ML_classification.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "76b9fd56",
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import pickle\n",
    "#import torch\n",
    "#import transformers\n",
    "#from torch.utils.data import Dataset, DataLoader\n",
    "from sklearn.model_selection import train_test_split\n",
    "import sys\n",
    "import csv\n",
    "import re\n",
    "from tqdm.notebook import tqdm\n",
    "import sys\n",
    "import csv\n",
    "from keras.preprocessing.text import Tokenizer\n",
    "from keras.preprocessing.sequence import pad_sequences\n",
    "from nltk.corpus import stopwords\n",
    "from nltk import word_tokenize\n",
    "STOPWORDS = set(stopwords.words('english'))\n",
    "from bs4 import BeautifulSoup\n",
    "from src.preprocess import *\n",
    "from src.models import *\n",
    "from src.lib import *"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d0e79b42",
   "metadata": {},
   "source": [
    "## Explore data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "1781aeb4",
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "\n",
    "## Training and Testing files\n",
    "read_file=pd.read_csv('./data/IntentSDCNL_Training.csv', encoding = \"ISO-8859-1\")\n",
    "read_file1=pd.read_csv('./data/IntentSDCNL_Testing.csv', encoding = \"ISO-8859-1\")\n",
    "\n",
    "\n",
    "\n",
    "#training data\n",
    "df_train=pd.DataFrame(read_file,columns= ['selftext','ANNOTATIONS'])### \n",
    "df_train = df_train.rename(columns={'selftext': 'text','ANNOTATIONS':'label'})\n",
    "df_train= df_train.convert_dtypes()\n",
    "\n",
    "#testing data\n",
    "df_test=pd.DataFrame(read_file1,columns= ['selftext','ANNOTATIONS'])#### \n",
    "df_test = df_test.rename(columns={'selftext': 'text','ANNOTATIONS':'label'})\n",
    "df_test= df_test.convert_dtypes()\n",
    "\n",
    "## Sampling file\n",
    "df_samples=pd.read_csv('./data/sample.csv', encoding = \"ISO-8859-1\")\n",
    "\n",
    "## add samples to training file\n",
    "\n",
    "frame = [df_train, df_samples]\n",
    "df_train =  pd.concat(frame)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "a81b1b81",
   "metadata": {},
   "outputs": [],
   "source": [
    "## some preprocessings due to wrong entries\n",
    "df_test.drop(df_test[df_test.text==\"emptypost\"].index, inplace=True)\n",
    "df_train.drop(df_train[df_train.text==\"emptypost\"].index, inplace=True)\n",
    "df_test.drop(df_test[df_test.label==\" \"].index, inplace=True)\n",
    "df_train.drop(df_train[df_train.label==\" \"].index, inplace=True)\n",
    "df_test.dropna(inplace=True)\n",
    "df_train.dropna(inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "84a878dd",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "Int64Index: 1817 entries, 0 to 359\n",
      "Data columns (total 2 columns):\n",
      " #   Column  Non-Null Count  Dtype \n",
      "---  ------  --------------  ----- \n",
      " 0   text    1817 non-null   object\n",
      " 1   label   1817 non-null   Int64 \n",
      "dtypes: Int64(1), object(1)\n",
      "memory usage: 44.4+ KB\n",
      "<class 'pandas.core.frame.DataFrame'>\n",
      "Int64Index: 1817 entries, 0 to 359\n",
      "Data columns (total 2 columns):\n",
      " #   Column  Non-Null Count  Dtype \n",
      "---  ------  --------------  ----- \n",
      " 0   text    1817 non-null   object\n",
      " 1   label   1817 non-null   Int64 \n",
      "dtypes: Int64(1), object(1)\n",
      "memory usage: 44.4+ KB\n"
     ]
    }
   ],
   "source": [
    "df_train.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "ce4c734c",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "###### Class distribution of training data #####\n",
      "###### Class distribution of training data #####\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:ylabel='label'>"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:ylabel='label'>"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "print(\"###### Class distribution of training data #####\")\n",
    "#labels = [\"0\", \"1\", \"2\", \"3\", \"4\", \"5\"]\n",
    "fig, ax = plt.subplots(1, 1, figsize=(5,5))\n",
    "df_train['label'].value_counts().plot.pie(autopct = '%1.1f%%')#, labels=labels)\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "731db500",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "###### Class distribution of testing data \n",
      " #####\n",
      "###### Class distribution of testing data \n",
      " #####\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:ylabel='label'>"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:ylabel='label'>"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "\n",
    "print(\"###### Class distribution of testing data \\n #####\")\n",
    "#labels = [\"0\", \"1\", \"2\", \"3\", \"4\", \"5\"]\n",
    "fig, ax = plt.subplots(1, 1, figsize=(5,5))\n",
    "df_test['label'].value_counts().plot.pie(autopct = '%1.1f%%')#, labels=labels)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "cca269e9",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "5    92\n",
       "4    91\n",
       "0    61\n",
       "2    48\n",
       "3    43\n",
       "1    35\n",
       "Name: label, dtype: Int64"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "text/plain": [
       "5    92\n",
       "4    91\n",
       "0    61\n",
       "2    48\n",
       "3    43\n",
       "1    35\n",
       "Name: label, dtype: Int64"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_test['label'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "6fd664e7",
   "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>label</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Indubitably, I am not worth anything as a bein...</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Almost everyday I wake up hating myself, why w...</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>To kill myself if im ugly?</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>God help me.... I know I should go to the hosp...</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>I really don't want to do anything anymore. Wh...</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>355</th>\n",
       "      <td>IÃ¢ÂÂve posted here a lot over the last year...</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>356</th>\n",
       "      <td>JUst need to ramble a little. Over 30 years of...</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>357</th>\n",
       "      <td>I think the relationship btw money and ur ment...</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>358</th>\n",
       "      <td>Hi everyone. I just wanted to clear something ...</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>359</th>\n",
       "      <td>They follow me everywhere I go</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>2187 rows × 2 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                  text  label\n",
       "0    Indubitably, I am not worth anything as a bein...      1\n",
       "1    Almost everyday I wake up hating myself, why w...      5\n",
       "2                           To kill myself if im ugly?      1\n",
       "3    God help me.... I know I should go to the hosp...      1\n",
       "4    I really don't want to do anything anymore. Wh...      2\n",
       "..                                                 ...    ...\n",
       "355  IÃ¢ÂÂve posted here a lot over the last year...      3\n",
       "356  JUst need to ramble a little. Over 30 years of...      2\n",
       "357  I think the relationship btw money and ur ment...      2\n",
       "358  Hi everyone. I just wanted to clear something ...      3\n",
       "359                     They follow me everywhere I go      1\n",
       "\n",
       "[2187 rows x 2 columns]"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "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>label</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Indubitably, I am not worth anything as a bein...</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Almost everyday I wake up hating myself, why w...</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>To kill myself if im ugly?</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>God help me.... I know I should go to the hosp...</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>I really don't want to do anything anymore. Wh...</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>355</th>\n",
       "      <td>IÃ¢ÂÂve posted here a lot over the last year...</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>356</th>\n",
       "      <td>JUst need to ramble a little. Over 30 years of...</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>357</th>\n",
       "      <td>I think the relationship btw money and ur ment...</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>358</th>\n",
       "      <td>Hi everyone. I just wanted to clear something ...</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>359</th>\n",
       "      <td>They follow me everywhere I go</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>2187 rows × 2 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                  text  label\n",
       "0    Indubitably, I am not worth anything as a bein...      1\n",
       "1    Almost everyday I wake up hating myself, why w...      5\n",
       "2                           To kill myself if im ugly?      1\n",
       "3    God help me.... I know I should go to the hosp...      1\n",
       "4    I really don't want to do anything anymore. Wh...      2\n",
       "..                                                 ...    ...\n",
       "355  IÃ¢ÂÂve posted here a lot over the last year...      3\n",
       "356  JUst need to ramble a little. Over 30 years of...      2\n",
       "357  I think the relationship btw money and ur ment...      2\n",
       "358  Hi everyone. I just wanted to clear something ...      3\n",
       "359                     They follow me everywhere I go      1\n",
       "\n",
       "[2187 rows x 2 columns]"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "frame = [df_test, df_train]\n",
    "data =  pd.concat(frame)\n",
    "data"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "44b85644",
   "metadata": {},
   "source": [
    "#### A Sample"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "be17c2c9",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "i have been so depressed these past couple weeks ever since i got back from vacation normal life and reality has hit me and its terrible. im 19 i havent moved out yet i dont really have a sufficient job i have one but its not enough money for me to support myself. at least i have a car and im taking online classes so. i feel stuck in this stupid fucking town. empty and alone man i just want someone to love me dude but im ugly and fat and worthless and no one is ever going to see im worth the chance. its just one thing after another to worry about. why not end it now before i really fuck up and become miserable? i dont see the point in going on im just really down right now i dont know what to do )): all there is darkness idk how to get out of this hole i hate myself i hate my body i hate my life i dont think it will ever get better\n",
      "text: 1\n",
      "i have been so depressed these past couple weeks ever since i got back from vacation normal life and reality has hit me and its terrible. im 19 i havent moved out yet i dont really have a sufficient job i have one but its not enough money for me to support myself. at least i have a car and im taking online classes so. i feel stuck in this stupid fucking town. empty and alone man i just want someone to love me dude but im ugly and fat and worthless and no one is ever going to see im worth the chance. its just one thing after another to worry about. why not end it now before i really fuck up and become miserable? i dont see the point in going on im just really down right now i dont know what to do )): all there is darkness idk how to get out of this hole i hate myself i hate my body i hate my life i dont think it will ever get better\n",
      "text: 1\n"
     ]
    }
   ],
   "source": [
    "print_plot(df_train,0)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "72b8bf9b",
   "metadata": {},
   "source": [
    "## Clean and preprocess the text"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "440f4d70",
   "metadata": {},
   "source": [
    "####  Clean Training and Testing data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "6e2af759",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/var/folders/36/14vf2r1n2q92gt0lb6xj29nm0000gn/T/ipykernel_16645/1264519194.py:21: FutureWarning: The default value of regex will change from True to False in a future version.\n",
      "  df_test['text'] = df_test['text'].str.replace('\\d+', '')\n",
      "/var/folders/36/14vf2r1n2q92gt0lb6xj29nm0000gn/T/ipykernel_16645/1264519194.py:21: FutureWarning: The default value of regex will change from True to False in a future version.\n",
      "  df_test['text'] = df_test['text'].str.replace('\\d+', '')\n"
     ]
    }
   ],
   "source": [
    "REPLACE_BY_SPACE_RE = re.compile('[/(){}\\[\\]\\|@,;]')\n",
    "BAD_SYMBOLS_RE = re.compile('[^0-9a-z #+_]')\n",
    "STOPWORDS = set(stopwords.words('english'))\n",
    "\n",
    "def clean_text(text):\n",
    "    text = text.lower() \n",
    "    text = REPLACE_BY_SPACE_RE.sub(' ', text) \n",
    "    text = BAD_SYMBOLS_RE.sub('', text)     \n",
    "    text = text.replace('x', '')\n",
    "    text = ' '.join(word for word in text.split() if word not in STOPWORDS) # remove stopwors from text\n",
    "    return text\n",
    "\n",
    "\n",
    "#training data\n",
    "df_train = df_train.reset_index(drop=True)\n",
    "df_train['text'] = df_train['text'].apply(clean_text)\n",
    "df_train['text'] = df_train['text'].apply(clean_text)\n",
    "##testing data\n",
    "df_train = df_train.reset_index(drop=True)\n",
    "df_test['text'] = df_test['text'].apply(clean_text)\n",
    "df_test['text'] = df_test['text'].str.replace('\\d+', '')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4b991f54",
   "metadata": {},
   "source": [
    "#### A sample after cleanup"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "46136ca9",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "depressed past couple weeks ever since got back vacation normal life reality hit terrible im 19 havent moved yet dont really sufficient job one enough money support least car im taking online classes feel stuck stupid fucking town empty alone man want someone love dude im ugly fat worthless one ever going see im worth chance one thing another worry end really fuck become miserable dont see point going im really right dont know darkness idk get hole hate hate body hate life dont think ever get better\n",
      "text: 1\n",
      "depressed past couple weeks ever since got back vacation normal life reality hit terrible im 19 havent moved yet dont really sufficient job one enough money support least car im taking online classes feel stuck stupid fucking town empty alone man want someone love dude im ugly fat worthless one ever going see im worth chance one thing another worry end really fuck become miserable dont see point going im really right dont know darkness idk get hole hate hate body hate life dont think ever get better\n",
      "text: 1\n"
     ]
    }
   ],
   "source": [
    "print_plot(df_train,0)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e1a980db",
   "metadata": {},
   "source": [
    "#### Intialize parameters"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "9e5dd333",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Define some Global Variables\n",
    "max_nb_words = 60000\n",
    "max_features = 10543 # Maximum Number of words we want to include in our dictionary\n",
    "maxlen = 500 # No of words in question we want to create a sequence with\n",
    "emb_dim = 100# Size of word to vec embedding we are usingtokenizer = Tokenizer(num_words=max_nb_words, filters='!\"#$%&()*+,-./:;<=>?@[\\]^_`{|}~', lower=True)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d50a645a",
   "metadata": {},
   "source": [
    "#### Initial Tokenizer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "354197dc",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 14692 unique tokens.\n",
      "Found 14692 unique tokens.\n"
     ]
    }
   ],
   "source": [
    "tokenizer = Tokenizer(num_words=max_nb_words, filters='!\"#$%&()*+,-./:;<=>?@[\\]^_`{|}~', lower=True)\n",
    "tokenizer.fit_on_texts(data['text'].values)# tokenizer on whole training and testing data\n",
    "word_index = tokenizer.word_index\n",
    "print('Found %s unique tokens.' % len(word_index))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "4a4e6cf3",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Shape of data tensor: (1817, 500)\n",
      "Shape of data tensor: (1817, 500)\n"
     ]
    }
   ],
   "source": [
    "X_train = tokenizer.texts_to_sequences(df_train['text'].values)\n",
    "X_train = pad_sequences(X_train, maxlen=maxlen)\n",
    "print('Shape of data tensor:', X_train.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "bdd114a1",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Shape of label tensor: (1817, 6)\n",
      "Shape of label tensor: (1817, 6)\n"
     ]
    }
   ],
   "source": [
    "Y_train = pd.get_dummies(df_train['label']).values\n",
    "print('Shape of label tensor:', Y_train.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "b2c6178b",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Shape of data tensor: (370, 500)\n",
      "Shape of data tensor: (370, 500)\n"
     ]
    }
   ],
   "source": [
    "X_test = tokenizer.texts_to_sequences(df_test['text'].values)\n",
    "X_test = pad_sequences(X_test, maxlen=maxlen)\n",
    "print('Shape of data tensor:', X_test.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "8fcc9864",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Shape of label tensor: (370, 6)\n",
      "Shape of label tensor: (370, 6)\n"
     ]
    }
   ],
   "source": [
    "Y_test = pd.get_dummies(df_test['label']).values\n",
    "print('Shape of label tensor:', Y_test.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4daadd61",
   "metadata": {},
   "source": [
    "### Finally some validation data from Training Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "749fdee7",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train, X_val, Y_train, Y_val = train_test_split(X_train,Y_train,random_state=0,test_size=0.1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "7170b77e",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "TrainingData (1635, 500) (1635, 6)\n",
      "TestingData (370, 500) (370, 6)\n",
      "ValidationData (182, 500) (182, 6)\n",
      "TrainingData (1635, 500) (1635, 6)\n",
      "TestingData (370, 500) (370, 6)\n",
      "ValidationData (182, 500) (182, 6)\n"
     ]
    }
   ],
   "source": [
    "print('TrainingData',X_train.shape,Y_train.shape)\n",
    "print('TestingData',X_test.shape,Y_test.shape)\n",
    "print('ValidationData',X_val.shape,Y_val.shape)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "4464a0fd",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "test example: [   0    0    0    0    0    0    0    0    0    0    0    0    0    0\n",
      "    0    0    0    0    0    0    0    0    0    0    0    0    0    0\n",
      "    0    0    0    0    0    0    0    0    0    0    0    0    0    0\n",
      "    0    0    0    0    0    0    0    0    0    0    0    0    0    0\n",
      "    0    0    0    0    0    0    0    0    0    0    0    0    0    0\n",
      "    0    0    0    0    0    0    0    0    0    0    0    0    0    0\n",
      "    0    0    0    0    0    0    0    0    0    0    0    0    0    0\n",
      "    0    0    0    0    0    0    0    0    0    0    0    0    0    0\n",
      "    0    0    0    0    0    0    0    0    0    0    0    0    0    0\n",
      "    0    0    0    0    0    0    0    0    0    0    0    0    0    0\n",
      "    0    0    0    0    0    0    0    0    0    0    0    0    0    0\n",
      "    0    0    0    0    0    0    0    0    0    0    0    0    0    0\n",
      "    0    0    0    0    0    0    0    0    0    0    0    0    0    0\n",
      "    0    0    0    0    0    0    0    0    0    0    0    0    0    0\n",
      "    0    0    0    0    0    0    0    0    0    0    0    0    0    0\n",
      "    0    0    0    0    0    0    0    0    0    0    0    0    0    0\n",
      "    0    0    0    0    0    0    0    0    0    0    0    0    0  113\n",
      "  352  130   60  163  269  361  163  139  503  179  179  253 2221  286\n",
      "  170   26  145 1105 1174  711   83  190  145  498   99 1680  185  209\n",
      "  582 1593  136  582  228  380  159  269  361  929   46  215 1187  190\n",
      "  145 1519  200  114  244  602  286  279  113  352  176 1902  113 5883\n",
      "   60 1059  511 1558  127  157   68  136  136   63  122  136   95 2352\n",
      "   54  279  279  457  230   68  362  369  240  214  457   68 1964  159\n",
      "  181  185  113   22  504  309   36  296  251  169 1362 4048  126  214\n",
      "  682  425   40 1903 8030  590  268 1808  233  595  107 1396   70  438\n",
      "   95  402   55  503  241  149  503   26   22  269 1681  503  191   69\n",
      " 1352 3181  505  163  180 1298  235  365  200   36 5314  190   22  169\n",
      "   68  139   26   94   92  414 1069 1353  117  233   94  185   26   22\n",
      "  132  193  804  387  130   71  782  602  286  244  602 3154 1362  309\n",
      "  191  803  149  210  174  151   60  247   85   46  237  664  167  241\n",
      "  195  195   68  114  356  130   34  111  308 7327  200  200  692  130\n",
      "  110  167  139  286  169   33   98  256  139   98  359   34  267  692\n",
      " 3566 1049  126  114  111 8031  174  211  768  176 1970 3567 8032   60\n",
      "   63  111 2099  408  176 1148 1311   60  924  176 1970  290  111  151\n",
      "   55 3182  644  328  107  122 1410  170  119   26 1061  794   26   55\n",
      "  122   46  163  139  190 1415  172  185  110   95]\n",
      "test example: [   0    0    0    0    0    0    0    0    0    0    0    0    0    0\n",
      "    0    0    0    0    0    0    0    0    0    0    0    0    0    0\n",
      "    0    0    0    0    0    0    0    0    0    0    0    0    0    0\n",
      "    0    0    0    0    0    0    0    0    0    0    0    0    0    0\n",
      "    0    0    0    0    0    0    0    0    0    0    0    0    0    0\n",
      "    0    0    0    0    0    0    0    0    0    0    0    0    0    0\n",
      "    0    0    0    0    0    0    0    0    0    0    0    0    0    0\n",
      "    0    0    0    0    0    0    0    0    0    0    0    0    0    0\n",
      "    0    0    0    0    0    0    0    0    0    0    0    0    0    0\n",
      "    0    0    0    0    0    0    0    0    0    0    0    0    0    0\n",
      "    0    0    0    0    0    0    0    0    0    0    0    0    0    0\n",
      "    0    0    0    0    0    0    0    0    0    0    0    0    0    0\n",
      "    0    0    0    0    0    0    0    0    0    0    0    0    0    0\n",
      "    0    0    0    0    0    0    0    0    0    0    0    0    0    0\n",
      "    0    0    0    0    0    0    0    0    0    0    0    0    0    0\n",
      "    0    0    0    0    0    0    0    0    0    0    0    0    0    0\n",
      "    0    0    0    0    0    0    0    0    0    0    0    0    0  113\n",
      "  352  130   60  163  269  361  163  139  503  179  179  253 2221  286\n",
      "  170   26  145 1105 1174  711   83  190  145  498   99 1680  185  209\n",
      "  582 1593  136  582  228  380  159  269  361  929   46  215 1187  190\n",
      "  145 1519  200  114  244  602  286  279  113  352  176 1902  113 5883\n",
      "   60 1059  511 1558  127  157   68  136  136   63  122  136   95 2352\n",
      "   54  279  279  457  230   68  362  369  240  214  457   68 1964  159\n",
      "  181  185  113   22  504  309   36  296  251  169 1362 4048  126  214\n",
      "  682  425   40 1903 8030  590  268 1808  233  595  107 1396   70  438\n",
      "   95  402   55  503  241  149  503   26   22  269 1681  503  191   69\n",
      " 1352 3181  505  163  180 1298  235  365  200   36 5314  190   22  169\n",
      "   68  139   26   94   92  414 1069 1353  117  233   94  185   26   22\n",
      "  132  193  804  387  130   71  782  602  286  244  602 3154 1362  309\n",
      "  191  803  149  210  174  151   60  247   85   46  237  664  167  241\n",
      "  195  195   68  114  356  130   34  111  308 7327  200  200  692  130\n",
      "  110  167  139  286  169   33   98  256  139   98  359   34  267  692\n",
      " 3566 1049  126  114  111 8031  174  211  768  176 1970 3567 8032   60\n",
      "   63  111 2099  408  176 1148 1311   60  924  176 1970  290  111  151\n",
      "   55 3182  644  328  107  122 1410  170  119   26 1061  794   26   55\n",
      "  122   46  163  139  190 1415  172  185  110   95]\n"
     ]
    }
   ],
   "source": [
    "#print a test sample\n",
    "print('test example:', X_test[100])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "c8539e19",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "500"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "text/plain": [
       "500"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    " len(X_test[100])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c3f8e9b0",
   "metadata": {},
   "source": [
    "#### Intialize Embedding"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "ce41274e",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Size of vocabulary in GloVe: 400000\n",
      "Size of vocabulary in GloVe: 400000\n"
     ]
    }
   ],
   "source": [
    "embedding_vectors = {}\n",
    "\n",
    "with open('./embeddings/glove.6B.100d.txt','r',encoding='utf-8') as file:\n",
    "\n",
    "    for row in file:\n",
    "        values = row.split(' ')\n",
    "        word = values[0]\n",
    "        weights = np.asarray([float(val) for val in values[1:]])\n",
    "        embedding_vectors[word] = weights\n",
    "print(f\"Size of vocabulary in GloVe: {len(embedding_vectors)}\")   "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "f1cc5f7f",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "A few out of valubulary words: [\"i'm\", \"don't\", \"i've\", \"can't\", \"it's\"]\n",
      "A few out of valubulary words: [\"i'm\", \"don't\", \"i've\", \"can't\", \"it's\"]\n"
     ]
    }
   ],
   "source": [
    "#initialize the embedding_matrix with zeros\n",
    "emb_dim = 100\n",
    "if max_nb_words is not None: \n",
    "    vocab_len = max_nb_words\n",
    "else:\n",
    "    vocab_len = len(word_index)+1\n",
    "embedding_matrix = np.zeros((vocab_len, emb_dim))\n",
    "oov_count = 0\n",
    "oov_words = []\n",
    "for word, idx in word_index.items():\n",
    "    if idx < vocab_len:\n",
    "        embedding_vector = embedding_vectors.get(word)\n",
    "        if embedding_vector is not None:\n",
    "            embedding_matrix[idx] = embedding_vector\n",
    "        else:\n",
    "            oov_count += 1 \n",
    "            oov_words.append(word)\n",
    "#Check some of the out of vocabulary words\n",
    "print(f'A few out of valubulary words: {oov_words[0:5]}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "9286940a",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "3015 out of 60000 words were OOV.\n",
      "3015 out of 60000 words were OOV.\n"
     ]
    }
   ],
   "source": [
    "print(f'{oov_count} out of {vocab_len} words were OOV.')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d0a6a8c1",
   "metadata": {},
   "source": [
    "## NN models"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "6ce9711f",
   "metadata": {},
   "outputs": [],
   "source": [
    "##Global variables\n",
    "batch_size = 256\n",
    "epochs  = 20"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9330e923",
   "metadata": {},
   "source": [
    "### a) LSTM Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "c946bd87",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2022-01-08 13:43:48.128516: I tensorflow/core/platform/cpu_feature_guard.cc:151] This TensorFlow binary is optimized with oneAPI Deep Neural Network Library (oneDNN) to use the following CPU instructions in performance-critical operations:  AVX2 FMA\n",
      "To enable them in other operations, rebuild TensorFlow with the appropriate compiler flags.\n",
      "2022-01-08 13:43:48.128516: I tensorflow/core/platform/cpu_feature_guard.cc:151] This TensorFlow binary is optimized with oneAPI Deep Neural Network Library (oneDNN) to use the following CPU instructions in performance-critical operations:  AVX2 FMA\n",
      "To enable them in other operations, rebuild TensorFlow with the appropriate compiler flags.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " embedding (Embedding)       (None, None, 100)         6000000   \n",
      "                                                                 \n",
      " lstm (LSTM)                 (None, 128)               117248    \n",
      "                                                                 \n",
      " dropout (Dropout)           (None, 128)               0         \n",
      "                                                                 \n",
      " dense (Dense)               (None, 6)                 774       \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 6,118,022\n",
      "Trainable params: 118,022\n",
      "Non-trainable params: 6,000,000\n",
      "_________________________________________________________________\n",
      "None\n",
      "Model: \"sequential\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " embedding (Embedding)       (None, None, 100)         6000000   \n",
      "                                                                 \n",
      " lstm (LSTM)                 (None, 128)               117248    \n",
      "                                                                 \n",
      " dropout (Dropout)           (None, 128)               0         \n",
      "                                                                 \n",
      " dense (Dense)               (None, 6)                 774       \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 6,118,022\n",
      "Trainable params: 118,022\n",
      "Non-trainable params: 6,000,000\n",
      "_________________________________________________________________\n",
      "None\n"
     ]
    }
   ],
   "source": [
    "lstm_model = LSTM_model(embedding_matrix,vocab_len,emb_dim)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "30bbf47c",
   "metadata": {},
   "source": [
    "#### Train the model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "4ce8e228",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/5\n",
      "Epoch 1/5\n",
      "7/7 [==============================] - 12s 1s/step - loss: 1.8192 - accuracy: 0.1670 - val_loss: 1.7765 - val_accuracy: 0.2198\n",
      "7/7 [==============================] - 12s 1s/step - loss: 1.8192 - accuracy: 0.1670 - val_loss: 1.7765 - val_accuracy: 0.2198\n",
      "Epoch 2/5\n",
      "Epoch 2/5\n",
      "7/7 [==============================] - 11s 2s/step - loss: 1.7795 - accuracy: 0.2018 - val_loss: 1.7731 - val_accuracy: 0.1923\n",
      "7/7 [==============================] - 11s 2s/step - loss: 1.7795 - accuracy: 0.2018 - val_loss: 1.7731 - val_accuracy: 0.1923\n",
      "Epoch 3/5\n",
      "Epoch 3/5\n",
      "7/7 [==============================] - 16s 2s/step - loss: 1.7640 - accuracy: 0.2245 - val_loss: 1.7564 - val_accuracy: 0.2033\n",
      "7/7 [==============================] - 16s 2s/step - loss: 1.7640 - accuracy: 0.2245 - val_loss: 1.7564 - val_accuracy: 0.2033\n",
      "Epoch 4/5\n",
      "Epoch 4/5\n",
      "7/7 [==============================] - 11s 2s/step - loss: 1.7386 - accuracy: 0.2477 - val_loss: 1.7320 - val_accuracy: 0.2527\n",
      "7/7 [==============================] - 11s 2s/step - loss: 1.7386 - accuracy: 0.2477 - val_loss: 1.7320 - val_accuracy: 0.2527\n",
      "Epoch 5/5\n",
      "Epoch 5/5\n",
      "7/7 [==============================] - 15s 2s/step - loss: 1.6885 - accuracy: 0.3046 - val_loss: 1.6238 - val_accuracy: 0.3571\n",
      "7/7 [==============================] - 15s 2s/step - loss: 1.6885 - accuracy: 0.3046 - val_loss: 1.6238 - val_accuracy: 0.3571\n",
      "CPU times: user 1min 47s, sys: 25.3 s, total: 2min 12s\n",
      "Wall time: 1min 5s\n",
      "CPU times: user 1min 47s, sys: 25.3 s, total: 2min 12s\n",
      "Wall time: 1min 5s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "history = lstm_model.fit(X_train, np.asarray(Y_train), validation_data=(X_val, np.asarray(Y_val)), batch_size = batch_size, epochs = epochs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "1c327305",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x504 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#plot accuracy\n",
    "plot_accuracy(history,epochs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "0e3daa6a",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train accuracy: 36.76\n",
      "Test accuracy: 38.11\n",
      "Train accuracy: 36.76\n",
      "Test accuracy: 38.11\n"
     ]
    }
   ],
   "source": [
    "# Accuracy\n",
    "train_test_accuracy(lstm_model,X_train,Y_train,X_test,Y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "1b5cfae7",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.50      0.61      0.55        61\n",
      "           1       0.40      0.11      0.18        35\n",
      "           2       0.30      0.56      0.39        48\n",
      "           3       0.19      0.16      0.18        43\n",
      "           4       0.45      0.38      0.41        91\n",
      "           5       0.37      0.34      0.35        92\n",
      "\n",
      "    accuracy                           0.38       370\n",
      "   macro avg       0.37      0.36      0.34       370\n",
      "weighted avg       0.39      0.38      0.37       370\n",
      "\n",
      "accuracy: 0.381\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.50      0.61      0.55        61\n",
      "           1       0.40      0.11      0.18        35\n",
      "           2       0.30      0.56      0.39        48\n",
      "           3       0.19      0.16      0.18        43\n",
      "           4       0.45      0.38      0.41        91\n",
      "           5       0.37      0.34      0.35        92\n",
      "\n",
      "    accuracy                           0.38       370\n",
      "   macro avg       0.37      0.36      0.34       370\n",
      "weighted avg       0.39      0.38      0.37       370\n",
      "\n",
      "accuracy: 0.381\n"
     ]
    }
   ],
   "source": [
    "#Classification report\n",
    "report_nn(lstm_model,X_test,Y_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "57ba0f5c",
   "metadata": {},
   "source": [
    "### b) CNN"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "4958a265",
   "metadata": {},
   "outputs": [],
   "source": [
    "cnn_model=CNN_model(embedding_matrix,vocab_len,emb_dim)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f07a61c1",
   "metadata": {},
   "source": [
    "#### Train the model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "7b2178e8",
   "metadata": {},
   "outputs": [],
   "source": [
    "%%time\n",
    "history = cnn_model.fit(X_train, np.asarray(Y_train), validation_data=(X_val, np.asarray(Y_val)),batch_size = batch_size, epochs = epochs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "bf65a127",
   "metadata": {},
   "outputs": [],
   "source": [
    "#plot accuracy\n",
    "plot_accuracy(history,epochs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "6facdda7",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Accuracy\n",
    "train_test_accuracy(cnn_model,X_train,Y_train,X_test,Y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "284f0d5c",
   "metadata": {},
   "outputs": [],
   "source": [
    "#Classification report\n",
    "report_nn(cnn_model,X_test,Y_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "41211722",
   "metadata": {},
   "source": [
    "### d) GRU"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "f3f75416",
   "metadata": {},
   "outputs": [],
   "source": [
    "gru_model = GRU_model(embedding_matrix,vocab_len,emb_dim)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "93f1934e",
   "metadata": {},
   "outputs": [],
   "source": [
    "## tranining\n",
    "%%time\n",
    "history = gru_model.fit(X_train, np.asarray(Y_train), validation_data=(X_val, np.asarray(Y_val)), batch_size = batch_size, epochs = epochs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "1f800d30",
   "metadata": {},
   "outputs": [],
   "source": [
    "#plot accuracy\n",
    "plot_accuracy(history,epochs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "798392d3",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Accuracy\n",
    "train_test_accuracy(gru_model,X_train,Y_train,X_test,Y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "3346397e",
   "metadata": {},
   "outputs": [],
   "source": [
    "#Classification report\n",
    "report_nn(gru_model,X_test,Y_test)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2bf3e6f3",
   "metadata": {},
   "source": [
    "## e) CNN+GRU"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "31cc6538",
   "metadata": {},
   "outputs": [],
   "source": [
    "cnn_gru_model =CNN_GRU_model(embedding_matrix,vocab_len,emb_dim)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "4217fb1d",
   "metadata": {},
   "outputs": [],
   "source": [
    "## tranining\n",
    "%%time\n",
    "history=cnn_gru_model.fit(X_train, np.asarray(Y_train), validation_data=(X_val, np.asarray(Y_val)),epochs=epochs, batch_size=batch_size)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "b447a9ab",
   "metadata": {},
   "outputs": [],
   "source": [
    "#plot accuracy\n",
    "plot_accuracy(history,epochs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "3917939f",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Accuracy\n",
    "train_test_accuracy(cnn_gru_model,X_train,Y_train,X_test,Y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "e65751a8",
   "metadata": {},
   "outputs": [],
   "source": [
    "#Classification report\n",
    "report_nn(cnn_gru_model,X_test,Y_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "eca4da46",
   "metadata": {},
   "source": [
    "### f) CNN+LSTM"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "3124034e",
   "metadata": {},
   "outputs": [],
   "source": [
    "cnn_lstm_model = CNN_LSTM_model(embedding_matrix,vocab_len,emb_dim)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 103,
   "id": "a71d3b12",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/25\n",
      "7/7 [==============================] - 16s 2s/step - loss: 1.7746 - accuracy: 0.2183 - val_loss: 1.7336 - val_accuracy: 0.2473\n",
      "Epoch 2/25\n",
      "7/7 [==============================] - 11s 2s/step - loss: 1.7192 - accuracy: 0.2550 - val_loss: 1.6896 - val_accuracy: 0.2802\n",
      "Epoch 3/25\n",
      "7/7 [==============================] - 16s 2s/step - loss: 1.6660 - accuracy: 0.2936 - val_loss: 1.6104 - val_accuracy: 0.3462\n",
      "Epoch 4/25\n",
      "7/7 [==============================] - 13s 2s/step - loss: 1.6132 - accuracy: 0.3284 - val_loss: 1.5845 - val_accuracy: 0.3407\n",
      "Epoch 5/25\n",
      "7/7 [==============================] - 13s 2s/step - loss: 1.5725 - accuracy: 0.3621 - val_loss: 1.6001 - val_accuracy: 0.3242\n",
      "Epoch 6/25\n",
      "7/7 [==============================] - 14s 2s/step - loss: 1.5185 - accuracy: 0.4024 - val_loss: 1.5440 - val_accuracy: 0.4066\n",
      "Epoch 7/25\n",
      "7/7 [==============================] - 12s 2s/step - loss: 1.4281 - accuracy: 0.4495 - val_loss: 1.5644 - val_accuracy: 0.4066\n",
      "Epoch 8/25\n",
      "7/7 [==============================] - 12s 2s/step - loss: 1.4005 - accuracy: 0.4489 - val_loss: 1.5809 - val_accuracy: 0.3571\n",
      "Epoch 9/25\n",
      "7/7 [==============================] - 12s 2s/step - loss: 1.3415 - accuracy: 0.4789 - val_loss: 1.5764 - val_accuracy: 0.3846\n",
      "Epoch 10/25\n",
      "7/7 [==============================] - 12s 2s/step - loss: 1.2864 - accuracy: 0.5089 - val_loss: 1.6248 - val_accuracy: 0.3791\n",
      "Epoch 11/25\n",
      "7/7 [==============================] - 13s 2s/step - loss: 1.2497 - accuracy: 0.5248 - val_loss: 1.5621 - val_accuracy: 0.4231\n",
      "Epoch 12/25\n",
      "7/7 [==============================] - 11s 2s/step - loss: 1.2486 - accuracy: 0.5211 - val_loss: 1.5435 - val_accuracy: 0.3956\n",
      "Epoch 13/25\n",
      "7/7 [==============================] - 13s 2s/step - loss: 1.2235 - accuracy: 0.5309 - val_loss: 1.4728 - val_accuracy: 0.4286\n",
      "Epoch 14/25\n",
      "7/7 [==============================] - 12s 2s/step - loss: 1.1824 - accuracy: 0.5590 - val_loss: 1.6231 - val_accuracy: 0.4121\n",
      "Epoch 15/25\n",
      "7/7 [==============================] - 13s 2s/step - loss: 1.1525 - accuracy: 0.5492 - val_loss: 1.5729 - val_accuracy: 0.3846\n",
      "Epoch 16/25\n",
      "7/7 [==============================] - 12s 2s/step - loss: 1.1172 - accuracy: 0.5731 - val_loss: 1.5940 - val_accuracy: 0.3846\n",
      "Epoch 17/25\n",
      "7/7 [==============================] - 11s 2s/step - loss: 1.1023 - accuracy: 0.5951 - val_loss: 1.6433 - val_accuracy: 0.3626\n",
      "Epoch 18/25\n",
      "7/7 [==============================] - 11s 2s/step - loss: 1.0467 - accuracy: 0.5976 - val_loss: 1.6533 - val_accuracy: 0.3956\n",
      "Epoch 19/25\n",
      "7/7 [==============================] - 12s 2s/step - loss: 1.0499 - accuracy: 0.5982 - val_loss: 1.6312 - val_accuracy: 0.4341\n",
      "Epoch 20/25\n",
      "7/7 [==============================] - 11s 2s/step - loss: 1.0363 - accuracy: 0.6177 - val_loss: 1.6564 - val_accuracy: 0.4011\n",
      "Epoch 21/25\n",
      "7/7 [==============================] - 12s 2s/step - loss: 0.9587 - accuracy: 0.6489 - val_loss: 1.6681 - val_accuracy: 0.4066\n",
      "Epoch 22/25\n",
      "7/7 [==============================] - 13s 2s/step - loss: 0.9030 - accuracy: 0.6636 - val_loss: 1.8701 - val_accuracy: 0.3626\n",
      "Epoch 23/25\n",
      "7/7 [==============================] - 12s 2s/step - loss: 0.9216 - accuracy: 0.6593 - val_loss: 1.8432 - val_accuracy: 0.3516\n",
      "Epoch 24/25\n",
      "7/7 [==============================] - 11s 2s/step - loss: 0.8941 - accuracy: 0.6709 - val_loss: 1.7294 - val_accuracy: 0.4121\n",
      "Epoch 25/25\n",
      "7/7 [==============================] - 12s 2s/step - loss: 0.8725 - accuracy: 0.6691 - val_loss: 1.6837 - val_accuracy: 0.3956\n"
     ]
    }
   ],
   "source": [
    "## tranining\n",
    "%%time\n",
    "history=cnn_lstm_model.fit(X_train, np.asarray(Y_train), validation_data=(X_val, np.asarray(Y_val)),epochs=epochs, batch_size=batch_size)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "id": "998f7586",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train accuracy: 70.21\n",
      "Test accuracy: 47.84\n"
     ]
    }
   ],
   "source": [
    "#plot accuracy\n",
    "plot_accuracy(history)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 105,
   "id": "0a0697dc",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.54      0.54      0.54        61\n",
      "           1       0.18      0.29      0.22        35\n",
      "           2       0.45      0.69      0.54        48\n",
      "           3       0.59      0.40      0.47        43\n",
      "           4       0.57      0.52      0.54        91\n",
      "           5       0.55      0.40      0.47        92\n",
      "\n",
      "    accuracy                           0.48       370\n",
      "   macro avg       0.48      0.47      0.46       370\n",
      "weighted avg       0.51      0.48      0.48       370\n",
      "\n",
      "accuracy: 0.478\n"
     ]
    }
   ],
   "source": [
    "# Accuracy\n",
    "train_test_accuracy(cnn_lstm_model,X_train,Y_train,X_test,Y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "3cc249b2",
   "metadata": {},
   "outputs": [],
   "source": [
    "#Classification report\n",
    "report_nn(cnn_lstm_model,X_test,Y_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8b9d2838",
   "metadata": {},
   "source": [
    "##### Confusion Matrix"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 107,
   "id": "74ced08b",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "--Confusion Matrix for CNN+LSTM--\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x1080 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#cnn_lstm_model0=cnn_lstm_model\n",
    "print('--Confusion Matrix for CNN+LSTM--')\n",
    "print_confusion(cnn_lstm_model,X_test,Y_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "32e1d892",
   "metadata": {},
   "source": [
    "## Bidirectional LSTM"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "29038356",
   "metadata": {},
   "outputs": [],
   "source": [
    "lstmbd_model = BID_LSTM_model(embedding_matrix,vocab_len,emb_dim)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a21856f2",
   "metadata": {},
   "outputs": [],
   "source": [
    "## tranining\n",
    "%%time\n",
    "history=lstmbd_model.fit(X_train, np.asarray(Y_train), validation_data=(X_val, np.asarray(Y_val)),epochs=epochs, batch_size=batch_size)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "e76ab166",
   "metadata": {},
   "outputs": [],
   "source": [
    "#plot accuracy\n",
    "plot_accuracy(history,epochs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "8247af34",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Accuracy\n",
    "train_test_accuracy(cnn_lstm_model,X_train,Y_train,X_test,Y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "926f9f00",
   "metadata": {},
   "outputs": [],
   "source": [
    "#Classification report\n",
    "report_nn(cnn_lstm_model,X_test,Y_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "da6ddb5e",
   "metadata": {},
   "source": [
    "## Bidirectional GRU"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "b04a1db7",
   "metadata": {},
   "outputs": [],
   "source": [
    "grubd_model = BID_GRU_model(embedding_matrix,vocab_len,emb_dim)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "72ce52a6",
   "metadata": {},
   "outputs": [],
   "source": [
    "## tranining\n",
    "%%time\n",
    "history=grubd_model.fit(X_train, np.asarray(Y_train), validation_data=(X_val, np.asarray(Y_val)),epochs=epochs, batch_size=batch_size)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ef7f10aa",
   "metadata": {},
   "outputs": [],
   "source": [
    "#plot accuracy\n",
    "plot_accuracy(history,epochs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "df09b1f9",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Accuracy\n",
    "train_test_accuracy(grubd_model,X_train,Y_train,X_test,Y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "fe095440",
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "#Classification report\n",
    "report_nn(grubd_model,X_test,Y_test)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.9.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}