",
       "      <td>0.930000</td>\n",
       "      <td>0.913706</td>\n",
       "      <td>0.897959</td>\n",
       "      <td>0.905958</td>\n",
       "      <td>0.016060</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    mean_fit_time  std_fit_time  mean_score_time  std_score_time param_kernel  \\\n",
       "1        0.208500      0.002050         0.047819        0.000487       linear   \n",
       "16       0.206866      0.003210         0.047958        0.000473       linear   \n",
       "14       0.208077      0.003306         0.047730        0.000688       linear   \n",
       "3        0.170360      0.003602         0.038231        0.000382       linear   \n",
       "36       0.147350      0.006508         0.030201        0.001086       linear   \n",
       "\n",
       "   param_gamma param_degree param_class_weight param_C  \\\n",
       "1        scale            3           balanced     0.5   \n",
       "16        auto            4           balanced     0.5   \n",
       "14        auto            3           balanced     0.5   \n",
       "3        scale            3           balanced       1   \n",
       "36        auto            3           balanced      10   \n",
       "\n",
       "                                               params  split0_test_score  \\\n",
       "1   {'kernel': 'linear', 'gamma': 'scale', 'degree...           0.902913   \n",
       "16  {'kernel': 'linear', 'gamma': 'auto', 'degree'...           0.902913   \n",
       "14  {'kernel': 'linear', 'gamma': 'auto', 'degree'...           0.902913   \n",
       "3   {'kernel': 'linear', 'gamma': 'scale', 'degree...           0.912621   \n",
       "36  {'kernel': 'linear', 'gamma': 'auto', 'degree'...           0.906404   \n",
       "\n",
       "    split1_test_score  split2_test_score  split3_test_score  \\\n",
       "1            0.903226           0.920792           0.911765   \n",
       "16           0.903226           0.920792           0.911765   \n",
       "14           0.903226           0.920792           0.911765   \n",
       "3            0.898396           0.925373           0.897561   \n",
       "36           0.881720           0.930000           0.913706   \n",
       "\n",
       "    split4_test_score  mean_test_score  std_test_score  rank_test_score  \n",
       "1            0.908163         0.909372        0.006589                1  \n",
       "16           0.908163         0.909372        0.006589                1  \n",
       "14           0.908163         0.909372        0.006589                1  \n",
       "3            0.908163         0.908423        0.010233                4  \n",
       "36           0.897959         0.905958        0.016060                5  "
      ]
     },
     "execution_count": 108,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "param_grid_svc = {\n",
    "    'C': [0.1,  0.5,  1, 10, 100],\n",
    "    'kernel': ['linear', 'rbf', 'poly'],\n",
    "    'gamma': ['scale', 'auto'],\n",
    "    'degree': [2, 3, 4],\n",
    "    'class_weight': ['balanced', None]\n",
    "}\n",
    "\n",
    "param_search_svc = RandomizedSearchCV(\n",
    "    SVC(random_state=RS),\n",
    "    param_distributions=param_grid_svc,\n",
    "    cv=5,\n",
    "    scoring='f1',\n",
    "    n_iter=40,\n",
    "    random_state=RS\n",
    ")\n",
    "\n",
    "param_search_svc.fit(X_train_final, y_train_final)\n",
    "\n",
    "results_svc = pd.DataFrame(param_search_svc.cv_results_)\n",
    "results_svc = results_svc.sort_values('rank_test_score')\n",
    "results_svc.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 113,
   "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>accuracy</th>\n",
       "      <th>precision</th>\n",
       "      <th>recall</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>ETC</th>\n",
       "      <td>0.981625</td>\n",
       "      <td>0.977273</td>\n",
       "      <td>0.889655</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MNB</th>\n",
       "      <td>0.972921</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.806897</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>RFC</th>\n",
       "      <td>0.976789</td>\n",
       "      <td>0.984</td>\n",
       "      <td>0.848276</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SVC</th>\n",
       "      <td>0.98646</td>\n",
       "      <td>0.985185</td>\n",
       "      <td>0.917241</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     accuracy precision    recall\n",
       "ETC  0.981625  0.977273  0.889655\n",
       "MNB  0.972921       1.0  0.806897\n",
       "RFC  0.976789     0.984  0.848276\n",
       "SVC   0.98646  0.985185  0.917241"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "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>accuracy</th>\n",
       "      <th>precision</th>\n",
       "      <th>recall</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>SVC</th>\n",
       "      <td>0.981625</td>\n",
       "      <td>0.92</td>\n",
       "      <td>0.951724</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     accuracy precision    recall\n",
       "SVC  0.981625      0.92  0.951724"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "svc_final = SVC(\n",
    "    **param_search_svc.best_params_,\n",
    "    probability=True,\n",
    "    random_state=RS\n",
    ")\n",
    "svc_final.fit(X_train_final, y_train_final)\n",
    "\n",
    "svc_scores = calc_scores(svc_final, X_test_final, y_test_final)\n",
    "svc_perf_df = create_perf_df({'SVC': svc_scores})\n",
    "\n",
    "display(text_augmented_perf_df)\n",
    "display(svc_perf_df)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Ensemble"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 114,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.ensemble import StackingClassifier, VotingClassifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 118,
   "metadata": {},
   "outputs": [],
   "source": [
    "clfs_tuned_final = {\n",
    "    'MNB': mnb_final,\n",
    "    'ETC': etc_final,\n",
    "    'RFC': rfc_final,\n",
    "    'SVC': svc_final\n",
    "}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 120,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Voting Classifier\n",
      "Accuracy: 0.9912959381044487\n",
      "Precision: 0.9857142857142858\n",
      "Recall: 0.9517241379310345\n",
      "Confusion Matrix:\n",
      "[[887   2]\n",
      " [  7 138]]\n",
      "\n"
     ]
    }
   ],
   "source": [
    "clf_stack = StackingClassifier(\n",
    "    estimators=[*clfs_tuned_final.items()],\n",
    "    final_estimator=LogisticRegression(random_state=RS)\n",
    ")\n",
    "\n",
    "clf_stack.fit(X_train_final, y_train_final)\n",
    "scores = calc_scores(clf_stack, X_test_final, y_test_final)\n",
    "\n",
    "print('Voting Classifier')\n",
    "print_scores(*scores)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 121,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Voting Classifier\n",
      "Accuracy: 0.988394584139265\n",
      "Precision: 0.9925925925925926\n",
      "Recall: 0.9241379310344827\n",
      "Confusion Matrix:\n",
      "[[888   1]\n",
      " [ 11 134]]\n",
      "\n"
     ]
    }
   ],
   "source": [
    "clf_voting = VotingClassifier(\n",
    "    estimators=[*clfs_tuned_final.items()],\n",
    "    voting='soft',\n",
    ")\n",
    "\n",
    "clf_voting.fit(X_train_final, y_train_final)\n",
    "scores = calc_scores(clf_voting, X_test_final, y_test_final)\n",
    "\n",
    "print('Voting Classifier')\n",
    "print_scores(*scores)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Dumping"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Export model and vectorizer to files."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 122,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pickle\n",
    "\n",
    "pickle.dump(vectorizer_final, open('./model/vectorizer.pkl', 'wb'))\n",
    "pickle.dump(clf_voting, open('./model/classifier.pkl', 'wb'))"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "sklearn-env",
   "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": "3.11.5"
  },
  "vscode": {
   "interpreter": {
    "hash": "a5c098468f3ab8ce71de17d6097928b8e5b513ec64ad93e22897b4170adfd838"
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}