Advanced_nlp_techniques.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "aee2cf3f",
   "metadata": {},
   "source": [
    "### Read and Clean Text "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "d739f9d0",
   "metadata": {},
   "outputs": [],
   "source": [
    "import nltk\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "\n",
    "import re\n",
    "import string\n",
    "\n",
    "from sklearn.model_selection import train_test_split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "134d4729",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[nltk_data] Downloading package stopwords to /home/dkarl/nltk_data...\n",
      "[nltk_data]   Unzipping corpora/stopwords.zip.\n"
     ]
    }
   ],
   "source": [
    "nltk.download('stopwords')\n",
    "stopwords = nltk.corpus.stopwords.words('english')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "411ade10",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/dkarl/anaconda3/envs/nlp-prep-env/lib/python3.7/site-packages/ipykernel_launcher.py:1: ParserWarning: Falling back to the 'python' engine because the 'c' engine does not support regex separators (separators > 1 char and different from '\\s+' are interpreted as regex); you can avoid this warning by specifying engine='python'.\n",
      "  \"\"\"Entry point for launching an IPython kernel.\n"
     ]
    },
    {
     "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>label</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": [
       "   label                                               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": [
    "data = pd.read_csv('dataset/SMSSpamCollection.csv', sep='delimiter')\n",
    "messages = pd.DataFrame(columns=['label','text'])\n",
    "messages[['label','text']] = data['v1\\tv2'].str.split('\\t', expand=True)\n",
    "\n",
    "messages['label'] = np.where(messages['label'] == 'spam', 1, 0)\n",
    "\n",
    "messages.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "3b1b26f4",
   "metadata": {},
   "outputs": [],
   "source": [
    "def clean_text(text):\n",
    "    text = \"\".join([word.lower() for word in text if word not in string.punctuation])\n",
    "    tokens = re.split('\\W+', text)\n",
    "    text = [word for word in tokens if word not in stopwords]\n",
    "    return text"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "125d7e93",
   "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>label</th>\n",
       "      <th>text</th>\n",
       "      <th>clean_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",
       "      <td>[go, jurong, point, crazy, available, bugis, n...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>Ok lar... Joking wif u oni...</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",
       "      <td>[free, entry, 2, wkly, comp, win, fa, cup, fin...</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>[u, dun, say, early, hor, u, c, already, 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",
       "      <td>[nah, dont, think, goes, usf, lives, around, t...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   label                                               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",
       "                                          clean_text  \n",
       "0  [go, jurong, point, crazy, available, bugis, n...  \n",
       "1                     [ok, lar, joking, wif, u, oni]  \n",
       "2  [free, entry, 2, wkly, comp, win, fa, cup, fin...  \n",
       "3      [u, dun, say, early, hor, u, c, already, say]  \n",
       "4  [nah, dont, think, goes, usf, lives, around, t...  "
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "messages['clean_text'] = messages['text'].apply(lambda x: clean_text(x))\n",
    "messages.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "48b08861",
   "metadata": {},
   "source": [
    "### Splitting train and test data "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "cfed8683",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(\n",
    "    messages['clean_text'],\n",
    "    messages['label'],\n",
    "    test_size=0.2\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "bbe1eb28",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "2444    [ltgt, great, loxahatchee, xmas, tree, burning...\n",
       "1474    [gonna, blakes, night, might, able, get, littl...\n",
       "2073                                   [wondarfull, song]\n",
       "5171                                   [im, going, lunch]\n",
       "2724                 [tunde, wishing, great, day, abiola]\n",
       "3635    [big, difference, ltgt, versus, ltgt, every, l...\n",
       "2593    [friend, got, says, hes, upping, order, grams,...\n",
       "4567    [buy, blackberry, bold, 2, torch, buy, new, us...\n",
       "4530        [hiya, probably, coming, home, weekend, next]\n",
       "4193                       [ltgt, w, jetton, ave, forgot]\n",
       "Name: clean_text, dtype: object"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train[:10]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "e1946f27",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "2444    0\n",
       "1474    0\n",
       "2073    0\n",
       "5171    0\n",
       "2724    0\n",
       "3635    0\n",
       "2593    0\n",
       "4567    0\n",
       "4530    0\n",
       "4193    0\n",
       "Name: label, dtype: int64"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_train[:10]"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "cde10d6e",
   "metadata": {},
   "source": [
    "#### saving each set to a csv file"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "272144a6",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train.to_csv('dataset/X_train.csv', index=False, header=True)\n",
    "X_test.to_csv('dataset/X_test.csv', index=False, header=True)\n",
    "y_train.to_csv('dataset/y_train.csv', index=False, header=True)\n",
    "y_test.to_csv('dataset/y_test.csv', index=False, header=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2def88dc",
   "metadata": {},
   "source": [
    "-----------------------------------------------\n",
    "\n",
    "### TF-IDF Vectors\n",
    "    - on top of which a RandomForest Classifier can be built"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "8940d2c7",
   "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>clean_text</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>['ltgt', 'great', 'loxahatchee', 'xmas', 'tree...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>['gonna', 'blakes', 'night', 'might', 'able', ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>['wondarfull', 'song']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>['im', 'going', 'lunch']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>['tunde', 'wishing', 'great', 'day', 'abiola']</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                          clean_text\n",
       "0  ['ltgt', 'great', 'loxahatchee', 'xmas', 'tree...\n",
       "1  ['gonna', 'blakes', 'night', 'might', 'able', ...\n",
       "2                             ['wondarfull', 'song']\n",
       "3                           ['im', 'going', 'lunch']\n",
       "4     ['tunde', 'wishing', 'great', 'day', 'abiola']"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train = pd.read_csv('dataset/X_train.csv')\n",
    "X_test = pd.read_csv('dataset/X_test.csv')\n",
    "y_train = pd.read_csv('dataset/y_train.csv')\n",
    "y_test = pd.read_csv('dataset/y_test.csv')\n",
    "\n",
    "X_train.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "2a3771c7",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.feature_extraction.text import TfidfVectorizer\n",
    "\n",
    "tfidf_vect = TfidfVectorizer()\n",
    "tfidf_vect.fit(X_train['clean_text'])\n",
    "\n",
    "X_train_vect = tfidf_vect.transform(X_train['clean_text'])\n",
    "X_test_vect = tfidf_vect.transform(X_test['clean_text'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "3a7d8b3d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'ltgt': 4490,\n",
       " 'great': 3391,\n",
       " 'loxahatchee': 4477,\n",
       " 'xmas': 8094,\n",
       " 'tree': 7429,\n",
       " 'burning': 1657,\n",
       " 'update': 7573,\n",
       " 'totally': 7387,\n",
       " 'see': 6326,\n",
       " 'stars': 6807,\n",
       " 'gonna': 3332,\n",
       " 'blakes': 1457,\n",
       " 'night': 5041,\n",
       " 'might': 4735,\n",
       " 'able': 833,\n",
       " 'get': 3258,\n",
       " 'little': 4379,\n",
       " 'early': 2633,\n",
       " 'wondarfull': 7968,\n",
       " 'song': 6663,\n",
       " 'im': 3800,\n",
       " 'going': 3320,\n",
       " 'lunch': 4504,\n",
       " 'tunde': 7469,\n",
       " 'wishing': 7936,\n",
       " 'day': 2305,\n",
       " 'abiola': 831,\n",
       " 'big': 1423,\n",
       " 'difference': 2442,\n",
       " 'versus': 7667,\n",
       " 'every': 2790,\n",
       " 'hrs': 3709,\n",
       " 'friend': 3143,\n",
       " 'got': 3354,\n",
       " 'says': 6273,\n",
       " 'hes': 3581,\n",
       " 'upping': 7582,\n",
       " 'order': 5280,\n",
       " 'grams': 3375,\n",
       " 'buy': 1672,\n",
       " 'blackberry': 1453,\n",
       " 'bold': 1502,\n",
       " 'torch': 7378,\n",
       " 'new': 5019,\n",
       " 'used': 7608,\n",
       " 'let': 4308,\n",
       " 'know': 4179,\n",
       " 'plus': 5565,\n",
       " 'saying': 6272,\n",
       " 'wifi': 7908,\n",
       " 'ipad': 3922,\n",
       " 'hiya': 3615,\n",
       " 'probably': 5743,\n",
       " 'coming': 2039,\n",
       " 'home': 3644,\n",
       " 'weekend': 7833,\n",
       " 'next': 5027,\n",
       " 'jetton': 4020,\n",
       " 'ave': 1234,\n",
       " 'forgot': 3080,\n",
       " 'wish': 7932,\n",
       " 'hey': 3586,\n",
       " 'fancy': 2898,\n",
       " 'meetin': 4669,\n",
       " 'cha': 1828,\n",
       " 'hav': 3510,\n",
       " 'lil': 4344,\n",
       " 'beverage': 1412,\n",
       " 'txt': 7488,\n",
       " 'ring': 6122,\n",
       " 'meet': 4667,\n",
       " 'l8r': 4203,\n",
       " 'quite': 5855,\n",
       " 'tired': 7290,\n",
       " 'vpist': 7722,\n",
       " 'love': 4459,\n",
       " 'pete': 5472,\n",
       " 'much': 4903,\n",
       " 'bold2': 1503,\n",
       " 'nigeria': 5039,\n",
       " 'free': 3114,\n",
       " 'msg': 4887,\n",
       " 'gnarls': 3303,\n",
       " 'barkleys': 1308,\n",
       " 'crazy': 2188,\n",
       " 'ringtone': 6124,\n",
       " 'reply': 6044,\n",
       " 'go': 3304,\n",
       " 'message': 4711,\n",
       " 'right': 6116,\n",
       " 'yeah': 8126,\n",
       " 'totes': 7388,\n",
       " 'wanna': 7773,\n",
       " 'studying': 6911,\n",
       " 'ill': 3797,\n",
       " 'congrats': 2090,\n",
       " 'year': 8128,\n",
       " 'special': 6715,\n",
       " 'cinema': 1948,\n",
       " 'pass': 5402,\n",
       " 'call': 1699,\n",
       " '09061209465': 191,\n",
       " 'suprman': 6993,\n",
       " 'matrix3': 4630,\n",
       " 'starwars3': 6816,\n",
       " 'etc': 2772,\n",
       " 'bx420ip45we': 1681,\n",
       " '150pm': 326,\n",
       " 'dont': 2539,\n",
       " 'miss': 4772,\n",
       " 'sunshine': 6974,\n",
       " 'quiz': 5856,\n",
       " 'wkly': 7950,\n",
       " 'win': 7916,\n",
       " 'top': 7371,\n",
       " 'sony': 6665,\n",
       " 'dvd': 2624,\n",
       " 'player': 5545,\n",
       " 'country': 2160,\n",
       " 'liverpool': 4382,\n",
       " 'played': 5544,\n",
       " 'mid': 4731,\n",
       " 'week': 7832,\n",
       " 'ansr': 1061,\n",
       " '82277': 726,\n",
       " '150': 316,\n",
       " 'sptyrone': 6773,\n",
       " 'wait': 7747,\n",
       " 'thats': 7190,\n",
       " 'still': 6848,\n",
       " 'clear': 1967,\n",
       " 'sure': 6995,\n",
       " 'sarcastic': 6245,\n",
       " 'doesnt': 2516,\n",
       " 'want': 7775,\n",
       " 'live': 4380,\n",
       " 'us': 7603,\n",
       " 'okok': 5214,\n",
       " 'okthenwhats': 5217,\n",
       " 'ur': 7589,\n",
       " 'todays': 7325,\n",
       " 'plan': 5535,\n",
       " 'mums': 4919,\n",
       " 'present': 5705,\n",
       " 'ar': 1122,\n",
       " 'arrived': 1153,\n",
       " 'couple': 2162,\n",
       " 'days': 2310,\n",
       " 'lt3': 4485,\n",
       " 'yeshe': 8145,\n",
       " 'really': 5932,\n",
       " 'greatbhaji': 3392,\n",
       " 'told': 7334,\n",
       " 'kallis': 4091,\n",
       " 'best': 1403,\n",
       " 'cricketer': 2200,\n",
       " 'sachin': 6205,\n",
       " 'worldvery': 7996,\n",
       " 'tough': 7391,\n",
       " 'usual': 7619,\n",
       " 'guys': 3443,\n",
       " 'town': 7395,\n",
       " 'therere': 7208,\n",
       " 'definitely': 2358,\n",
       " 'people': 5444,\n",
       " 'around': 1145,\n",
       " 'painful': 5355,\n",
       " 'words': 7983,\n",
       " 'thought': 7240,\n",
       " 'happy': 3494,\n",
       " 'toughest': 7392,\n",
       " 'thing': 7218,\n",
       " 'earth': 2637,\n",
       " 'acting': 874,\n",
       " 'unspoken': 7563,\n",
       " 'pain': 5354,\n",
       " 'inside': 3882,\n",
       " 'class': 1959,\n",
       " 'holla': 3640,\n",
       " 'later': 4252,\n",
       " 'thanks': 7177,\n",
       " 'bomb': 1506,\n",
       " 'date': 2293,\n",
       " 'phone': 5489,\n",
       " 'wanted': 7777,\n",
       " 'say': 6268,\n",
       " 'urgent': 7592,\n",
       " '2nd': 458,\n",
       " 'attempt': 1201,\n",
       " 'contact': 2105,\n",
       " '900': 798,\n",
       " 'prize': 5737,\n",
       " 'yesterday': 8150,\n",
       " 'awaiting': 1242,\n",
       " 'collection': 2021,\n",
       " 'claim': 1953,\n",
       " '09061702893': 198,\n",
       " 'acl03530150pm': 867,\n",
       " 'watching': 7795,\n",
       " 'cartoon': 1780,\n",
       " 'listening': 4373,\n",
       " 'music': 4929,\n",
       " 'amp': 1030,\n",
       " 'eve': 2783,\n",
       " 'temple': 7139,\n",
       " 'church': 1945,\n",
       " 'ok': 5205,\n",
       " 'b4': 1256,\n",
       " 'leave': 4283,\n",
       " 'house': 3690,\n",
       " 'hello': 3550,\n",
       " 'boytoy': 1565,\n",
       " 'geeee': 3240,\n",
       " 'already': 1003,\n",
       " 'woke': 7962,\n",
       " 'bed': 1362,\n",
       " 'cuddling': 2228,\n",
       " 'ive': 3968,\n",
       " 'booked': 1513,\n",
       " 'pilates': 5516,\n",
       " 'yoga': 8159,\n",
       " 'lesson': 4306,\n",
       " 'haha': 3453,\n",
       " 'dude': 2612,\n",
       " 'sux': 7008,\n",
       " 'snake': 6621,\n",
       " 'old': 5223,\n",
       " 'raiden': 5870,\n",
       " 'buff': 1641,\n",
       " 'awww': 1250,\n",
       " 'dat': 2292,\n",
       " 'sweet': 7023,\n",
       " 'think': 7221,\n",
       " 'something': 6651,\n",
       " 'nice': 5033,\n",
       " 'time': 7276,\n",
       " 'tonight': 7357,\n",
       " 'cos': 2144,\n",
       " 'lonely': 4420,\n",
       " 'xxx': 8102,\n",
       " 'rct': 5909,\n",
       " 'thnq': 7234,\n",
       " 'adrian': 903,\n",
       " 'text': 7161,\n",
       " 'rgds': 6108,\n",
       " 'vatian': 7650,\n",
       " 'shell': 6421,\n",
       " 'guess': 3431,\n",
       " 'daal': 2260,\n",
       " 'voda': 7708,\n",
       " 'numbers': 5142,\n",
       " 'ending': 2715,\n",
       " '7548': 686,\n",
       " 'selected': 6339,\n",
       " 'receive': 5950,\n",
       " '350': 502,\n",
       " 'award': 1244,\n",
       " 'match': 4618,\n",
       " 'please': 5553,\n",
       " '08712300220': 91,\n",
       " 'quoting': 5861,\n",
       " 'code': 2004,\n",
       " '4041': 539,\n",
       " 'standard': 6799,\n",
       " 'rates': 5898,\n",
       " 'app': 1096,\n",
       " 'kkthis': 4167,\n",
       " 'month': 4843,\n",
       " 'kotees': 4191,\n",
       " 'birthday': 1443,\n",
       " 'oh': 5198,\n",
       " 'take': 7062,\n",
       " 'shopping': 6453,\n",
       " 'whatever': 7872,\n",
       " 'juliana': 4068,\n",
       " 'better': 1410,\n",
       " 'bb': 1332,\n",
       " 'wont': 7973,\n",
       " 'use': 7607,\n",
       " 'wife': 7905,\n",
       " 'doctor': 2510,\n",
       " 'sent': 6361,\n",
       " 'bucks': 1635,\n",
       " 'dunno': 2620,\n",
       " 'juz': 4081,\n",
       " 'askin': 1173,\n",
       " 'card': 1752,\n",
       " '20': 385,\n",
       " 'salon': 6228,\n",
       " 'called': 1709,\n",
       " 'hair': 3458,\n",
       " 'sense': 6357,\n",
       " 'tot': 7385,\n",
       " 'da': 2259,\n",
       " 'one': 5232,\n",
       " 'cut': 2248,\n",
       " 'lemme': 4298,\n",
       " 'youre': 8177,\n",
       " 'orange': 5276,\n",
       " 'customer': 2245,\n",
       " 'may': 4640,\n",
       " 'camera': 1726,\n",
       " 'upgrade': 7575,\n",
       " 'loyalty': 4479,\n",
       " '0207': 8,\n",
       " '153': 347,\n",
       " '9996': 807,\n",
       " 'offer': 5184,\n",
       " 'ends': 2718,\n",
       " '14thmarch': 315,\n",
       " 'tcs': 7103,\n",
       " 'apply': 1103,\n",
       " 'optout': 5273,\n",
       " 'availa': 1228,\n",
       " 'reach': 5914,\n",
       " 'orchard': 5279,\n",
       " 'wan': 7770,\n",
       " 'tickets': 7265,\n",
       " 'first': 3005,\n",
       " 'olol': 5225,\n",
       " 'printed': 5727,\n",
       " 'forum': 3090,\n",
       " 'post': 5641,\n",
       " 'guy': 3442,\n",
       " 'exact': 2816,\n",
       " 'prob': 5742,\n",
       " 'fixed': 3012,\n",
       " 'gpu': 3366,\n",
       " 'replacement': 6041,\n",
       " 'hopefully': 3669,\n",
       " 'ignore': 3790,\n",
       " 'sry': 6783,\n",
       " 'cant': 1743,\n",
       " 'talk': 7070,\n",
       " 'parents': 5384,\n",
       " 'need': 4990,\n",
       " 'strong': 6895,\n",
       " 'arms': 1141,\n",
       " 'laptop': 4237,\n",
       " 'matric': 4629,\n",
       " 'wif': 7904,\n",
       " 'lei': 4297,\n",
       " 'cool': 2128,\n",
       " 'inches': 3830,\n",
       " 'long': 4421,\n",
       " 'hope': 3666,\n",
       " 'like': 4338,\n",
       " 'persons': 5467,\n",
       " 'story': 6873,\n",
       " 'kit': 4153,\n",
       " 'strip': 6891,\n",
       " 'billed': 1428,\n",
       " '150p': 322,\n",
       " 'netcollex': 5008,\n",
       " 'ltd': 4486,\n",
       " 'po': 5570,\n",
       " 'box': 1544,\n",
       " '1013': 268,\n",
       " 'ig11': 3789,\n",
       " 'oja': 5204,\n",
       " 'anyway': 1082,\n",
       " 'sis': 6534,\n",
       " 'done': 2536,\n",
       " 'yet': 8151,\n",
       " 'dun': 2619,\n",
       " 'disturb': 2491,\n",
       " 'liao': 4315,\n",
       " 'anyone': 1074,\n",
       " 'calls': 1718,\n",
       " 'treadmill': 7424,\n",
       " 'youll': 8172,\n",
       " 'make': 4565,\n",
       " 'working': 7989,\n",
       " 'found': 3094,\n",
       " 'ad': 884,\n",
       " 'craigslist': 2179,\n",
       " 'selling': 6346,\n",
       " 'finally': 2980,\n",
       " 'today': 7320,\n",
       " 'project': 5765,\n",
       " 'started': 6810,\n",
       " 'aha': 947,\n",
       " 'sorry': 6679,\n",
       " 'meeting': 4670,\n",
       " 'sir': 6529,\n",
       " 'hmm': 3620,\n",
       " 'forgiven': 3078,\n",
       " 'havent': 3514,\n",
       " 'topic': 7372,\n",
       " 'rite': 6134,\n",
       " 'gobi': 3310,\n",
       " 'arts': 1158,\n",
       " 'college': 2022,\n",
       " 'give': 3285,\n",
       " 'pick': 5502,\n",
       " 'tuesday': 7465,\n",
       " 'evening': 2787,\n",
       " 'mm': 4794,\n",
       " 'umma': 7521,\n",
       " 'ask': 1170,\n",
       " 'vava': 7651,\n",
       " 'also': 1009,\n",
       " 'come': 2031,\n",
       " 'tell': 7128,\n",
       " 'play': 5543,\n",
       " 'together': 7328,\n",
       " 'good': 3334,\n",
       " 'ttyl': 7462,\n",
       " 'wylie': 8082,\n",
       " 'tampa': 7077,\n",
       " 'sarasota': 6243,\n",
       " 'planning': 5539,\n",
       " 'usually': 7621,\n",
       " 'stops': 6866,\n",
       " 'find': 2983,\n",
       " 'hella': 3549,\n",
       " 'weed': 7831,\n",
       " 'smoke': 6607,\n",
       " 'wine': 7921,\n",
       " 'dine': 2454,\n",
       " 'lady': 4217,\n",
       " 'x2': 8083,\n",
       " 'pls': 5560,\n",
       " 'nelson': 5003,\n",
       " 'bbs': 1337,\n",
       " 'longer': 4422,\n",
       " 'comin': 2038,\n",
       " 'money': 4832,\n",
       " 'expecting': 2841,\n",
       " 'aint': 960,\n",
       " 'lar': 4238,\n",
       " 'sleep': 6564,\n",
       " 'nite': 5057,\n",
       " 'afternoon': 927,\n",
       " 'sexy': 6390,\n",
       " 'buns': 1651,\n",
       " 'goes': 3317,\n",
       " 'job': 4028,\n",
       " 'search': 6311,\n",
       " 'wake': 7754,\n",
       " 'always': 1015,\n",
       " 'fine': 2985,\n",
       " 'adore': 899,\n",
       " 'opinion': 5261,\n",
       " 'jada': 3980,\n",
       " 'kusruthi': 4198,\n",
       " 'lovable': 4458,\n",
       " 'silent': 6508,\n",
       " 'spl': 6744,\n",
       " 'character': 1846,\n",
       " 'matured': 4634,\n",
       " 'stylish': 6923,\n",
       " 'simple': 6514,\n",
       " 'whos': 7895,\n",
       " 'hi': 3591,\n",
       " 'drugdealer': 2597,\n",
       " 'likely': 4340,\n",
       " 'mittelschmertz': 4787,\n",
       " 'google': 3349,\n",
       " 'paracetamol': 5375,\n",
       " 'worry': 8000,\n",
       " 'chat': 1860,\n",
       " 'set': 6378,\n",
       " '09096102316': 249,\n",
       " 'cum': 2231,\n",
       " '2moro': 450,\n",
       " 'luv': 4510,\n",
       " 'jane': 3991,\n",
       " 'xx': 8099,\n",
       " '1minmoremobsemspobox45po139wa': 372,\n",
       " 'dad': 2262,\n",
       " 'oredi': 5283,\n",
       " 'line': 4351,\n",
       " 'said': 6218,\n",
       " 'till': 7274,\n",
       " 'forget': 3075,\n",
       " 'realize': 5929,\n",
       " 'cannot': 1741,\n",
       " 'gn': 3302,\n",
       " 'yup': 8198,\n",
       " 'noe': 5071,\n",
       " 'leh': 4295,\n",
       " 'outta': 5319,\n",
       " 'gym': 3445,\n",
       " 'asked': 1172,\n",
       " 'anand': 1039,\n",
       " 'number': 5140,\n",
       " 'noooooooo': 5092,\n",
       " 'last': 4245,\n",
       " 'stress': 6883,\n",
       " 'life': 4326,\n",
       " 'fair': 2881,\n",
       " 'goal': 3308,\n",
       " 'arsenal': 1155,\n",
       " 'henry': 3567,\n",
       " 'scores': 6292,\n",
       " 'shot': 6464,\n",
       " 'yards': 8119,\n",
       " 'bergkamp': 1401,\n",
       " 'margin': 4598,\n",
       " '78': 688,\n",
       " 'mins': 4757,\n",
       " 'went': 7854,\n",
       " 'wats': 7800,\n",
       " 'dear': 2326,\n",
       " 'sleeping': 6566,\n",
       " 'ah': 946,\n",
       " 'yes': 8138,\n",
       " 'send': 6353,\n",
       " 'requests': 6052,\n",
       " 'came': 1725,\n",
       " 'back': 1274,\n",
       " 'double': 2552,\n",
       " 'coins': 2010,\n",
       " 'factory': 2873,\n",
       " 'gotta': 3359,\n",
       " 'cash': 1782,\n",
       " 'nitros': 5058,\n",
       " 'hours': 3689,\n",
       " 'tops': 7376,\n",
       " 'deal': 2319,\n",
       " 'care': 1756,\n",
       " 'aight': 959,\n",
       " 'chillin': 1919,\n",
       " 'friends': 3145,\n",
       " 'room': 6154,\n",
       " 'way': 7802,\n",
       " 'edison': 2654,\n",
       " 'rightly': 6118,\n",
       " 'fool': 3062,\n",
       " 'questions': 5849,\n",
       " 'wise': 7931,\n",
       " 'man': 4575,\n",
       " 'answer': 1062,\n",
       " 'speechless': 6724,\n",
       " 'viva': 7703,\n",
       " 'gmgngegnt': 3300,\n",
       " 'others': 5299,\n",
       " 'oi': 5201,\n",
       " 'mobile': 4806,\n",
       " '11mths': 289,\n",
       " 'oranges': 5278,\n",
       " 'latest': 4254,\n",
       " 'colour': 2024,\n",
       " 'mobiles': 4807,\n",
       " 'unlimited': 7557,\n",
       " 'upd8': 7571,\n",
       " 'freefone': 3118,\n",
       " '08000839402': 44,\n",
       " '2stoptx': 469,\n",
       " 'nokia': 5079,\n",
       " '3650': 505,\n",
       " 'video': 7676,\n",
       " '09066382422': 238,\n",
       " 'cost': 2146,\n",
       " '150ppm': 334,\n",
       " '3mins': 523,\n",
       " 'vary': 7647,\n",
       " '16': 350,\n",
       " 'close': 1979,\n",
       " '300603': 482,\n",
       " 'bcm4284': 1342,\n",
       " 'ldn': 4270,\n",
       " 'wc1n3xx': 7806,\n",
       " 'speak': 6713,\n",
       " 'email': 2698,\n",
       " 'oops': 5253,\n",
       " 'roommates': 6157,\n",
       " 'balance': 1289,\n",
       " '500': 603,\n",
       " 'question': 5847,\n",
       " 'sang': 6236,\n",
       " 'uptown': 7588,\n",
       " 'girl': 3279,\n",
       " '80s': 719,\n",
       " '83600': 741,\n",
       " 'luck': 4494,\n",
       " 'mine': 4749,\n",
       " 'possession': 5635,\n",
       " 'property': 5782,\n",
       " 'mmm': 4795,\n",
       " 'childish': 1912,\n",
       " 'smile': 6601,\n",
       " 'wen': 7852,\n",
       " 'tel': 7125,\n",
       " 'vikky': 7686,\n",
       " 'mail': 4558,\n",
       " 'il': 3795,\n",
       " 'aftr': 929,\n",
       " '6pm': 669,\n",
       " 'calicut': 1696,\n",
       " 'tried': 7432,\n",
       " '750': 684,\n",
       " 'anytime': 1081,\n",
       " 'network': 5011,\n",
       " 'half': 3463,\n",
       " 'price': 5720,\n",
       " 'rental': 6031,\n",
       " 'camcorder': 1724,\n",
       " '08000930705': 45,\n",
       " 'delivery': 2372,\n",
       " 'wed': 7824,\n",
       " 'smash': 6595,\n",
       " 'bros': 1612,\n",
       " 'religiously': 6009,\n",
       " 'grab': 3370,\n",
       " 'eat': 2644,\n",
       " 'mu': 4902,\n",
       " 'private': 5734,\n",
       " '2003': 389,\n",
       " 'account': 860,\n",
       " 'statement': 6818,\n",
       " '07808': 31,\n",
       " 'xxxxxx': 8105,\n",
       " 'shows': 6479,\n",
       " '800': 703,\n",
       " 'unredeemed': 7560,\n",
       " 'points': 5593,\n",
       " '08719899217': 157,\n",
       " 'identifier': 3780,\n",
       " '41685': 545,\n",
       " 'expires': 2849,\n",
       " '071104': 23,\n",
       " 'fifteen': 2958,\n",
       " 'fucking': 3173,\n",
       " 'minutes': 4764,\n",
       " 'respond': 6074,\n",
       " 'well': 7847,\n",
       " 'shes': 6424,\n",
       " 'surprise': 7001,\n",
       " 'idea': 3777,\n",
       " 'getting': 3265,\n",
       " 'soon': 6668,\n",
       " 'jot': 4054,\n",
       " 'things': 7220,\n",
       " 'remember': 6015,\n",
       " 'tv': 7476,\n",
       " 'charge': 1848,\n",
       " 'morning': 4856,\n",
       " 'designation': 2397,\n",
       " 'software': 6636,\n",
       " 'developer': 2414,\n",
       " 'chennai': 1886,\n",
       " 'wwwapplausestorecom': 8041,\n",
       " 'monthlysubscription50pmsg': 4845,\n",
       " 'max6month': 4638,\n",
       " 'tcsc': 7104,\n",
       " 'web': 7817,\n",
       " 'age16': 934,\n",
       " '2stop': 468,\n",
       " 'stop': 6860,\n",
       " 'arnt': 1144,\n",
       " 'pissed': 5525,\n",
       " 'id': 3775,\n",
       " 'would': 8010,\n",
       " 'tomorrow': 7346,\n",
       " 'xxxxxxxxxxxxxx': 8108,\n",
       " 'loverboy': 4467,\n",
       " 'thank': 7176,\n",
       " 'sister': 6535,\n",
       " 'law': 4264,\n",
       " 'meatballs': 4659,\n",
       " 'grins': 3404,\n",
       " 'babe': 1264,\n",
       " 'ha': 3448,\n",
       " 'nan': 4957,\n",
       " 'yalrigu': 8114,\n",
       " 'heltiniiyo': 3564,\n",
       " 'kothi': 4192,\n",
       " 'chikku': 1902,\n",
       " 'shared': 6413,\n",
       " 'many': 4589,\n",
       " 'wit': 7938,\n",
       " 'meso': 4709,\n",
       " 'far': 2903,\n",
       " 'didnt': 2434,\n",
       " 'body': 1499,\n",
       " 'even': 2786,\n",
       " 'uttered': 7624,\n",
       " 'word': 7981,\n",
       " 'abt': 838,\n",
       " 'trusting': 7447,\n",
       " 'plz': 5567,\n",
       " 'nxt': 5154,\n",
       " 'meok': 4701,\n",
       " 'chikkub': 1904,\n",
       " '6230': 642,\n",
       " 'digital': 2448,\n",
       " 'auction': 1210,\n",
       " 'part': 5392,\n",
       " '83383': 739,\n",
       " 'pobox11414tcrw1': 5574,\n",
       " 'lost': 4442,\n",
       " '10': 258,\n",
       " 'kilos': 4141,\n",
       " 'buzz': 1677,\n",
       " 'moment': 4826,\n",
       " 'againloving': 932,\n",
       " 'nah': 4946,\n",
       " 'help': 3555,\n",
       " 'never': 5016,\n",
       " 'iphone': 3924,\n",
       " 'feel': 2934,\n",
       " 'bad': 1278,\n",
       " 'rejected': 6000,\n",
       " 'kinda': 4144,\n",
       " 'cutest': 2251,\n",
       " 'world': 7993,\n",
       " 'sad': 6210,\n",
       " 'puppy': 5816,\n",
       " 'noise': 5075,\n",
       " 'darren': 2291,\n",
       " 'lor': 4436,\n",
       " 'oso': 5297,\n",
       " 'sian': 6491,\n",
       " 'tmr': 7309,\n",
       " 'haf': 3452,\n",
       " 'lect': 4286,\n",
       " 'velachery': 7658,\n",
       " 'playng': 5550,\n",
       " 'doors': 2547,\n",
       " 'game': 3208,\n",
       " 'gt': 3419,\n",
       " 'racing': 5864,\n",
       " 'lol': 4414,\n",
       " 'priest': 5722,\n",
       " 'restrict': 6084,\n",
       " 'eating': 2647,\n",
       " 'anythin': 1078,\n",
       " 'likes': 4341,\n",
       " 'two': 7486,\n",
       " 'knowing': 4181,\n",
       " '09095350301': 248,\n",
       " 'girls': 3282,\n",
       " 'erotic': 2755,\n",
       " 'ecstacy': 2650,\n",
       " '60pmin': 639,\n",
       " 'texts': 7169,\n",
       " '08712460324': 104,\n",
       " 'nat': 4965,\n",
       " 'rate': 5897,\n",
       " 'family': 2894,\n",
       " 'merry': 4706,\n",
       " 'mas': 4611,\n",
       " 'advance': 907,\n",
       " 'drink': 2582,\n",
       " 'plenty': 5558,\n",
       " 'water': 7796,\n",
       " 'usualiam': 7620,\n",
       " 'theres': 7209,\n",
       " 'lot': 4443,\n",
       " 'terms': 7148,\n",
       " 'mids': 4734,\n",
       " 'daddy': 2263,\n",
       " 'yo': 8158,\n",
       " 'car': 1751,\n",
       " 'accept': 847,\n",
       " 'dayu': 2314,\n",
       " 'brother': 1615,\n",
       " 'lover': 4465,\n",
       " 'dear1': 2327,\n",
       " 'best1': 1404,\n",
       " 'clos1': 1978,\n",
       " 'lvblefrnd': 4515,\n",
       " 'jstfrnd': 4062,\n",
       " 'cutefrnd': 2250,\n",
       " 'lifpartnr': 4331,\n",
       " 'belovd': 1392,\n",
       " 'swtheart': 7036,\n",
       " 'bstfrnd': 1628,\n",
       " 'rply': 6171,\n",
       " 'means': 4653,\n",
       " 'enemy': 2720,\n",
       " 'earlier': 2630,\n",
       " 'networking': 5012,\n",
       " 'gud': 3425,\n",
       " 'nyt': 5158,\n",
       " 'parentsi': 5385,\n",
       " 'hand': 3472,\n",
       " 'text82228': 7162,\n",
       " 'ringtones': 6128,\n",
       " 'logos': 4411,\n",
       " 'games': 3209,\n",
       " 'wwwtxt82228com': 8080,\n",
       " 'infotxt82228couk': 3865,\n",
       " 'ure': 7590,\n",
       " 'busy': 1666,\n",
       " 'bored': 1524,\n",
       " 'wat': 7791,\n",
       " 'choose': 1931,\n",
       " 'anything': 1079,\n",
       " 'tiime': 7272,\n",
       " 'hug': 3725,\n",
       " 'break': 1581,\n",
       " 'tears': 7116,\n",
       " 'rose': 6160,\n",
       " 'redred': 5973,\n",
       " 'bloodblood': 1479,\n",
       " 'heartheart': 3537,\n",
       " 'tis': 7295,\n",
       " 'including': 3835,\n",
       " '1u': 379,\n",
       " 'poor': 5616,\n",
       " 'relation': 6002,\n",
       " '2u': 471,\n",
       " 'support': 6986,\n",
       " '3u': 529,\n",
       " 'frnd': 3153,\n",
       " '4some1': 591,\n",
       " 'luvs': 4513,\n",
       " 'some1': 6644,\n",
       " 'praying': 5681,\n",
       " 'god': 3311,\n",
       " 'marry': 4607,\n",
       " 'try': 7450,\n",
       " 'dog': 2517,\n",
       " 'chile': 1916,\n",
       " 'ltdecimalgt': 4487,\n",
       " 'hour': 3688,\n",
       " 'drive': 2587,\n",
       " 'subletting': 6926,\n",
       " 'febapril': 2931,\n",
       " 'audition': 1212,\n",
       " 'season': 6313,\n",
       " 'whats': 7873,\n",
       " 'beer': 1371,\n",
       " 'msgs': 4890,\n",
       " 'iam': 3762,\n",
       " 'addicted': 889,\n",
       " 'msging': 4889,\n",
       " 'wrong': 8028,\n",
       " 'bcoz': 1343,\n",
       " 'bslvyl': 1625,\n",
       " 'england': 2726,\n",
       " 'macedonia': 4534,\n",
       " 'goalsteam': 3309,\n",
       " 'news': 5023,\n",
       " 'national': 4967,\n",
       " 'team': 7113,\n",
       " '87077': 766,\n",
       " 'eg': 2667,\n",
       " 'trywales': 7453,\n",
       " 'scotland': 6295,\n",
       " '4txtú120': 597,\n",
       " 'poboxox36504w45wq': 5584,\n",
       " 'dang': 2279,\n",
       " 'mean': 4649,\n",
       " 'supposed': 6991,\n",
       " 'couldnt': 2156,\n",
       " 'though': 7239,\n",
       " 'popping': 5623,\n",
       " 'ibuprofens': 3769,\n",
       " 'freemsg': 3121,\n",
       " 'replied': 6043,\n",
       " 'randy': 5891,\n",
       " 'female': 2945,\n",
       " 'local': 4397,\n",
       " 'hear': 3530,\n",
       " '08700621170150p': 64,\n",
       " 'per': 5446,\n",
       " 'end': 2713,\n",
       " 'ass': 1179,\n",
       " 'south': 6696,\n",
       " 'preferably': 5690,\n",
       " 'kegger': 4116,\n",
       " 'winner': 7923,\n",
       " 'specially': 6720,\n",
       " '1000': 260,\n",
       " 'holiday': 3638,\n",
       " 'flights': 3027,\n",
       " 'inc': 3828,\n",
       " 'operator': 5260,\n",
       " '0871277810810': 108,\n",
       " 'sea': 6310,\n",
       " 'lays': 4267,\n",
       " 'rock': 6142,\n",
       " 'envelope': 2746,\n",
       " 'paper': 5372,\n",
       " 'ard': 1127,\n",
       " 'christmas': 1940,\n",
       " 'enjoy': 2728,\n",
       " 'bring': 1600,\n",
       " 'food': 3061,\n",
       " 'kim': 4142,\n",
       " 'alright': 1005,\n",
       " 'thinkin': 7223,\n",
       " 'difficulties': 2446,\n",
       " 'works': 7992,\n",
       " 'another': 1059,\n",
       " 'request': 6051,\n",
       " 'jus': 4076,\n",
       " 'reached': 5916,\n",
       " 'bathe': 1323,\n",
       " 'using': 7615,\n",
       " 'net': 5007,\n",
       " 'finishes': 2994,\n",
       " 'goin': 3319,\n",
       " 'wrks': 8025,\n",
       " 'wot': 8007,\n",
       " 'bout': 1539,\n",
       " 'work': 7986,\n",
       " 'min': 4742,\n",
       " 'someone': 6647,\n",
       " 'contacted': 2106,\n",
       " 'dating': 2296,\n",
       " 'service': 6374,\n",
       " 'entered': 2736,\n",
       " 'landline': 4226,\n",
       " '09111032124': 255,\n",
       " 'pobox12n146tf150p': 5576,\n",
       " 'inperialmusic': 3878,\n",
       " 'listening2the': 4374,\n",
       " 'weirdest': 7841,\n",
       " 'track': 7400,\n",
       " 'ever': 2789,\n",
       " 'leafcutter': 4277,\n",
       " 'john': 4034,\n",
       " 'sounds': 6693,\n",
       " 'insects': 3880,\n",
       " 'molestedsomeone': 4824,\n",
       " 'plumbingremixed': 5563,\n",
       " 'evil': 2805,\n",
       " 'men': 4692,\n",
       " 'acid': 865,\n",
       " 'shu': 6484,\n",
       " 'looking': 4429,\n",
       " 'singapore': 6521,\n",
       " 'lets': 4309,\n",
       " 'supports': 6988,\n",
       " 'srt': 6782,\n",
       " 'thnk': 7233,\n",
       " 'ps3': 5795,\n",
       " 'usb': 7604,\n",
       " 'rajipls': 5879,\n",
       " 'favour': 2925,\n",
       " 'convey': 2120,\n",
       " 'wishes': 7934,\n",
       " 'nimya': 5053,\n",
       " 'switch': 7032,\n",
       " 'fone': 3057,\n",
       " 'dammit': 2272,\n",
       " 'took': 7364,\n",
       " 'omw': 5228,\n",
       " 'galsu': 3206,\n",
       " 'dinner': 2459,\n",
       " 'nìte': 5162,\n",
       " 'yar': 8118,\n",
       " 'keep': 4113,\n",
       " 'raining': 5873,\n",
       " 'non': 5085,\n",
       " 'elsewhere': 2695,\n",
       " 'dying': 2626,\n",
       " 'gone': 3329,\n",
       " 'late': 4248,\n",
       " 'since': 6517,\n",
       " 'side': 6496,\n",
       " 'fever': 2949,\n",
       " 'vomitin': 7715,\n",
       " 'arrested': 1150,\n",
       " 'shit': 6439,\n",
       " 'lttimegt': 4491,\n",
       " 'pm': 5568,\n",
       " 'ax': 1251,\n",
       " 'generally': 3246,\n",
       " 'brothas': 1614,\n",
       " 'meim': 4677,\n",
       " 'office': 5188,\n",
       " 'lmao': 4389,\n",
       " 'fun': 3184,\n",
       " 'mate': 4621,\n",
       " 'hows': 3700,\n",
       " 'honeydid': 3655,\n",
       " 'gimmi': 3277,\n",
       " 'de': 2317,\n",
       " 'gossx': 3353,\n",
       " 'teaching': 7111,\n",
       " 'til': 7273,\n",
       " 'twelve': 7478,\n",
       " 'lecture': 4287,\n",
       " 'damn': 2273,\n",
       " 'gd': 3235,\n",
       " 'wah': 7741,\n",
       " 'yan': 8117,\n",
       " 'jiu': 4025,\n",
       " 'fast': 2908,\n",
       " 'website': 7822,\n",
       " 'careful': 1760,\n",
       " 'easy': 2643,\n",
       " 'quick': 5851,\n",
       " 'bit': 1445,\n",
       " 'ambitious': 1018,\n",
       " 'wiskey': 7937,\n",
       " 'brandy': 1572,\n",
       " 'rum': 6190,\n",
       " 'gin': 3278,\n",
       " 'vodka': 7710,\n",
       " 'scotch': 6294,\n",
       " 'shampain': 6408,\n",
       " 'kudiyarasu': 4197,\n",
       " 'dhina': 2421,\n",
       " ...}"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tfidf_vect.vocabulary_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "cbd92266",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<1x8210 sparse matrix of type '<class 'numpy.float64'>'\n",
       "\twith 12 stored elements in Compressed Sparse Row format>"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_test_vect[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "208beefe",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[0., 0., 0., ..., 0., 0., 0.]])"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "### converting the sparse vector to an array\n",
    "X_test_vect[0].toarray()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "7a5fd7f1",
   "metadata": {},
   "outputs": [],
   "source": [
    "### fitting a random forest on the vectors\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "\n",
    "rf = RandomForestClassifier()\n",
    "\n",
    "rf_model = rf.fit(X_train_vect, y_train.values.ravel())\n",
    "    # values.ravel converts the pandas column (series) to an object type it can process (array-like)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "ff611554",
   "metadata": {},
   "outputs": [],
   "source": [
    "y_pred = rf_model.predict(X_test_vect)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "9cfbdd34",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "precision:  1.0 ;recall: 0.8417721518987342 accuracy: 0.978\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import precision_score, recall_score\n",
    "\n",
    "precision = precision_score(y_test, y_pred)\n",
    "recall = recall_score(y_test, y_pred)\n",
    "\n",
    "print(\"precision: \",precision, \";recall:\",recall, \"accuracy:\",round((y_pred==y_test['label']).sum()/len(y_pred), 3))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b14d1186",
   "metadata": {},
   "source": [
    "---------------------------------\n",
    "\n",
    "### word2vec Vectors"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "b76cf734",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/dkarl/anaconda3/envs/nlp-prep-env/lib/python3.7/site-packages/gensim/similarities/__init__.py:15: UserWarning: The gensim.similarities.levenshtein submodule is disabled, because the optional Levenshtein package <https://pypi.org/project/python-Levenshtein/> is unavailable. Install Levenhstein (e.g. `pip install python-Levenshtein`) to suppress this warning.\n",
      "  warnings.warn(msg)\n"
     ]
    }
   ],
   "source": [
    "import gensim "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "4c94fef9",
   "metadata": {},
   "outputs": [],
   "source": [
    "w2v_model = gensim.models.Word2Vec(\n",
    "    X_train,\n",
    "    vector_size=100,\n",
    "    window=5,\n",
    "    min_count=2\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "e9f27c12",
   "metadata": {},
   "outputs": [],
   "source": [
    "words = set(w2v_model.wv.index_to_key)\n",
    "X_train_vect = np.array([np.array([w2v_model.wv[i] for i in ls if i in words]) \n",
    "                        for ls in X_train['clean_text']])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "id": "6e3ba8aa",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_test_vect = np.array([np.array([w2v_model.wv[i] for i in ls if i in words]) \n",
    "                        for ls in X_test['clean_text']])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "51de7603",
   "metadata": {},
   "outputs": [],
   "source": [
    "### creating an element-wise average\n",
    "X_train_vect_avg = []\n",
    "\n",
    "for v in X_train_vect:\n",
    "    if v.size:\n",
    "        X_train_vect_avg.append(v.mean(axis=0))\n",
    "    else:\n",
    "        X_train_vect_avg.append(np.zeros(100, dtype=float))\n",
    "        \n",
    "        \n",
    "X_test_vect_avg = []\n",
    "\n",
    "for v in X_test_vect:\n",
    "    if v.size:\n",
    "        X_test_vect_avg.append(v.mean(axis=0))\n",
    "    else:\n",
    "        X_test_vect_avg.append(np.zeros(100, dtype=float))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "id": "d8db8012",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[-0.00053623,  0.00023643,  0.00510335, ..., -0.00704156,\n",
       "         0.00090146,  0.00639253],\n",
       "       [-0.00053623,  0.00023643,  0.00510335, ..., -0.00704156,\n",
       "         0.00090146,  0.00639253],\n",
       "       [-0.00861969,  0.00366574,  0.00518988, ..., -0.00239151,\n",
       "        -0.0095101 ,  0.00450588],\n",
       "       ...,\n",
       "       [-0.00861969,  0.00366574,  0.00518988, ..., -0.00239151,\n",
       "        -0.0095101 ,  0.00450588],\n",
       "       [-0.00861969,  0.00366574,  0.00518988, ..., -0.00239151,\n",
       "        -0.0095101 ,  0.00450588],\n",
       "       [-0.00053623,  0.00023643,  0.00510335, ..., -0.00704156,\n",
       "         0.00090146,  0.00639253]], dtype=float32)"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train_vect[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "id": "4673a752",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([-4.3402081e-03,  1.8502220e-03,  5.1440713e-03,  7.4717035e-03,\n",
       "       -1.4112473e-03, -6.6701593e-03,  3.9396919e-03,  7.5961840e-03,\n",
       "       -3.9917221e-03, -4.8975609e-03,  3.7142802e-03, -4.7501680e-03,\n",
       "       -5.0370372e-03,  6.8131043e-03, -9.9536055e-04,  2.4388935e-03,\n",
       "        4.7230111e-03,  4.0689884e-03, -6.1694235e-03, -5.2666953e-03,\n",
       "        4.9760528e-03,  5.5765238e-04,  7.5252396e-03, -4.2352667e-03,\n",
       "        6.5455968e-03, -4.3135061e-04, -2.8223696e-03,  5.1236860e-03,\n",
       "       -4.8006675e-03,  1.0744776e-03,  7.1262108e-04, -2.5452906e-03,\n",
       "        4.7675502e-03, -6.5551540e-03,  5.7662657e-04,  2.8549024e-04,\n",
       "        7.5191488e-03, -2.6878231e-04,  4.5126127e-03,  1.8472779e-03,\n",
       "       -1.3674948e-03, -6.3825905e-04, -8.9460704e-03, -2.4924977e-03,\n",
       "       -1.4773315e-03,  3.5581691e-03, -1.2613282e-03,  4.3627871e-03,\n",
       "        3.3486011e-03,  5.7305060e-03, -6.4004079e-04, -2.0952548e-03,\n",
       "       -2.2869380e-03,  2.0690809e-03,  4.0573976e-03,  1.5820337e-03,\n",
       "        6.6319872e-03, -5.1709474e-04, -2.2134928e-03,  8.6041046e-03,\n",
       "       -4.8513860e-03,  1.6793644e-03, -4.3746852e-03, -6.4621083e-03,\n",
       "        8.9087419e-04,  3.8347561e-03,  3.1861092e-03,  2.1593552e-04,\n",
       "        2.0458871e-03,  4.3143854e-03,  1.1425731e-03,  3.0284922e-04,\n",
       "        1.7891808e-03, -1.8114850e-03,  1.9432092e-03, -2.8548704e-03,\n",
       "        6.0527917e-04, -1.6056491e-03,  1.0235473e-03, -6.3587641e-03,\n",
       "       -4.8629106e-03,  3.2972337e-03,  5.1581576e-03,  7.0855243e-04,\n",
       "       -2.4505667e-03,  7.6671540e-06,  3.2349983e-03, -7.1912087e-03,\n",
       "       -3.7819587e-04,  1.3634003e-03,  1.3755554e-03, -6.4120902e-04,\n",
       "        2.8727122e-03, -3.5936844e-03,  3.8136442e-03,  3.9324127e-03,\n",
       "       -2.2048862e-03, -4.8533040e-03, -3.9980975e-03,  5.5046952e-03],\n",
       "      dtype=float32)"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train_vect_avg[0]"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1b8f7bfa",
   "metadata": {},
   "source": [
    "#### Fitting random forest classifier on top of word2vec model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "id": "3d248550",
   "metadata": {},
   "outputs": [],
   "source": [
    "rf_model = rf.fit(X_train_vect_avg, y_train.values.ravel())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "id": "d99247ba",
   "metadata": {},
   "outputs": [],
   "source": [
    "y_pred = rf_model.predict(X_test_vect_avg)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "id": "f35b46a2",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "precision:  0.58 ;recall: 0.18354430379746836 accuracy: 0.865\n"
     ]
    }
   ],
   "source": [
    "precision = precision_score(y_test, y_pred)\n",
    "recall = recall_score(y_test, y_pred)\n",
    "\n",
    "print(\"precision: \",precision, \";recall:\",recall, \"accuracy:\",round((y_pred==y_test['label']).sum()/len(y_pred), 3))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "392e6ac2",
   "metadata": {},
   "source": [
    "--------------------------------------\n",
    "\n",
    "### Creating doc2vec Vectors "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "id": "32155595",
   "metadata": {},
   "outputs": [],
   "source": [
    "tagged_docs_train = [gensim.models.doc2vec.TaggedDocument(v, [i]) \n",
    "                     for i,v in enumerate(X_train['clean_text'])]\n",
    "\n",
    "tagged_docs_test = [gensim.models.doc2vec.TaggedDocument(v, [i]) \n",
    "                    for i,v in enumerate(X_test['clean_text'])]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "id": "8ced10a4",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[TaggedDocument(words=\"['ltgt', 'great', 'loxahatchee', 'xmas', 'tree', 'burning', 'update', 'totally', 'see', 'stars']\", tags=[0]),\n",
       " TaggedDocument(words=\"['gonna', 'blakes', 'night', 'might', 'able', 'get', 'little', 'early']\", tags=[1]),\n",
       " TaggedDocument(words=\"['wondarfull', 'song']\", tags=[2]),\n",
       " TaggedDocument(words=\"['im', 'going', 'lunch']\", tags=[3]),\n",
       " TaggedDocument(words=\"['tunde', 'wishing', 'great', 'day', 'abiola']\", tags=[4]),\n",
       " TaggedDocument(words=\"['big', 'difference', 'ltgt', 'versus', 'ltgt', 'every', 'ltgt', 'hrs']\", tags=[5]),\n",
       " TaggedDocument(words=\"['friend', 'got', 'says', 'hes', 'upping', 'order', 'grams', 'hes', 'got', 'ltgt', 'get']\", tags=[6]),\n",
       " TaggedDocument(words=\"['buy', 'blackberry', 'bold', '2', 'torch', 'buy', 'new', 'used', 'let', 'know', 'plus', 'saying', 'buy', 'ltgt', 'g', 'wifi', 'ipad', 'saying', 'ltgt', 'g']\", tags=[7]),\n",
       " TaggedDocument(words=\"['hiya', 'probably', 'coming', 'home', 'weekend', 'next']\", tags=[8]),\n",
       " TaggedDocument(words=\"['ltgt', 'w', 'jetton', 'ave', 'forgot']\", tags=[9])]"
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tagged_docs_train[:10]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "id": "954dde63",
   "metadata": {},
   "outputs": [],
   "source": [
    "d2v_model = gensim.models.Doc2Vec(\n",
    "    tagged_docs_train,\n",
    "    vector_size = 100,\n",
    "    window = 5,\n",
    "    min_count = 2\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "id": "b3b5dfb8",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Infer the vectors to be used in training and testing\n",
    "    # eval function evaluates a string and converts it to a list\n",
    "train_vectors = [d2v_model.infer_vector(eval(v.words)) for v in tagged_docs_train]\n",
    "test_vectors = [d2v_model.infer_vector(eval(v.words)) for v in tagged_docs_test]"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "316026c7",
   "metadata": {},
   "source": [
    "#### fitting random forest on top of doc2vec"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "id": "dfa6178c",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "precision:  0.7 ;recall: 0.17721518987341772 accuracy: 0.873\n"
     ]
    }
   ],
   "source": [
    "rf_model = rf.fit(train_vectors, y_train.values.ravel())\n",
    "y_pred = rf_model.predict(test_vectors)\n",
    "\n",
    "precision = precision_score(y_test, y_pred)\n",
    "recall = recall_score(y_test, y_pred)\n",
    "\n",
    "print(\"precision: \",precision, \";recall:\",recall, \"accuracy:\",round((y_pred==y_test['label']).sum()/len(y_pred), 3))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "607382b3",
   "metadata": {},
   "source": [
    "------------------------------------------------\n",
    "\n",
    "### Building a classic RNN "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "id": "7d65b5c3",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train = pd.read_csv('dataset/X_train.csv')\n",
    "X_test = pd.read_csv('dataset/X_test.csv')\n",
    "y_train = pd.read_csv('dataset/y_train.csv')\n",
    "y_test = pd.read_csv('dataset/y_test.csv')\n",
    "\n",
    "\n",
    "\n",
    "from keras.preprocessing.text import Tokenizer\n",
    "from keras.preprocessing.sequence import pad_sequences\n",
    "\n",
    "tokenizer = Tokenizer()\n",
    "tokenizer.fit_on_texts(X_train['clean_text'])\n",
    "\n",
    "X_train_seq = tokenizer.texts_to_sequences(X_train['clean_text'])\n",
    "X_test_seq = tokenizer.texts_to_sequences(X_test['clean_text']) \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "id": "9e4899a4",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train_seq_padded = pad_sequences(X_train_seq, 50)\n",
    "X_test_seq_padded = pad_sequences(X_test_seq, 50)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "id": "04af4b27",
   "metadata": {},
   "outputs": [],
   "source": [
    "import keras.backend as K\n",
    "from keras.layers import Dense,Embedding,LSTM\n",
    "from keras.models import Sequential"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "id": "ab9a531a",
   "metadata": {},
   "outputs": [],
   "source": [
    "def recall_m(y_true, y_pred):\n",
    "    true_positives = K.sum(K.round(K.clip(y_true * y_pred, 0, 1)))\n",
    "    possible_positives = K.sum(K.round(K.clip(y_true, 0, 1)))\n",
    "    \n",
    "    recall = true_positives / (possible_positives + K.epsilon())\n",
    "    return recall\n",
    "\n",
    "\n",
    "def precision_m(y_true, y_pred):\n",
    "    true_positives = K.sum(K.round(K.clip(y_true * y_pred, 0, 1)))\n",
    "    predicted_positives = K.sum(K.round(K.clip(y_pred, 0, 1)))\n",
    "    \n",
    "    precision = true_positives / (predicted_positives + K.epsilon())\n",
    "    return precision"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "id": "170432cc",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_3\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "embedding_3 (Embedding)      (None, None, 32)          263712    \n",
      "_________________________________________________________________\n",
      "lstm_2 (LSTM)                (None, 32)                8320      \n",
      "_________________________________________________________________\n",
      "dense_4 (Dense)              (None, 32)                1056      \n",
      "_________________________________________________________________\n",
      "dense_5 (Dense)              (None, 1)                 33        \n",
      "=================================================================\n",
      "Total params: 273,121\n",
      "Trainable params: 273,121\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model = Sequential()\n",
    "\n",
    "model.add(Embedding(len(tokenizer.index_word)+1, 32))\n",
    "model.add(LSTM(32, dropout=0.2, recurrent_dropout=0))\n",
    "\n",
    "model.add(Dense(32, activation='relu'))\n",
    "model.add(Dense(1, activation='sigmoid'))\n",
    "\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "id": "40730518",
   "metadata": {},
   "outputs": [],
   "source": [
    "model.compile(\n",
    "    optimizer = 'adam',\n",
    "    loss = 'binary_crossentropy',\n",
    "    metrics = ['accuracy', precision_m, recall_m]\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "id": "a42517e2",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(4459, 50)\n",
      "(1115, 50)\n",
      "(4459, 1)\n",
      "(1115, 1)\n"
     ]
    }
   ],
   "source": [
    "print(X_train_seq_padded.shape)\n",
    "print(X_test_seq_padded.shape)\n",
    "\n",
    "print(y_train.shape)\n",
    "print(y_test.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "id": "4c5b7431",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/10\n",
      "140/140 [==============================] - 56s 352ms/step - loss: 0.4352 - accuracy: 0.8695 - precision_m: 0.1039 - recall_m: 0.0626 - val_loss: 0.0891 - val_accuracy: 0.9785 - val_precision_m: 0.9714 - val_recall_m: 0.8377\n",
      "Epoch 2/10\n",
      "140/140 [==============================] - 29s 204ms/step - loss: 0.0569 - accuracy: 0.9827 - precision_m: 0.9410 - recall_m: 0.8759 - val_loss: 0.0546 - val_accuracy: 0.9848 - val_precision_m: 0.9602 - val_recall_m: 0.8954\n",
      "Epoch 3/10\n",
      "140/140 [==============================] - 35s 248ms/step - loss: 0.0150 - accuracy: 0.9966 - precision_m: 0.9986 - recall_m: 0.9806 - val_loss: 0.0514 - val_accuracy: 0.9883 - val_precision_m: 0.9423 - val_recall_m: 0.9309\n",
      "Epoch 4/10\n",
      "140/140 [==============================] - 47s 338ms/step - loss: 0.0068 - accuracy: 0.9984 - precision_m: 0.9944 - recall_m: 0.9949 - val_loss: 0.0657 - val_accuracy: 0.9865 - val_precision_m: 0.9405 - val_recall_m: 0.9095\n",
      "Epoch 5/10\n",
      "140/140 [==============================] - 44s 314ms/step - loss: 0.0016 - accuracy: 0.9998 - precision_m: 0.9886 - recall_m: 0.9876 - val_loss: 0.0742 - val_accuracy: 0.9865 - val_precision_m: 0.9500 - val_recall_m: 0.9038\n",
      "Epoch 6/10\n",
      "140/140 [==============================] - 38s 268ms/step - loss: 7.3543e-04 - accuracy: 0.9998 - precision_m: 0.9584 - recall_m: 0.9574 - val_loss: 0.0788 - val_accuracy: 0.9874 - val_precision_m: 0.9500 - val_recall_m: 0.9095\n",
      "Epoch 7/10\n",
      "140/140 [==============================] - 45s 319ms/step - loss: 7.1124e-04 - accuracy: 1.0000 - precision_m: 0.9982 - recall_m: 0.9982 - val_loss: 0.0850 - val_accuracy: 0.9874 - val_precision_m: 0.9500 - val_recall_m: 0.9095\n",
      "Epoch 8/10\n",
      "140/140 [==============================] - 48s 337ms/step - loss: 2.2229e-04 - accuracy: 1.0000 - precision_m: 0.9828 - recall_m: 0.9828 - val_loss: 0.0857 - val_accuracy: 0.9874 - val_precision_m: 0.9500 - val_recall_m: 0.9095\n",
      "Epoch 9/10\n",
      "140/140 [==============================] - 38s 272ms/step - loss: 1.1705e-04 - accuracy: 1.0000 - precision_m: 0.9979 - recall_m: 0.9979 - val_loss: 0.0887 - val_accuracy: 0.9874 - val_precision_m: 0.9500 - val_recall_m: 0.9095\n",
      "Epoch 10/10\n",
      "140/140 [==============================] - 41s 290ms/step - loss: 8.4126e-05 - accuracy: 1.0000 - precision_m: 0.9942 - recall_m: 0.9942 - val_loss: 0.0942 - val_accuracy: 0.9874 - val_precision_m: 0.9500 - val_recall_m: 0.9095\n"
     ]
    }
   ],
   "source": [
    "history = model.fit(X_train_seq_padded, # training data\n",
    "                    y_train['label'],   # training target\n",
    "                    batch_size=32,      # batch size\n",
    "                    epochs=10,          # the number of epochs\n",
    "                    validation_data=(X_test_seq_padded, y_test)\n",
    "                   )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "id": "eba412a2",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXQAAAEICAYAAABPgw/pAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuNCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8QVMy6AAAACXBIWXMAAAsTAAALEwEAmpwYAAAtpElEQVR4nO3de3xU9bnv8c+Tyf1CAiThFiQgd6ogRqyXKt3WLdZWtGKVvW1FW2+tVXTbbmtri9t6tvvoPq22Vg8qWqwHvGLVjTeqoPVKUFCYoGIIEDOEECST+22e88esJJOQyyRMmMzM83695pWZtX6z5pkhfOeX31rrt0RVMcYYE/niwl2AMcaY0LBAN8aYKGGBbowxUcIC3RhjooQFujHGRAkLdGOMiRIW6CYsRGS9iPw4RNv6nYjsF5G9odjekSIiL4nIpUG0qxGRSUeiJhPZLNANIlIiIvVOcOwVkUdFJP0Ivv4SEfnHAJ87Hvg3YKaqjg5tZYNLVc9W1b8E0S5dVYuPRE0mslmgmzbfVdV0YA5wHPDL8JYTtAlAparu6+8TRST+cF88FNswJlQs0E0nqroXeAV/sAMgIl8XkXdE5KCIbBGR+QHrlohIsYhUi8hOEflXZ/kyEflrQLt8EdGuASgiM4AHgJOcvxAOOsu/LSJuZ7tfishNXWsVkW8BrwFjnec+6iw/V0S2OfWud16j7TklIvLvIvIxUNtdIDt1Xue8r/0icpeIxAW837dF5PcicgBYJiJJInK3iOwWkXIReUBEUgK2t1BENouIV0S+EJEFzvL2YScRmSwiG0SkynnNJ7rUM9m5nykiK0WkQkR2icivu9T2D6eWr5x/j7N7+edu2/56Z9jqHedzfEFERorI407NG0Ukv6/tmPCzQDediEgecDaww3k8Dvgf4HfACOAm4BkRyRGRNOBe4GxVzQBOBjb35/VUtQi4GnjXGVrIclY9DFzlbPdrwOvdPHedU2uZ89wlIjIVWAUsBXKAtcALIpIY8NTFwDlAlqq29FDa+UABMBdYCFwesO5EoBjIBe4A/guYiv9LcDIwDvgNgIjMA1YCPweygNOAkm5e73bgVWA4kAf8sYe6/ghkApOA04EfApd1qe1TIBv438DDIiI9bCvQxcAPnNqPBt4FHsH/b14E/DaIbZgws0A3bZ4TkWpgD7CPjv/AlwBrVXWtqvpU9TWgEPi2s94HfE1EUlTVo6rbQlRPMzBTRIap6leq+mGQz7sI+B9VfU1Vm4G7gRT8XzZt7lXVPapa38t2/ktVD6jqbuAP+L8E2pSp6h+dL4MG4ArgBqd9NfC/8AckwI+AFU49PlX9UlW39/B+JwBjVbVBVQ/ZpyAiLuf9/VJVq1W1BPhv/EHcZpeqPqiqrcBfgDHAqF7eZ5tHVPULVa0CXgK+UNV1znt8Cv8wnBniLNBNm/Oc3vB8YDr+Hh74Q+ZCZ/jioDMkciowRlVr8QfM1YBHRP5HRKaHqJ4L8H9p7HKGIk4K8nljgV1tD1TVh/9LalxAmz1BbCewzS5nu92tywFSgU0Bn8/LznKA8cAXQbzeLwABPnCGiy7vpk02kEjA+3PuB7639iN9VLXOuRvMDu7ygPv13Tw+YjvJzcBZoJtOVHUD8Cj+ni34w+sxVc0KuKWp6p1O+1dU9Uz8PcHtwIPO82rxB12b3o5AOWTKT1XdqKoL8Q9rPAc8GeRbKMP/JQSAM9wwHviyt9frxviA+0c52+3u+fvxB96sgM8n09nBDP7P7+i+XkxV96rqFao6FrgK+HPbuHmX12rryQfW9iXGYIFuuvcH4EwRmQP8FfiuiJwlIi4RSRaR+SKSJyKjnB2QaUAjUAO0OtvYDJwmIkeJSCa9HzVTDuS1jXOLSKKI/KuIZDrDJt6A7fblSeAcETlDRBLwH9LYCLzTj/cP8HMRGS7+wyKvB57orpHzF8CDwO9FJNepf5yInOU0eRi4zKknzll3yF8xInKhs/8C4Cv8Xxqd3rMzjPIkcIeIZIjIBOBG/P9Gxligm0OpagX+HXm3quoe/DsFbwEq8Pc4f47/dycOf2CWAQfw76T7ibON1/CH4MfAJuDFXl7ydWAbsFdE9jvLfgCUiIgX/5DOJUHW/qnT9o/4e7TfxX9IZlMwzw/wN6fuzfh3Cj/cS9t/x78T+T2n3nXANKeeD/DvtPw9UAVsoHMPu80JwPsiUgM8D1yvqju7afcz/H/9FAP/AP4fsKKf781EKbELXBjTmYgoMEVVd4S7FmP6w3roxhgTJewsN2NigDOU052zVfWtI1qMGTQ25GKMMVHChlyMMSZKhG3IJTs7W/Pz88P18sYYE5E2bdq0X1VzulsXtkDPz8+nsLAwXC9vjDERSUR29bTOhlyMMSZKWKAbY0yUsEA3xpgoYYFujDFRos9AF5EVIrJPRLb2sF5E5F4R2SEiH4vI3NCXaYwxpi/B9NAfBRb0sv5sYIpzuxK4//DLMsYY0199Brqqvol/Jr2eLARWqt97QJaIjAlVgcYYY4ITiuPQx9H5Ci6lzjJP14YiciX+XjxHHXVUCF46trX6lFe37aVobzUuEeJdgitOcIn/Z7xLiBMhPk6Ii/P/dDm3+DhnnUtwxcW1PydwvSvu0GXtz2l7jbg44uIgPi6O5IQ4grt8pTFmMIQi0Lv7H9ztBDGquhxYDlBQUGCTyAxQXVMLTxWW8tA/itlzoLfLYh5ZGcnxTMlNZ+qoDKaMymDqKP/93IwkC/owUFVqm1oZCvM1JbjiSE5whbuMsFNVGpp9+FRJSwr9eZ2h2GIpnS/XlUfny3WZEKmobmTluyU89t4uDtY1M/eoLH717ZmcOdN/DeAWnw+fz/+z1acdN1VaWhWfKi2By51bi89Z19rRvtXno9VH+8/uthn4uLlVKTtYz2fl1bzqLmf1xo4/2oYlxx8S8lNGpZOTbkF/uHw+xeNtYFdlLbsq65xbLSWVdeyurKW2KdgLPQ2+pPg4slITyEpJJDMlgczUBLJSEvzLUhMZlhLwOKBNRlI8cXFD6/fE51OqG1o4WN/EwbpmDtY3U1XfTFVdx+ODdc6yLm2aWnxc+83J3HTWtJDXFYpAfx64VkRWAycCVap6yHCLGbgvKmp46K2dPPNhKc2tPs6cMYorT5tEQf6ITu1ccW09oPD3hPbXNPJZeTWfl9e0/3xpq4dVHzS3t8lKTWBKbro/6J2e/dTRGWSnJ4Wx8qGnudVH6Vf1h4T2rgN17D5QR1OLr71tgksYPyKVCSNSOXHiCMZkJuMaAmHY2OLD64RcWwjuOVDHVmdZfXPPXzxxApkp3YV+Apmp/uDv+GJI8H8ROF8IifG97yZsbGl1gtgJ3PYwbqIqIJQPtoW1s8zb0Exvf/ikJbqcL6REslISmJybTlZqglN/IgX5wwf6Ufaqz0AXkVX4rwSfLSKlwG+BBABVfQBYi//q7DuAOvyX2zKHSVUp3PUV/3dDMeuKykmMj2PR8Xn86NSJHJ0z9C/Anp2eRHZ6Eicfnd2+TFWpqGlsD/nPymv4vLyaF7eU4W1oaW83Ii2RybnpHb35XH/PfmQUB319Uyu7D9R1hPYB/8+SylrKDjbQ6utIj5QEFxNGpnJ0ThpnTM9lwsg0JoxMZcLIVMZkpgyJAO+vrsF6sK4jVKvavwj8yw7WNVFSWRt0sAZ+ESjaEdJBfpG0B3NqIvnZaf4vkoCwbvsS8f8M7otksIRtPvSCggK1ybkO1bajc/lbxXy0+yDDUxP4wUn5/PCkCVHbc1VV9lU3dgr5tl59dWNH0I9MS2TKqIAxeqdXPzwtMYzVB8/b0Myu/R1h3TY0squylnJvY6e2w5LjmZidxlEj08gfmcpRI1LJz/YHtw1VdWj1KTVdhj4O1jUF/DXQ1sv2rxeBLCeI20O4azCnJA7ZoR4AEdmkqgXdrrNAHxrqm1p5etMeHvrHTnZV1nHUiFSu+MZEFh0/npTE8A+hhIOqstfb0Cnk2+4Hjg1npye19+Yn56aTPgg7m/qryRkm2R0Q2l/VNXdqk5ORxIQRqUxoC+2RqeQ7ve2s1Mj4kjJHngX6EFZZ08hf3t3FY++W8FVdM3PGZ3HVaZP451mjI/JP5yNBVSmranB68R0h//m+GuqG0E5AERibmeIMh/iDOt+5f9SI1EE5ysFEv94CPfJ+o1qbAQFX5JUeaOf+Wh58q5hnNpXS2OLjWzNGcdXpkyiYMNz+nO6DiDAuK4VxWSl8c1pu+3Kfz9+jbwzYSRguLhFGZSaRFB+bf12Z8Ii8VPzsFXjyB5A+CoaNdW7juvwcCxljIH7ojTlv2nWA5W8W86q7nARXHBfMHcePTp3E5Nyhv6NzqIuLE8ZmpYS7DGPCJvICfcQk+MZN4C0D75dQ8Rl88QY0dXNR87ScQ4M+MPwzxkBi6qCX3OpTXnOXs/zNL/hw90GyUhO49puT+eFJ+eRkDL0vHWNMZIq8QB8103/rqsHbEfLess73v9oFu96BhoOHPi9leEfIZ4zpPvyThw2o1IbmVp7eVMrD/9jJzv21jB+Rwm3nzuLCgjxSEyPvozfGDG3RkyrJw/y33Ok9t2mqBa8Hqsu6D/+yj6C24tDnJWZ008MfC1njIXem/4sgYNz7QG0TK98tYeW7uzhQ28TsvEzu+5e5nDVrFPEum4LeGDM4oifQg5GYBtmT/beetDRCtefQXn7bzy+2Q/VeOk1Xk5wFuTPxDpvC+oPZPLErnU+ax3HC9ElccdokTpw4wnZ0GmMGXWwFejDik2B4vv/Wk9ZmqCmHAzthXxEVxR9Rtetjcnc9yblSz7ku/GffV46Fd2bCjhmQOwtyZ0DONEiwHXfGmNCzQB8IVwK+jHGsK43nwY9cbCwZx7Dk8/jBSUdx2TGJZNd+AfvcHbedb0GrcyagxPl37ObO9N9GOT9HTII4O8TNGDNwFugDtPLdEpa94GZcVgq//e5Mvl8wPuBEkaNh6j93NG5tgQPFnUO+fBsUvUD70E18MmRPhVFOT76tRz9sbKfx+ZjV2uzfv1FTDjX7An4699vW1Vb4P29jhrKTfgr/9KuQb9YCfYDe/qKSidlpvHbDaX3v6HTFQ85U/23WeR3Lm+pg/6ewr8gf8PuKoHgDbFnV0SY5s6M3nzujI/BTBme2tiPK1wp1lZ0DujYgpGvKocYJ6voeLpqVlAnpuf7zEkYf4z9U1WWnzZshLq/bEz0PmwX6ALnLvMydMPzwjlpJTIWxx/lvgeoOQMX2jpDf54ZPnobGqo42GWOdnvwMyBzv/9KIi4e4BHAl+Idv2u/Hd9xcCf7lca4e7sc720oIaN+PoSBVqP+qj960c7+2ArSbszoTUjtCOnsyTDjZfz89t2N5ei6k5UJC8sA+e2OikAX6AFTVN/PlwXr+9euDdBm91BH+EJtwcscyVf9RNvuKYF9A0G98CFoaBqeOdhLcl0FjtT+sW5sO3YQr0R/A6bkwLM//JZY+qnM4t4V1kp01a8xAWKAPQJHHC8DMMQM74WhARCBznP825Vsdy32t0FAFvhb/rbW5y/1mf5v2+y3+Meb2+13b9+e5XbaTlOEf8uiuN52cZfsCjBlkFugDEJZA70mcy9+jN8bEPDttcQDcZV6y0xNtHhZjzJBigT4ARXu9zBgzzM7+NMYMKRbo/dTc6uOzvTVDY7jFGGMCWKD3U3FFLU2tPmZYoBtjhhgL9H5ye/zHgs8ca4FujBlaLND7qchTTWJ8HJOy08JdijHGdGKB3k/uMi/TRmXYvObGmCHHUqkfVJUij5cZYzLCXYoxxhwiqEAXkQUi8qmI7BCRm7tZP1xE1ojIxyLygYh8LfSlht++6kYqa5vsCBdjzJDUZ6CLiAu4DzgbmAksFpGuF/W8BdisqscCPwTuCXWhQ4HbOUPUjnAxxgxFwfTQ5wE7VLVYVZuA1cDCLm1mAn8HUNXtQL6IjApppUOAu8wJdDvCxRgzBAUT6OOAPQGPS51lgbYA3wMQkXnABCCv64ZE5EoRKRSRwoqKbi7GPMQVebzkDU9hWHJCuEsxxphDBBPo3Z3frl0e3wkMF5HNwM+Aj4BDLhujqstVtUBVC3Jycvpba9i5PV4bPzfGDFnBzLZYCowPeJwHlAU2UFUvcBmA+Cc42encokZdUws799fy3WPHhrsUY4zpVjA99I3AFBGZKCKJwMXA84ENRCTLWQfwY+BNJ+Sjxqd7q1G1M0SNMUNXnz10VW0RkWuBVwAXsEJVt4nI1c76B4AZwEoRaQXcwI8GseawKPJUA0NkDnRjjOlGUBe4UNW1wNouyx4IuP8uMCW0pQ0tbk8VGUnx5A1PCXcpxhjTLTtTNEhFnmqbA90YM6RZoAfB5/Of8m/j58aYocwCPQi7D9RR19Rqc7gYY4Y0C/QguNsvCp0Z5kqMMaZnFuhBKPJ4ccUJU0alh7sUY4zpkQV6ENxlXo7OSSM5wRXuUowxpkcW6EHwz4FuO0SNMUObBXofDtY1UVbVYCcUGWOGPAv0Ptgc6MaYSGGB3of2OdAt0I0xQ5wFeh+KPNXkZCSRk5EU7lKMMaZXFuh9sDnQjTGRwgK9F00tPnbsq7bhFmNMRLBA78WOfTU0t6rN4WKMiQgW6L0oaj/l3+ZwMcYMfRbovXB7vCQnxDEx2075N8YMfRbovSjyeJk2KgNXnM2BbowZ+izQe6Cq/iNcbPzcGBMhLNB7sNfbwMG6ZjvCxRgTMSzQe9B2hqgdg26MiRQW6D1oO8JlugW6MSZCWKD3wO3xMmFkKulJ8eEuxRhjgmKB3oMiTzUzRlvv3BgTOSzQu1Hb2EJJZa0d4WKMiShBBbqILBCRT0Vkh4jc3M36TBF5QUS2iMg2Ebks9KUeOdv3VqNqU+YaYyJLn4EuIi7gPuBsYCawWERmdmn2U8CtqrOB+cB/i0hiiGs9YtouamE9dGNMJAmmhz4P2KGqxaraBKwGFnZpo0CGiAiQDhwAWkJa6RFU5PEyLDmesZnJ4S7FGGOCFkygjwP2BDwudZYF+hMwAygDPgGuV1Vf1w2JyJUiUigihRUVFQMsefC5y/xniPq/n4wxJjIEE+jdpZp2eXwWsBkYC8wB/iQih4xXqOpyVS1Q1YKcnJx+lnpktPqUT/faHOjGmMgTTKCXAuMDHufh74kHugx4Vv12ADuB6aEp8cgqqaylvrnVzhA1xkScYAJ9IzBFRCY6OzovBp7v0mY3cAaAiIwCpgHFoSz0SGk7Q9R66MaYSNPnaZCq2iIi1wKvAC5ghapuE5GrnfUPALcDj4rIJ/iHaP5dVfcPYt2Dxl3mJT5OmDLK5kA3xkSWoM5rV9W1wNouyx4IuF8G/HNoSwuPIo+XybnpJMW7wl2KMcb0i50p2oXb47Xxc2NMRLJAD1BZ00i5t9HGz40xEckCPUCRpxqwM0SNMZHJAj2AHeFijIlkFugB3B4vo4clMyItYqehMcbEMAv0AEUeLzPGZIS7DGOMGRALdEdjSys79tXY+LkxJmJZoDs+L6+hxac2fm6MiVgW6I72OdAt0I0xEcoC3VHk8ZKS4GLCyLRwl2KMMQNige5wl3mZPiYDV5zNgW6MiUwW6ICqOke42HCLMSZyWaADXx6sx9vQYuPnxpiIZoFOxyn/1kM3xkQyC3T84+ciMH20nVRkjIlcFuj4j3DJH5lGWlJQ08MbY8yQZIGOzYFujIkOMR/o1Q3N7D5QZ3O4GGMiXswH+va9Nge6MSY6xHyg2xzoxphoEfOB7i7zMjw1gdHDksNdijHGHJaYD/S2M0RF7JR/Y0xki+lAb2n1sX1vtR3hYoyJCjEd6CWVtTS2+Gz83BgTFYIKdBFZICKfisgOEbm5m/U/F5HNzm2riLSKyIjQlxta28qcOdDtCBdjTBToM9BFxAXcB5wNzAQWi8jMwDaqepeqzlHVOcAvgQ2qemAQ6g2pIk81CS7h6Jz0cJdijDGHLZge+jxgh6oWq2oTsBpY2Ev7xcCqUBQ32NweL1NyM0iMj+mRJ2NMlAgmycYBewIelzrLDiEiqcAC4Jke1l8pIoUiUlhRUdHfWkPO5kA3xkSTYAK9u+P5tIe23wXe7mm4RVWXq2qBqhbk5OQEW+OgqKhupKK60U75N8ZEjWACvRQYH/A4Dyjroe3FRMhwS9sZorZD1BgTLYIJ9I3AFBGZKCKJ+EP7+a6NRCQTOB34W2hLHBzutkC3IRdjTJTocwJwVW0RkWuBVwAXsEJVt4nI1c76B5ym5wOvqmrtoFUbQkUeL2Mzk8lKTQx3KcYYExJBXdFBVdcCa7sse6DL40eBR0NV2GBzl9kOUWNMdInJ4/Uamlsp3l9r4+fGmKgSk4H+WXk1rT61HroxJqrEZKAX2Q5RY0wUislAd5d5SUt0cdSI1HCXYowxIROTgV7kqWb6mGHExdkc6MaY6BFzga6qzin/doaoMSa6xFygl35VT3VjCzPHZIa7FGOMCamYC/S2OdCth26MiTYxF+hFHi9xAtNH2xEuxpjoEnOB7vZ4yc9OIyXRFe5SjDEmpGIu0Is8Xjv+3BgTlWIq0Kvqmyn9qt7OEDXGRKWYCvTtNge6MSaKxVSg2xzoxphoFlOBXuTxMjItkdyMpHCXYowxIRdTge52LgotYqf8G2OiT8wEenOrj8/Ka2z83BgTtWIm0Isramlq8dkZosaYqBUzgd4xB7rN4WKMiU4xE+huj5dEVxyTctLCXYoxxgyKmAn0Io+XqaPTSXDFzFs2xsSYmEg3VcVd5mWGTchljIliMRHoFdWNVNY22REuxpioFhOBvs3TNge6BboxJnoFFegiskBEPhWRHSJycw9t5ovIZhHZJiIbQlvm4SmyQDfGxID4vhqIiAu4DzgTKAU2isjzquoOaJMF/BlYoKq7RSR3kOodEHeZl3FZKWSmJIS7FGOMGTTB9NDnATtUtVhVm4DVwMIubf4FeFZVdwOo6r7Qlnl4ijxeGz83xkS9YAJ9HLAn4HGpsyzQVGC4iKwXkU0i8sPuNiQiV4pIoYgUVlRUDKzifqpvamXn/lobbjHGRL1gAr27may0y+N44HjgHOAs4FYRmXrIk1SXq2qBqhbk5OT0u9iB+LS8Gp/alLnGmOjX5xg6/h75+IDHeUBZN232q2otUCsibwKzgc9CUuVhcJfZHOjGmNgQTA99IzBFRCaKSCJwMfB8lzZ/A74hIvEikgqcCBSFttSBKfJ4yUiKJ294SrhLMcaYQdVnD11VW0TkWuAVwAWsUNVtInK1s/4BVS0SkZeBjwEf8JCqbh3MwoPl9niZPiaDuDibA90YE92CGXJBVdcCa7sse6DL47uAu0JX2uHz+ZTtHi+Ljs8LdynGGDPoovpM0d0H6qhtarUjXIwxMSGqA719DnQ7Bt0YEwOiOtDdHi9xAlNH2VWKjDHRL6oDvcjj5eicdJITXOEuxRhjBl1UB7q7zGvj58aYmBG1gX6wromyqgYbPzfGxIyoDXS3TZlrjIkxURvoRZ5qwE75N8bEjqgNdHeZl+z0JHIyksJdijHGHBFRG+g2B7oxJtZEZaA3tfj4fF81M8bY8efGmNgRlYH+RUUNza1q4+fGmJgSlYFuc6AbY2JRVAZ6kcdLUnwcE7PTwl2KMcYcMVEZ6G6Pl2mjM4h3ReXbM8aYbkVd4qmq/wgXG24xxsSYqAv0vd4GvqprtjNEjTExJ+oC3eZAN8bEqqgL9LYjXKaPtmPQjTGxJeoCvchTzVEjUslITgh3KcYYc0RFXaC7PV47Q9QYE5OiKtBrG1soqaxl5pjMcJdijDFHXFQF+va91ahiPXRjTEyKqkC3I1yMMbEsqEAXkQUi8qmI7BCRm7tZP19EqkRks3P7TehL7Zvb42VYcjzjslLC8fLGGBNW8X01EBEXcB9wJlAKbBSR51XV3aXpW6r6nUGoMWhFHv9FoUUknGUYY0xYBNNDnwfsUNViVW0CVgMLB7es/mv1Kds91XaGqDEmZgUT6OOAPQGPS51lXZ0kIltE5CURmdXdhkTkShEpFJHCioqKAZTbs12VtdQ3t9r4uTEmZgUT6N2NX2iXxx8CE1R1NvBH4LnuNqSqy1W1QFULcnJy+lVoX9wemwPdGBPbggn0UmB8wOM8oCywgap6VbXGub8WSBCR7JBVGYQij5f4OGFybvqRfFljjBkyggn0jcAUEZkoIonAxcDzgQ1EZLQ4eyJFZJ6z3cpQF9sbd5mXo3PSSU5wHcmXNcaYIaPPo1xUtUVErgVeAVzAClXdJiJXO+sfABYB14hIC1APXKyqXYdlBlWRp5qTjh55JF/SGGOGlD4DHdqHUdZ2WfZAwP0/AX8KbWnBO1DbxF5vg50haoyJaUEF+lDXfoaozeFiIlRzczOlpaU0NDSEuxQzRCQnJ5OXl0dCQvAzx0ZFoLfNgW49dBOpSktLycjIID8/306MM6gqlZWVlJaWMnHixKCfFxVzuRR5vIwalsTI9KRwl2LMgDQ0NDBy5EgLcwOAiDBy5Mh+/8UWFYHudk75NyaSWZibQAP5fYj4QG9saWXHvho7ocgYE/MiPtA/L6+hxafWQzdmgCorK5kzZw5z5sxh9OjRjBs3rv1xU1NTr88tLCzkuuuu6/M1Tj755FCVC8D111/PuHHj8Pl8Id1upIv4naI2B7oxh2fkyJFs3rwZgGXLlpGens5NN93Uvr6lpYX4+O6joqCggIKCgj5f45133glJrQA+n481a9Ywfvx43nzzTebPnx+ybQdqbW3F5YqsExUjPtDdHi/JCXHkj0wLdynGhMRtL2xrP3IrVGaOHcZvv9vtnHndWrJkCSNGjOCjjz5i7ty5XHTRRSxdupT6+npSUlJ45JFHmDZtGuvXr+fuu+/mxRdfZNmyZezevZvi4mJ2797N0qVL23vv6enp1NTUsH79epYtW0Z2djZbt27l+OOP569//Ssiwtq1a7nxxhvJzs5m7ty5FBcX8+KLLx5S2xtvvMHXvvY1LrroIlatWtUe6OXl5Vx99dUUFxcDcP/993PyySezcuVK7r77bkSEY489lscee4wlS5bwne98h0WLFh1S32233caYMWPYvHkzbreb8847jz179tDQ0MD111/PlVdeCcDLL7/MLbfcQmtrK9nZ2bz22mtMmzaNd955h5ycHHw+H1OnTuW9994jO/vIzIQS8YFe5PEyffQwXHG2Q8mYUPrss89Yt24dLpcLr9fLm2++SXx8POvWreOWW27hmWeeOeQ527dv54033qC6uppp06ZxzTXXHHIc9UcffcS2bdsYO3Ysp5xyCm+//TYFBQVcddVVvPnmm0ycOJHFixf3WNeqVatYvHgxCxcu5JZbbqG5uZmEhASuu+46Tj/9dNasWUNrays1NTVs27aNO+64g7fffpvs7GwOHDjQ5/v+4IMP2Lp1a/vhgitWrGDEiBHU19dzwgkncMEFF+Dz+bjiiiva6z1w4ABxcXFccsklPP744yxdupR169Yxe/bsIxbmEOGBrqq4y7ycc+zYcJdiTMj0pyc9mC688ML2IYeqqiouvfRSPv/8c0SE5ubmbp9zzjnnkJSURFJSErm5uZSXl5OXl9epzbx589qXzZkzh5KSEtLT05k0aVJ7iC5evJjly5cfsv2mpibWrl3L73//ezIyMjjxxBN59dVXOeecc3j99ddZuXIlAC6Xi8zMTFauXMmiRYvaQ3XEiBF9vu958+Z1Ovb73nvvZc2aNQDs2bOHzz//nIqKCk477bT2dm3bvfzyy1m4cCFLly5lxYoVXHbZZX2+XihFdKCXVTXgbWix8XNjBkFaWscw5q233so3v/lN1qxZQ0lJSY/j1klJHeeCuFwuWlpagmoT7NRPL7/8MlVVVRxzzDEA1NXVkZqayjnnnNNte1Xt9vC/+Pj49h2qqtpp52/g+16/fj3r1q3j3XffJTU1lfnz59PQ0NDjdsePH8+oUaN4/fXXef/993n88ceDel+hEtFHubSNM860M0SNGVRVVVWMG+e/rs2jjz4a8u1Pnz6d4uJiSkpKAHjiiSe6bbdq1SoeeughSkpKKCkpYefOnbz66qvU1dVxxhlncP/99wP+HZper5czzjiDJ598kspK/+SvbUMu+fn5bNq0CYC//e1vPf7FUVVVxfDhw0lNTWX79u289957AJx00kls2LCBnTt3dtouwI9//GMuueQSvv/97x/xnaoRHehFHi8iMG209dCNGUy/+MUv+OUvf8kpp5xCa2tryLefkpLCn//8ZxYsWMCpp57KqFGjyMzsPDdTXV0dr7zySqfeeFpaGqeeeiovvPAC99xzD2+88QbHHHMMxx9/PNu2bWPWrFn86le/4vTTT2f27NnceOONAFxxxRVs2LCBefPm8f7773fqlQdasGABLS0tHHvssdx66618/etfByAnJ4fly5fzve99j9mzZ3PRRRe1P+fcc8+lpqbmiA+3AMgRnuW2XUFBgRYWFh7WNq5+bBPb93pZ//NvhqgqY8KjqKiIGTNmhLuMsKqpqSE9PR1V5ac//SlTpkzhhhtuCHdZ/VZYWMgNN9zAW2+9ddjb6u73QkQ2qWq3x4pGdg99r9fGz42JEg8++CBz5sxh1qxZVFVVcdVVV4W7pH678847ueCCC/jP//zPsLx+xPbQqxuaOWbZq/zbmVP52RlTQliZMUee9dBNd2Kmh/7p3mrAzhA1xpg2ERvobk/bHOgW6MYYAxEc6EUeL1mpCYzJTA53KcYYMyREbKC7y7zMGD3M5pA2xhhHRAZ6S6uP7XurbfzcmBCZP38+r7zySqdlf/jDH/jJT37S63PaDmz49re/zcGDBw9ps2zZMu6+++5eX/u5557D7Xa3P/7Nb37DunXr+lF972Jpqt2IDPSSyloaW3w2fm5MiCxevJjVq1d3WrZ69epeJ8kKtHbtWrKysgb02l0D/T/+4z/41re+NaBtddV1qt3BMhgnWw1ERM7l4vY4R7hYoJto9NLNsPeT0G5z9DFw9p09rl60aBG//vWvaWxsJCkpiZKSEsrKyjj11FO55ppr2LhxI/X19SxatIjbbrvtkOfn5+dTWFhIdnY2d9xxBytXrmT8+PHk5ORw/PHHA/7jzJcvX05TUxOTJ0/mscceY/PmzTz//PNs2LCB3/3udzzzzDPcfvvt7VPb/v3vf+emm26ipaWFE044gfvvv5+kpCTy8/O59NJLeeGFF2hubuapp55i+vTph9QVa1PtRmQP3V3mJcElTM5ND3cpxkSFkSNHMm/ePF5++WXA3zu/6KKLEBHuuOMOCgsL+fjjj9mwYQMff/xxj9vZtGkTq1ev5qOPPuLZZ59l48aN7eu+973vsXHjRrZs2cKMGTN4+OGHOfnkkzn33HO566672Lx5M0cffXR7+4aGBpYsWcITTzzBJ598QktLS/tcLQDZ2dl8+OGHXHPNNT0O67RNtXv++efz4osvts/Z0jbV7pYtW/jwww+ZNWtW+1S7r7/+Olu2bOGee+7p83P74IMPuOOOO9r/wlixYgWbNm2isLCQe++9l8rKSioqKrjiiit45pln2LJlC0899VSnqXaBkE21G1QPXUQWAPcALuAhVe32q15ETgDeAy5S1acPq7JeFHm8TM7NIDE+Ir+PjOldLz3pwdQ27LJw4UJWr17NihUrAHjyySdZvnw5LS0teDwe3G43xx57bLfbeOuttzj//PNJTU0F/POatNm6dSu//vWvOXjwIDU1NZx11lm91vPpp58yceJEpk6dCsCll17Kfffdx9KlSwH/FwTA8ccfz7PPPnvI82Nxqt0+A11EXMB9wJlAKbBRRJ5XVXc37f4LeOXQrYSW2+PlG1OO3KTxxsSC8847jxtvvJEPP/yQ+vp65s6dy86dO7n77rvZuHEjw4cPZ8mSJTQ0NPS6nZ6OPFuyZAnPPfccs2fP5tFHH2X9+vW9bqevs9jbpuHtaZreWJxqN5gu7jxgh6oWq2oTsBpY2E27nwHPAPsOu6peVFQ3UlHdaOPnxoRYeno68+fP5/LLL2/fGer1eklLSyMzM5Py8nJeeumlXrdx2mmnsWbNGurr66muruaFF15oX1ddXc2YMWNobm7uFF4ZGRlUV1cfsq3p06dTUlLCjh07AHjsscc4/fTTg34/sTjVbjCBPg7YE/C41FnWTkTGAecDD/S2IRG5UkQKRaSwoqKiv7UCAReFtkA3JuQWL17Mli1buPjiiwGYPXs2xx13HLNmzeLyyy/nlFNO6fX5bdcfnTNnDhdccAHf+MY32tfdfvvtnHjiiZx55pmddmBefPHF3HXXXRx33HF88cUX7cuTk5N55JFHuPDCCznmmGOIi4vj6quvDup9xOpUu31OziUiFwJnqeqPncc/AOap6s8C2jwF/LeqvicijwIv9jWGPtDJuQpLDvDAhmLuWnQsw9MS+/18Y4Yim5wrNvU11W5/J+cKZqdoKTA+4HEeUNalTQGw2hknyga+LSItqvpcENvvl4L8ETyU3/fOCmOMGcruvPNO7r///pBepi6YIZeNwBQRmSgiicDFwPOBDVR1oqrmq2o+8DTwk8EIc2OMiRY333wzu3bt4tRTTw3ZNvvsoatqi4hci//oFRewQlW3icjVzvpex82NMcHp6WgIE5sGcq2KoI5DV9W1wNouy7oNclVd0u8qjIlxycnJVFZWMnLkSAt1g6pSWVlJcnL/ZpONyFP/jYk2eXl5lJaWMtCjv0z0SU5OJi8vr1/PsUA3ZghISEjodMahMQNh584bY0yUsEA3xpgoYYFujDFRos8zRQfthUUqgF1hefHQyQb2h7uIIcQ+j87s8+hgn0Vnh/N5TFDVnO5WhC3Qo4GIFPZ0Cm4sss+jM/s8Othn0dlgfR425GKMMVHCAt0YY6KEBfrhWR7uAoYY+zw6s8+jg30WnQ3K52Fj6MYYEyWsh26MMVHCAt0YY6KEBfoAiMh4EXlDRIpEZJuIXB/umsJNRFwi8pGIvBjuWsJNRLJE5GkR2e78jpwU7prCSURucP6fbBWRVSLSvykEI5yIrBCRfSKyNWDZCBF5TUQ+d34OD8VrWaAPTAvwb6o6A/g68FMRmRnmmsLteqAo3EUMEfcAL6vqdGA2Mfy5ONcbvg4oUNWv4b+mwsXhreqIexRY0GXZzcDfVXUK8Hfn8WGzQB8AVfWo6ofO/Wr8/2HH9f6s6CUiecA5wEPhriXcRGQYcBrwMICqNqnqwbAWFX7xQIqIxAOpHHoJy6imqm8CB7osXgj8xbn/F+C8ULyWBfphEpF84Djg/TCXEk5/AH4B+MJcx1AwCagAHnGGoB4Ske4vEx8DVPVL4G5gN+ABqlT11fBWNSSMUlUP+DuIQG4oNmqBfhhEJB14Bliqqt5w1xMOIvIdYJ+qbgp3LUNEPDAXuF9VjwNqCdGf05HIGRteCEwExgJpInJJeKuKXhboAyQiCfjD/HFVfTbc9YTRKcC5IlICrAb+SUT+Gt6SwqoUKFXVtr/YnsYf8LHqW8BOVa1Q1WbgWeDkMNc0FJSLyBgA5+e+UGzUAn0AxH/Rx4eBIlX9P+GuJ5xU9Zeqmqeq+fh3dr2uqjHbA1PVvcAeEZnmLDoDcIexpHDbDXxdRFKd/zdnEMM7iQM8D1zq3L8U+FsoNmqXoBuYU4AfAJ+IyGZn2S3OxbSN+RnwuIgkAsXAZWGuJ2xU9X0ReRr4EP/RYR8RY9MAiMgqYD6QLSKlwG+BO4EnReRH+L/0LgzJa9mp/8YYEx1syMUYY6KEBboxxkQJC3RjjIkSFujGGBMlLNCNMSZKWKAbY0yUsEA3xpgo8f8Ber/rCaaNqLsAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "### visualizing the results data ###\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "\n",
    "\n",
    "for i in ['accuracy', 'precision_m', 'recall_m']:\n",
    "    acc = history.history[i]\n",
    "    val_acc = history.history['val_{}'.format(i)]\n",
    "    epochs = range(1, len(acc)+1)\n",
    "    \n",
    "    plt.figure()\n",
    "    plt.plot(epochs, acc, label='Training Accuracy')\n",
    "    plt.plot(epochs, val_acc, label='Validation Accuracy')\n",
    "    \n",
    "    plt.title(\"Results for {}\".format(i))\n",
    "    plt.legend()\n",
    "    \n",
    "    plt.show\n",
    "\n",
    "    \n",
    "### the plot will show that there is no significant rise in learning rate after epochs 2,3"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "37406432",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.10"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}