Notebook_analyse_sentiment.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "7907a0f1-4e9d-4d35-b17f-ba646152c3f1",
   "metadata": {
    "jp-MarkdownHeadingCollapsed": true,
    "tags": []
   },
   "source": [
    "<font color = 'red' ><h1 style = 'text-align: center; font-family: cambria'>NOTEBOOK D'ANALYSE DE SENTIMENT AVEC UTILISATION DES LSTM</h1></font>\n",
    "        <h3 style = 'text-decoration : underline; color: blue'>Objectifs : </h3>\n",
    "        <li>Importer le datasets et les modules de traitement du langage naturel</li>\n",
    "        <li>Pretraiter des données</li>\n",
    "        <li>Encodage et vectorisation des données</li>\n",
    "        <li>Creation du model</li>\n",
    "        <li>Entrainer le model</li>\n",
    "        <li>Prediction et analyse du resultat</li>\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4128f183-ee26-4f35-ac58-9dc6fc79c160",
   "metadata": {
    "tags": []
   },
   "source": [
    "<h1 style = 'color : blue; font-family: cambria; text-align: center'>I- Importation des modules</h1>"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4fd46276-1049-4c6d-a644-3443000d1fae",
   "metadata": {},
   "source": [
    "<h3 style = 'color : orange; font-family: cambria'>I.1- Importation des bibliotheques</h3>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "9e814283-4a70-488c-a0ea-de551e0f8a4c",
   "metadata": {},
   "outputs": [],
   "source": [
    "import os ,sys\n",
    "import warnings\n",
    "warnings.filterwarnings('ignore')\n",
    "\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "from matplotlib import cm\n",
    "import re\n",
    "import random\n",
    "import string\n",
    "import math\n",
    "import itertools\n",
    "\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.preprocessing import LabelEncoder\n",
    "from sklearn.feature_extraction.text import CountVectorizer , TfidfVectorizer\n",
    "\n",
    "# Modules pour la construction du modèle\n",
    "import tensorflow as tf\n",
    "from tensorflow import keras\n",
    "from keras.layers import Input,Dense,Activation,LSTM,Embedding,Conv1D,Bidirectional,SpatialDropout1D,Dropout\n",
    "from keras import Sequential\n",
    "from keras.preprocessing import sequence\n",
    "from keras.utils import pad_sequences\n",
    "from keras.callbacks import Callback , ModelCheckpoint , CSVLogger , ReduceLROnPlateau , LearningRateScheduler , EarlyStopping\n",
    "from keras.optimizers import Adam , RMSprop\n",
    "from keras.preprocessing.text import Tokenizer\n",
    "\n",
    "\n",
    "from sklearn.metrics import confusion_matrix , classification_report\n",
    "\n",
    "import nltk\n",
    "#nltk.download('stopwords')\n",
    "from nltk.corpus import stopwords\n",
    "from nltk.stem import SnowballStemmer , WordNetLemmatizer\n",
    "from nltk.tokenize import word_tokenize , RegexpTokenizer\n",
    "\n",
    "# Parametrages\n",
    "path = 'data/training.1600000.processed.noemoticon.csv'\n",
    "np.random.seed(31415)\n",
    "tf.random.set_seed(2)\n",
    "%matplotlib inline\n",
    "plt.style.use('ggplot')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ead699f4-9288-4a2c-b3c3-24d5bface75e",
   "metadata": {},
   "source": [
    "<h3 style = 'color : orange; font-family: cambria'>I.2- Importation du datasets</h3>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "dacbb9ca-e09b-46a8-b01f-be66ca02b6bf",
   "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>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>1467810369</td>\n",
       "      <td>Mon Apr 06 22:19:45 PDT 2009</td>\n",
       "      <td>NO_QUERY</td>\n",
       "      <td>_TheSpecialOne_</td>\n",
       "      <td>@switchfoot http://twitpic.com/2y1zl - Awww, t...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>1467810672</td>\n",
       "      <td>Mon Apr 06 22:19:49 PDT 2009</td>\n",
       "      <td>NO_QUERY</td>\n",
       "      <td>scotthamilton</td>\n",
       "      <td>is upset that he can't update his Facebook by ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>1467810917</td>\n",
       "      <td>Mon Apr 06 22:19:53 PDT 2009</td>\n",
       "      <td>NO_QUERY</td>\n",
       "      <td>mattycus</td>\n",
       "      <td>@Kenichan I dived many times for the ball. Man...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "      <td>1467811184</td>\n",
       "      <td>Mon Apr 06 22:19:57 PDT 2009</td>\n",
       "      <td>NO_QUERY</td>\n",
       "      <td>ElleCTF</td>\n",
       "      <td>my whole body feels itchy and like its on fire</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>1467811193</td>\n",
       "      <td>Mon Apr 06 22:19:57 PDT 2009</td>\n",
       "      <td>NO_QUERY</td>\n",
       "      <td>Karoli</td>\n",
       "      <td>@nationwideclass no, it's not behaving at all....</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   0           1                             2         3                4  \\\n",
       "0  0  1467810369  Mon Apr 06 22:19:45 PDT 2009  NO_QUERY  _TheSpecialOne_   \n",
       "1  0  1467810672  Mon Apr 06 22:19:49 PDT 2009  NO_QUERY    scotthamilton   \n",
       "2  0  1467810917  Mon Apr 06 22:19:53 PDT 2009  NO_QUERY         mattycus   \n",
       "3  0  1467811184  Mon Apr 06 22:19:57 PDT 2009  NO_QUERY          ElleCTF   \n",
       "4  0  1467811193  Mon Apr 06 22:19:57 PDT 2009  NO_QUERY           Karoli   \n",
       "\n",
       "                                                   5  \n",
       "0  @switchfoot http://twitpic.com/2y1zl - Awww, t...  \n",
       "1  is upset that he can't update his Facebook by ...  \n",
       "2  @Kenichan I dived many times for the ball. Man...  \n",
       "3    my whole body feels itchy and like its on fire   \n",
       "4  @nationwideclass no, it's not behaving at all....  "
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Utilisation de la bibliotheque pandas\n",
    "df = pd.read_csv(path  ,encoding = 'ISO-8859-1' , header = None)\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "ccc29480-eb6e-4c59-bdd4-8c95f728bbd1",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Datasets de 1600000 lignes et 6 colonnes\n"
     ]
    }
   ],
   "source": [
    "print(\"Datasets de {0} lignes et {1} colonnes\".format(df.shape[0] , df.shape[1]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "a996e80b-be2f-4b96-8292-3c86324f2cb6",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 1600000 entries, 0 to 1599999\n",
      "Data columns (total 6 columns):\n",
      " #   Column  Non-Null Count    Dtype \n",
      "---  ------  --------------    ----- \n",
      " 0   0       1600000 non-null  int64 \n",
      " 1   1       1600000 non-null  int64 \n",
      " 2   2       1600000 non-null  object\n",
      " 3   3       1600000 non-null  object\n",
      " 4   4       1600000 non-null  object\n",
      " 5   5       1600000 non-null  object\n",
      "dtypes: int64(2), object(4)\n",
      "memory usage: 73.2+ MB\n"
     ]
    }
   ],
   "source": [
    "# Informations sur le dataset\n",
    "df.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "b5b3c43f-4579-4435-8036-c8b39e63ac24",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Renommons le datasets grace a la description du datasets\n",
    "df.columns = ['label' , 'time' , 'date' , 'query' , 'user_id' , 'text']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "bf058d0c-af68-46c9-9741-cbd1c4bf060b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    800000\n",
       "4    800000\n",
       "Name: label, dtype: int64"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Contenu de la colonne 'sentiment'\n",
    "df['label'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "a007f66d-135b-4718-80ee-eff56c71adf7",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# recherche des valeyrs de Null\n",
    "np.sum(df.isnull().any(axis = 1))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "64385a3f-3ff7-4fad-822a-3ef39520a35b",
   "metadata": {},
   "source": [
    "<h1 style = 'color : blue; font-family: cambria; text-align: center'>II- Pretraitement des donnees</h1>\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ee776073-a707-46dc-bbc9-b6c204e37f15",
   "metadata": {},
   "source": [
    "<div style = 'background-color: #f05c07a9;border-radius:30px; border: 2px solid black'><br/><h3 style = 'text-align : center; font-family: cambria;font-size: 35px'>Data preparation</h3>\n",
    "    <li style = 'color : grey;text-align: center; font-size: 18px'>Reduction du dataset pour faciliter l'entrainement</li>\n",
    "    <li style = 'color : grey;text-align: center; font-size: 18px'>Selection des colonnes utiles</li>\n",
    "    <li style = 'color : grey;text-align: center; font-size: 18px'>Visualisation des differents labels</li>\n",
    "    <li style = 'color : grey;text-align: center; font-size: 18px'>Text processing</li>\n",
    "    <li style = 'color : grey;text-align: center; font-size: 18px'>conversion du texte pour l'utilisation</li>\n",
    "    <li style = 'color : grey;text-align: center; font-size: 18px'>Nettoyage du texte</li>\n",
    "    <li style = 'color : grey;text-align: center; font-size: 18px'>Tokenisation du texte</li>\n",
    "    <li style = 'color : grey;text-align: center; font-size: 18px'>Normalisation du texte</li>\n",
    "    <li style = 'color : grey;text-align: center; font-size: 18px'>Extraction et classification des donnees</li><br/><br/>\n",
    "</div>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "0e658b91-4896-4a26-9782-0c53bf472832",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    50000\n",
       "4    50000\n",
       "Name: label, dtype: int64"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = df.groupby('label').apply(lambda x: x.sample(n=50000)).reset_index(drop = True)\n",
    "\n",
    "df.label.value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "ccdbd16d-3d1e-4eb0-8355-9ed5ae3006e0",
   "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>time</th>\n",
       "      <th>date</th>\n",
       "      <th>query</th>\n",
       "      <th>user_id</th>\n",
       "      <th>text</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>99995</th>\n",
       "      <td>Positive</td>\n",
       "      <td>2057133488</td>\n",
       "      <td>Sat Jun 06 12:26:05 PDT 2009</td>\n",
       "      <td>NO_QUERY</td>\n",
       "      <td>taylormonet23</td>\n",
       "      <td>@TheRealLilZa &amp;quot;oh sayy can you seeeee&amp;quo...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>99996</th>\n",
       "      <td>Positive</td>\n",
       "      <td>1881538206</td>\n",
       "      <td>Fri May 22 04:56:02 PDT 2009</td>\n",
       "      <td>NO_QUERY</td>\n",
       "      <td>KatMcGraw</td>\n",
       "      <td>If it weren't for my kids, I'd ... nevermind.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>99997</th>\n",
       "      <td>Positive</td>\n",
       "      <td>1468421194</td>\n",
       "      <td>Tue Apr 07 01:33:57 PDT 2009</td>\n",
       "      <td>NO_QUERY</td>\n",
       "      <td>Lydisquidie</td>\n",
       "      <td>@Maggotron haha! we think it might have been a...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>99998</th>\n",
       "      <td>Positive</td>\n",
       "      <td>1962518643</td>\n",
       "      <td>Fri May 29 11:10:10 PDT 2009</td>\n",
       "      <td>NO_QUERY</td>\n",
       "      <td>MelissaHourigan</td>\n",
       "      <td>@hezness come to Hooked on Colfax and have som...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>99999</th>\n",
       "      <td>Positive</td>\n",
       "      <td>2000186591</td>\n",
       "      <td>Mon Jun 01 21:35:06 PDT 2009</td>\n",
       "      <td>NO_QUERY</td>\n",
       "      <td>KanJedi</td>\n",
       "      <td>Cleared all the bills for the month. Awesome! ...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          label        time                          date     query  \\\n",
       "99995  Positive  2057133488  Sat Jun 06 12:26:05 PDT 2009  NO_QUERY   \n",
       "99996  Positive  1881538206  Fri May 22 04:56:02 PDT 2009  NO_QUERY   \n",
       "99997  Positive  1468421194  Tue Apr 07 01:33:57 PDT 2009  NO_QUERY   \n",
       "99998  Positive  1962518643  Fri May 29 11:10:10 PDT 2009  NO_QUERY   \n",
       "99999  Positive  2000186591  Mon Jun 01 21:35:06 PDT 2009  NO_QUERY   \n",
       "\n",
       "               user_id                                               text  \n",
       "99995    taylormonet23  @TheRealLilZa &quot;oh sayy can you seeeee&quo...  \n",
       "99996        KatMcGraw     If it weren't for my kids, I'd ... nevermind.   \n",
       "99997      Lydisquidie  @Maggotron haha! we think it might have been a...  \n",
       "99998  MelissaHourigan  @hezness come to Hooked on Colfax and have som...  \n",
       "99999          KanJedi  Cleared all the bills for the month. Awesome! ...  "
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "labels_to_sentiment = {0:\"Negative\", 4:\"Positive\"}\n",
    "df.label = df.label.apply(lambda x: labels_to_sentiment[x])\n",
    "df.tail()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "fc11feab-ab35-4980-898b-d0bcbf03727d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 1.0, 'Sentiment Target Distribution')"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 800x400 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "### Visualisation des differents labels.\n",
    "val_count = df.label.value_counts()\n",
    "\n",
    "plt.figure(figsize=(8,4))\n",
    "plt.bar(val_count.index, val_count.values)\n",
    "plt.xlabel(\"Sentiment\")\n",
    "plt.ylabel(\"Count\")\n",
    "plt.title(\"Sentiment Target Distribution\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "83bf98c7-0d5f-4be3-9dff-08f482b659b5",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "842    The Veronicas were great! But its raining now ...\n",
       "717    just got home after 2 classes. had a great tim...\n",
       "996         @Anna_Montanna Yeah but it isn't activated. \n",
       "690    Cub run with sammers before she leaves for ric...\n",
       "565      @Dunncity  It wasn't suppose to be that easy!!!\n",
       "515    I better have something in the post....been 6 ...\n",
       "687    finally get to use internet after 5 days. inte...\n",
       "810      Blah moving sucks. Why do I have so much stuff \n",
       "47                             Ugh I'm not tired at all \n",
       "506    @Clumsyflic the kid that got a second chance s...\n",
       "Name: text, dtype: object"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# exploration des nouvelle donnees\n",
    "index = np.random.randint(0,1000,(10))\n",
    "df.iloc[index].text"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0e8e5630-ee54-4c9d-b943-dc2b9feb311a",
   "metadata": {},
   "source": [
    "<h3 style = 'color : orange; font-family: cambria'>II.1- Nettoyage des donnees</h3>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "019416de-6c3b-4ec2-8e29-0e5624549a3c",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "99995    @thereallilza &quot;oh sayy can you seeeee&quo...\n",
       "99996       if it weren't for my kids, i'd ... nevermind. \n",
       "99997    @maggotron haha! we think it might have been a...\n",
       "99998    @hezness come to hooked on colfax and have som...\n",
       "99999    cleared all the bills for the month. awesome! ...\n",
       "Name: text_trans, dtype: object"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Transformer le texte en minuscule\n",
    "df['text_trans'] = df['text'].str.lower()\n",
    "df['text_trans'].tail()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "250e5a64-bf0f-4a66-ab6e-c499cf035171",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "\"i, me, my, myself, we, our, ours, ourselves, you, you're, you've, you'll, you'd, your, yours, yourself, yourselves, he, him, his, himself, she, she's, her, hers, herself, it, it's, its, itself, they, them, their, theirs, themselves, what, which, who, whom, this, that, that'll, these, those, am, is, are, was, were, be, been, being, have, has, had, having, do, does, did, doing, a, an, the, and, but, if, or, because, as, until, while, of, at, by, for, with, about, against, between, into, through, during, before, after, above, below, to, from, up, down, in, out, on, off, over, under, again, further, then, once, here, there, when, where, why, how, all, any, both, each, few, more, most, other, some, such, no, nor, not, only, own, same, so, than, too, very, s, t, can, will, just, don, don't, should, should've, now, d, ll, m, o, re, ve, y, ain, aren, aren't, couldn, couldn't, didn, didn't, doesn, doesn't, hadn, hadn't, hasn, hasn't, haven, haven't, isn, isn't, ma, mightn, mightn't, mustn, mustn't, needn, needn't, shan, shan't, shouldn, shouldn't, wasn, wasn't, weren, weren't, won, won't, wouldn, wouldn't\""
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Nettoyage des donnees avec la methode stopwords de nltk\n",
    "stopwords_list = stopwords.words('english')\n",
    "\", \".join(stopwords.words('english'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "898926c8-a107-4a24-9a37-1caf022e2553",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0                @carol_carter1 think believe one time\n",
       "1    @nursebc1974 know!!! boys r wide awake, n i'm ...\n",
       "2    @maddieeann either mean send he's playing hard...\n",
       "3    @mswindows really wish microsoft would stop pu...\n",
       "4    @tommcfly still think brazilian's fans better....\n",
       "Name: text_trans, dtype: object"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "STOPWORDS = set(stopwords.words('english'))\n",
    "def cleaning_stopwords(text):\n",
    "    return \" \".join([word for word in str(text).split() if word not in STOPWORDS])\n",
    "df['text_trans'] = df['text_trans'].apply(lambda text: cleaning_stopwords(text))\n",
    "df['text_trans'].head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "7d5b42ca-8e89-4c3e-9700-22304097bfb8",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "99995    thereallilza quotoh sayy seeeeequot aha make h...\n",
       "99996                                   kids id  nevermind\n",
       "99997    maggotron haha think might little old lady liv...\n",
       "99998                           hezness come hooked colfax\n",
       "99999       cleared bills month awesome put rest aside fun\n",
       "Name: text_trans, dtype: object"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Nettoyage des ponctuations\n",
    "ponctuations = string.punctuation\n",
    "ponctuations_list = ponctuations\n",
    "\n",
    "def cleaning_punctuations(text):\n",
    "    translator = str.maketrans('', '', ponctuations_list)\n",
    "    return text.translate(translator)\n",
    "\n",
    "df['text_trans']= df['text_trans'].apply(lambda x: cleaning_punctuations(x))\n",
    "df['text_trans'].tail()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "1cdd36d4-70db-44e4-b7ad-3e5c41e74158",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "99995    therealilza quotoh say sequot aha make home sa...\n",
       "99996                                    kids id nevermind\n",
       "99997    magotron haha think might litle old lady lives...\n",
       "99998                             heznes come hoked colfax\n",
       "99999        cleared bils month awesome put rest aside fun\n",
       "Name: text_trans, dtype: object"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Nettoyage des caracteres repetitifs\n",
    "def cleaning_repeating_char(text):\n",
    "    return re.sub(r'(.)\\1+', r'\\1', text)\n",
    "\n",
    "df['text_trans'] = df['text_trans'].apply(lambda x: cleaning_repeating_char(x))\n",
    "df['text_trans'].tail()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "74decf6e-326a-4fcf-ac47-512d6c7e961d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "99995    therealilza quotoh say sequot aha make home sa...\n",
       "99996                                    kids id nevermind\n",
       "99997    magotron haha think might litle old lady lives...\n",
       "99998                             heznes come hoked colfax\n",
       "99999        cleared bils month awesome put rest aside fun\n",
       "Name: text_trans, dtype: object"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Nettayage des emails\n",
    "def cleaning_email(data):\n",
    "    return re.sub('@[^\\s]+', ' ', data)\n",
    "\n",
    "df['text_trans']= df['text_trans'].apply(lambda x: cleaning_email(x))\n",
    "df['text_trans'].tail()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "445b1424-70ce-4a3d-8366-be0b21ddd09d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "99995    therealilza quotoh say sequot aha make home sa...\n",
       "99996                                    kids id nevermind\n",
       "99997    magotron haha think might litle old lady lives...\n",
       "99998                             heznes come hoked colfax\n",
       "99999        cleared bils month awesome put rest aside fun\n",
       "Name: text_trans, dtype: object"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Nettoyage des URL\n",
    "def cleaning_URLs(data):\n",
    "    return re.sub('((www\\.[^\\s]+)|(https?://[^\\s]+))',' ',data)\n",
    "df['text_trans'] = df['text_trans'].apply(lambda x: cleaning_URLs(x))\n",
    "df['text_trans'].tail()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "57228855-adf3-422e-8a5e-3af115644e8a",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "99995    therealilza quotoh say sequot aha make home sa...\n",
       "99996                                    kids id nevermind\n",
       "99997    magotron haha think might litle old lady lives...\n",
       "99998                             heznes come hoked colfax\n",
       "99999        cleared bils month awesome put rest aside fun\n",
       "Name: text_trans, dtype: object"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Nettoyage des nombres\n",
    "def cleaning_numbers(data):\n",
    "    return re.sub('[0-9]+', '', data)\n",
    "\n",
    "df['text_trans'] = df['text_trans'].apply(lambda x: cleaning_numbers(x))\n",
    "df['text_trans'].tail()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9254674e-7432-41ad-9782-6c021f191fe1",
   "metadata": {},
   "source": [
    "<h1 style = 'color : blue; font-family: cambria; text-align: center'>III- Tokenisation des donnees</h1>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "d5cae8bc-ff04-426b-b2a6-d798f6455e3b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0             [carolcarter, think, believe, one, time]\n",
       "1    [nursebc, know, boys, r, wide, awake, n, im, f...\n",
       "2    [madiean, either, mean, send, hes, playing, ha...\n",
       "3    [mswindows, realy, wish, microsoft, would, sto...\n",
       "4    [tomcfly, stil, think, brazilians, fans, beter...\n",
       "Name: text_trans, dtype: object"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Tokenisation du text\n",
    "tokenizer = RegexpTokenizer(r'\\w+')\n",
    "df['text_trans'] = df['text_trans'].apply(tokenizer.tokenize)\n",
    "df['text_trans'].head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "3de05478-e52f-4579-8fd4-162fcc167c15",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0             [carolcarter, think, believe, one, time]\n",
       "1    [nursebc, know, boys, r, wide, awake, n, im, f...\n",
       "2    [madiean, either, mean, send, hes, playing, ha...\n",
       "3    [mswindows, realy, wish, microsoft, would, sto...\n",
       "4    [tomcfly, stil, think, brazilians, fans, beter...\n",
       "Name: text_trans, dtype: object"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Stemming\n",
    "\n",
    "st = nltk.PorterStemmer()\n",
    "def stemming_on_text(data):\n",
    "    text = [st.stem(word) for word in data]\n",
    "    return data\n",
    "\n",
    "df['text_trans']= df['text_trans'].apply(lambda x: stemming_on_text(x))\n",
    "df['text_trans'].head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "b11d3061-c198-466a-b5fc-927dbcf9972d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>text</th>\n",
       "      <th>text_trans</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>@carol_carter1 I don't think so  I believe it ...</td>\n",
       "      <td>[carolcarter, think, believe, one, time]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>@nursebc1974 I know!!!  My boys r wide awake, ...</td>\n",
       "      <td>[nursebc, know, boys, r, wide, awake, n, im, f...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>@Maddieeann either he didn't mean to send it t...</td>\n",
       "      <td>[madiean, either, mean, send, hes, playing, ha...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>@MSWindows I really wish microsoft would stop ...</td>\n",
       "      <td>[mswindows, realy, wish, microsoft, would, sto...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>@tommcfly I still think that brazilian's fans ...</td>\n",
       "      <td>[tomcfly, stil, think, brazilians, fans, beter...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>995</th>\n",
       "      <td>and when I got out my phone was dead :[[ it li...</td>\n",
       "      <td>[got, phone, dead, lived, weks, ugh]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>996</th>\n",
       "      <td>@Anna_Montanna Yeah but it isn't activated.</td>\n",
       "      <td>[anamontana, yeah, activated]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>997</th>\n",
       "      <td>Aww  i'm at retarded leadership shit</td>\n",
       "      <td>[aw, im, retarded, leadership, shit]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>998</th>\n",
       "      <td>@cehouck that's because we're like the same pe...</td>\n",
       "      <td>[cehouck, thats, were, like, person, ahaha, ps...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>999</th>\n",
       "      <td>Crap! I screwed up my entire Adobe Suite insta...</td>\n",
       "      <td>[crap, screwed, entire, adobe, suite, instal, ...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>1000 rows × 2 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                  text  \\\n",
       "0    @carol_carter1 I don't think so  I believe it ...   \n",
       "1    @nursebc1974 I know!!!  My boys r wide awake, ...   \n",
       "2    @Maddieeann either he didn't mean to send it t...   \n",
       "3    @MSWindows I really wish microsoft would stop ...   \n",
       "4    @tommcfly I still think that brazilian's fans ...   \n",
       "..                                                 ...   \n",
       "995  and when I got out my phone was dead :[[ it li...   \n",
       "996       @Anna_Montanna Yeah but it isn't activated.    \n",
       "997               Aww  i'm at retarded leadership shit   \n",
       "998  @cehouck that's because we're like the same pe...   \n",
       "999  Crap! I screwed up my entire Adobe Suite insta...   \n",
       "\n",
       "                                            text_trans  \n",
       "0             [carolcarter, think, believe, one, time]  \n",
       "1    [nursebc, know, boys, r, wide, awake, n, im, f...  \n",
       "2    [madiean, either, mean, send, hes, playing, ha...  \n",
       "3    [mswindows, realy, wish, microsoft, would, sto...  \n",
       "4    [tomcfly, stil, think, brazilians, fans, beter...  \n",
       "..                                                 ...  \n",
       "995               [got, phone, dead, lived, weks, ugh]  \n",
       "996                      [anamontana, yeah, activated]  \n",
       "997               [aw, im, retarded, leadership, shit]  \n",
       "998  [cehouck, thats, were, like, person, ahaha, ps...  \n",
       "999  [crap, screwed, entire, adobe, suite, instal, ...  \n",
       "\n",
       "[1000 rows x 2 columns]"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#texte initial vs texte final\n",
    "data = df.iloc[:1000].copy()\n",
    "data[[\"text\", \"text_trans\"]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "240f1bf0-3906-4901-be4c-ca19591599dd",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['label', 'time', 'date', 'query', 'user_id', 'text', 'text_trans'], dtype='object')"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.columns"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "75312316-dc79-4ac2-b22f-ece49bab9941",
   "metadata": {},
   "source": [
    "<h3 style = 'color : orange; font-family: cambria'>III.1- Separation du dataset</h3>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "02f2a2bd-7ac5-4612-bf13-d14051e61d3a",
   "metadata": {},
   "outputs": [],
   "source": [
    "## definition des parametres\n",
    "MAX_NB_WORDS = 10000  ## 100000\n",
    "MAX_SEQUENCE_LENGTH = 30  ## chaque sequence a une taille de 30 mots"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "68090ccd-c083-4b06-a225-78b3b8a7b827",
   "metadata": {},
   "source": [
    "### Prepartion des donnees d'entrées\n",
    "* Conversion du texte en sequence\n",
    "* Selection d'un vocabulaire de 10000 mots par rapport a l'importance de ceux-ci dans les differents labels"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "8d5eafe0-f869-4129-9725-8e15c6dbd9b5",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train Data size: 80000\n",
      "Test Data size 20000\n"
     ]
    }
   ],
   "source": [
    "train_data, test_data = train_test_split(df[[\"label\",\"text_trans\"]], test_size=0.2,\n",
    "                                         random_state=7) \n",
    "# separation du dataset en training et testing sets\n",
    "\n",
    "print(\"Train Data size:\", len(train_data))\n",
    "print(\"Test Data size\", len(test_data))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "57e98d5b-c528-4995-88a5-37c9e6b3567b",
   "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_trans</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>91526</th>\n",
       "      <td>Positive</td>\n",
       "      <td>[ned, work, word, esay, chapter]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19416</th>\n",
       "      <td>Negative</td>\n",
       "      <td>[etsywiki, coming, up, get, eror, page]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>99079</th>\n",
       "      <td>Positive</td>\n",
       "      <td>[ubercol, absolutely, blast, remember, harem, ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>45086</th>\n",
       "      <td>Negative</td>\n",
       "      <td>[final, exams]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>54659</th>\n",
       "      <td>Positive</td>\n",
       "      <td>[bigcam, lol, quen, victoria, precise]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50266</th>\n",
       "      <td>Positive</td>\n",
       "      <td>[stil, unhapy, lack, guitar, string, realy, wa...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>62172</th>\n",
       "      <td>Positive</td>\n",
       "      <td>[words, hanah, montana, nobodys, perfect]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>66637</th>\n",
       "      <td>Positive</td>\n",
       "      <td>[like, cover, environmental, science, bok, haha]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>95845</th>\n",
       "      <td>Positive</td>\n",
       "      <td>[ah, kim, kardashian, absolutely, love, her, a...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>98862</th>\n",
       "      <td>Positive</td>\n",
       "      <td>[fel, god, eating, dry, cherios, like, litle, ...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          label                                         text_trans\n",
       "91526  Positive                   [ned, work, word, esay, chapter]\n",
       "19416  Negative            [etsywiki, coming, up, get, eror, page]\n",
       "99079  Positive  [ubercol, absolutely, blast, remember, harem, ...\n",
       "45086  Negative                                     [final, exams]\n",
       "54659  Positive             [bigcam, lol, quen, victoria, precise]\n",
       "50266  Positive  [stil, unhapy, lack, guitar, string, realy, wa...\n",
       "62172  Positive          [words, hanah, montana, nobodys, perfect]\n",
       "66637  Positive   [like, cover, environmental, science, bok, haha]\n",
       "95845  Positive  [ah, kim, kardashian, absolutely, love, her, a...\n",
       "98862  Positive  [fel, god, eating, dry, cherios, like, litle, ..."
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_data.head(10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "ac59f1a0-c590-46a3-85af-1c406d3a106a",
   "metadata": {},
   "outputs": [],
   "source": [
    "## Creation du tokenizer\n",
    "tokenizer = Tokenizer()\n",
    "tokenizer.fit_on_texts(train_data.text_trans) \n",
    "### creation d'un dictionnaire de vocabulaire"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "9ad69820-3c94-4af2-a6c3-013bd41fd0ac",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Vocabulary Size : 84758\n"
     ]
    }
   ],
   "source": [
    "word_index = tokenizer.word_index\n",
    "vocab_size = len(tokenizer.word_index) + 1\n",
    "print(\"\\nVocabulary Size :\", vocab_size)  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "81c64a2c-9eaf-4100-a26e-37855b0bc0aa",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training X Shape: (80000, 30)\n",
      "Testing X Shape: (20000, 30)\n"
     ]
    }
   ],
   "source": [
    "x_train = pad_sequences(tokenizer.texts_to_sequences(train_data.text_trans),\n",
    "                        maxlen = MAX_SEQUENCE_LENGTH)\n",
    "x_test = pad_sequences(tokenizer.texts_to_sequences(test_data.text_trans),\n",
    "                       maxlen = MAX_SEQUENCE_LENGTH)\n",
    "\n",
    "print(\"Training X Shape:\",x_train.shape)\n",
    "print(\"Testing X Shape:\",x_test.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c80c8b00-89a3-49d3-a84d-f9dca667a7a6",
   "metadata": {},
   "source": [
    "## Encodage\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "id": "e35e50f6-9e13-4c2f-9cc4-56d32cc9f246",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "y_train shape: (80000, 1)\n",
      "y_test shape: (20000, 1)\n"
     ]
    }
   ],
   "source": [
    "encoder = LabelEncoder()\n",
    "encoder.fit(train_data.label.to_list())\n",
    "\n",
    "y_train = encoder.transform(train_data.label.to_list())\n",
    "y_test = encoder.transform(test_data.label.to_list())\n",
    "\n",
    "y_train = y_train.reshape(-1,1)\n",
    "y_test = y_test.reshape(-1,1)\n",
    "\n",
    "print(\"y_train shape:\", y_train.shape)\n",
    "print(\"y_test shape:\", y_test.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "fefcff49-c2fc-412f-a40d-25ef4916ac4f",
   "metadata": {},
   "source": [
    "## Vectorisation\n",
    "Ici nous allons utilisé le model pretrainé de CNN GloVe pour la representations des mots"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "id": "fb5cf7d9-0779-4bd5-b279-d5db4138d77a",
   "metadata": {},
   "outputs": [],
   "source": [
    "glove_path = 'glove.6B/glove.6B.300d.txt'\n",
    "embedding_dim = 300   ### pour controller l'espace entre les differents mots vectorisés\n",
    "model_path = 'save/model_avanced_lstm/model_diLstm.h5'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "id": "54e044b5-0427-4a69-bb39-9388e01364fe",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 400000 word vectors.\n"
     ]
    }
   ],
   "source": [
    "embeddings_index = {}\n",
    "\n",
    "### Processing the file\n",
    "for line in open(glove_path , encoding = 'utf-8'):\n",
    "    values = line.split()\n",
    "    word = value = values[0]\n",
    "    coefs = np.asarray(values[1:], dtype='float32')\n",
    "    embeddings_index[word] = coefs\n",
    "\n",
    "print('Found %s word vectors.' %len(embeddings_index))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "id": "f1f3492e-6f14-41ff-9632-17d862fa2d46",
   "metadata": {},
   "outputs": [],
   "source": [
    "### Creation de la matrice d'embedding\n",
    "embedding_matrix = np.zeros((vocab_size, embedding_dim))  \n",
    "for word, i in word_index.items():\n",
    "    embedding_vector = embeddings_index.get(word)  \n",
    "    if embedding_vector is not None:\n",
    "        embedding_matrix[i] = embedding_vector"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "id": "aa268318-1130-4f15-be80-79fc19fb10e8",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "For the word \"love\", its respective vector representation in a 300 Dimensional Space is:\n",
      "\n",
      " [-4.52050000e-01 -3.31220001e-01 -6.36069998e-02  2.83250008e-02\n",
      " -2.13719994e-01  1.68390006e-01 -1.71859991e-02  4.73090000e-02\n",
      " -5.23549989e-02 -9.87060010e-01  5.37620008e-01 -2.68929988e-01\n",
      " -5.42940021e-01  7.24869967e-02  6.61929995e-02 -2.18140006e-01\n",
      " -1.21129997e-01 -2.88320005e-01  4.81610000e-01  6.91850007e-01\n",
      " -2.00220004e-01  1.00820005e+00 -1.18649997e-01  5.87100029e-01\n",
      "  1.84819996e-01  4.57989983e-02 -1.78360008e-02 -3.39520007e-01\n",
      "  2.93139994e-01 -1.99509993e-01 -1.89300001e-01  4.32669997e-01\n",
      " -6.31810009e-01 -2.95100003e-01 -1.05470002e+00  1.82310000e-01\n",
      " -4.50399995e-01 -2.77999997e-01 -1.40210003e-01  3.67849991e-02\n",
      "  2.64869988e-01 -6.67119980e-01 -1.52040005e-01 -3.50010008e-01\n",
      "  4.08639997e-01 -7.36149997e-02  6.76299989e-01  1.82740003e-01\n",
      " -4.16599996e-02  1.50140002e-02  2.52160013e-01 -1.01089999e-01\n",
      "  3.19150016e-02 -1.12980001e-01 -4.01470006e-01  1.72739998e-01\n",
      "  1.84970000e-03  2.44560003e-01  6.87770009e-01 -2.70190001e-01\n",
      "  8.07280004e-01 -5.82959987e-02  4.05499995e-01  3.98930013e-01\n",
      " -9.16879997e-02 -5.20799994e-01  2.45700002e-01  6.30009994e-02\n",
      "  2.14210004e-01  3.31970006e-01 -3.42990011e-01 -4.87349987e-01\n",
      "  2.22640000e-02  2.78620005e-01  2.38810003e-01  9.77939963e-02\n",
      "  3.80230010e-01 -3.77440006e-02 -4.19660002e-01 -1.91450000e-01\n",
      " -9.58300009e-02  2.68709987e-01  5.28760016e-01 -2.68700004e-01\n",
      " -3.44500005e-01  2.54130006e-01  1.36059999e-01  3.65280002e-01\n",
      "  8.19600001e-02 -5.22239983e-01  3.41590010e-02  1.70190006e-01\n",
      " -1.05200000e-01 -8.08730006e-01 -1.80950001e-01 -6.08230010e-02\n",
      "  2.60140002e-01 -7.87170008e-02 -1.51610002e-01 -6.84379995e-01\n",
      "  2.18089998e-01  4.73299995e-02 -1.21469997e-01 -3.38919997e-01\n",
      " -1.86289996e-02  2.45460004e-01  2.73739994e-01 -2.70449996e-01\n",
      " -8.72329995e-02 -4.78709996e-01  1.96300000e-01  2.31240001e-02\n",
      "  1.94529995e-01 -4.68600005e-01  4.44990009e-01  1.83599994e-01\n",
      " -1.64800003e-01  3.85980010e-01  3.60700011e-01 -3.11080009e-01\n",
      " -3.46269995e-01  1.90040007e-01 -9.86829996e-02  3.38209987e-01\n",
      " -1.39780000e-01 -7.27230012e-01 -1.06600001e-01  1.92079996e-03\n",
      " -3.30929995e-01  3.51170003e-01  1.54909998e-01  1.71499997e-01\n",
      "  2.89330006e-01 -4.92129996e-02 -5.05670011e-01 -2.35129997e-01\n",
      " -2.80050009e-01 -2.84869999e-01 -2.43929997e-01 -4.08379994e-02\n",
      " -2.72229999e-01  9.45639983e-02  1.73319995e-01  1.86910003e-01\n",
      " -1.35849997e-01 -4.64390010e-01 -4.56770003e-01  9.78899971e-02\n",
      "  3.91130000e-02  1.79010004e-01 -4.56290007e-01  4.75199997e-01\n",
      " -2.03879997e-01  2.41229996e-01  6.75509989e-01  6.21050000e-02\n",
      " -1.89400002e-01  1.50539994e-01  1.58749998e-01 -2.33649999e-01\n",
      "  3.74289989e-01 -2.34630004e-01  4.06919986e-01  1.40349999e-01\n",
      " -4.29710001e-01  5.67130029e-01 -2.67060012e-01 -6.80280030e-02\n",
      " -5.12639999e-01 -3.09450001e-01 -3.90650004e-01 -2.70040005e-01\n",
      " -1.18019998e+00  6.64799988e-01 -2.43660003e-01  3.91829997e-01\n",
      " -2.89700001e-01 -1.88390002e-01 -4.92819995e-01  1.45380005e-01\n",
      "  2.44670004e-01 -4.43399977e-03  2.32649997e-01  7.45439976e-02\n",
      " -3.00060004e-01 -3.02720010e-01 -1.23939998e-01  3.34729999e-01\n",
      "  3.38800013e-01  9.74450037e-02 -3.37729990e-01 -5.43160021e-01\n",
      " -4.75140005e-01 -1.56959996e-01 -9.35159981e-01 -8.70340019e-02\n",
      " -2.67430007e-01  6.46409986e-04  3.19400012e-01 -6.25619991e-03\n",
      "  1.58539999e+00  1.24839999e-01  4.84809995e-01  7.53939971e-02\n",
      "  1.89630002e-01 -1.02260001e-01  4.74130005e-01  7.04020023e-01\n",
      " -6.44180030e-02 -1.01810005e-02 -6.86190009e-01  1.53350001e-02\n",
      "  4.78510000e-02  3.76500010e-01  1.04869999e-01  2.29739994e-01\n",
      "  4.53520000e-01  3.14819992e-01  5.88800013e-02  6.70960024e-02\n",
      "  1.56790003e-01  1.30989999e-01  3.45810018e-02 -7.38959983e-02\n",
      " -3.64329994e-01 -1.88470006e-01  4.15560007e-02 -2.01240003e-01\n",
      " -7.39179999e-02  3.87519985e-01  1.67590007e-01 -4.06269997e-01\n",
      " -1.32200001e-02 -7.17930019e-01 -2.29039997e-01  2.04740003e-01\n",
      " -1.36480004e-01  3.77790004e-01 -4.00029987e-01 -5.01060002e-02\n",
      " -3.77990007e-01  7.21089989e-02  2.30570007e-02  2.38790005e-01\n",
      " -2.24230006e-01 -8.48340020e-02 -7.07260013e-01 -1.61819994e-01\n",
      "  2.63729990e-01  1.22259997e-01  8.02820027e-02  7.59629980e-02\n",
      " -3.46949995e-01  3.77029985e-01  3.66120011e-01 -9.84160006e-02\n",
      "  4.74000007e-01 -2.50849992e-01  1.86409995e-01 -2.59930015e-01\n",
      "  3.90349999e-02 -6.13940001e-01  1.47210002e-01 -6.76190019e-01\n",
      " -1.24890000e-01  3.80309999e-01 -2.80609995e-01 -3.96800011e-01\n",
      "  2.41650008e-02  9.87059996e-02  2.48380005e-01 -4.58810002e-01\n",
      "  2.03099996e-01 -4.35149997e-01 -7.78040010e-03 -2.46420000e-02\n",
      " -1.32040000e+00 -4.10780013e-01  2.81450003e-01  2.81879995e-02\n",
      " -2.04699993e-01  1.60370007e-01 -1.66960001e-01  1.37610003e-01\n",
      " -3.67139995e-01  1.38009995e-01 -2.06880003e-01  3.82739991e-01\n",
      "  2.12889999e-01 -7.50010014e-02 -5.03669977e-01 -2.92949993e-02\n",
      " -2.12539993e-02 -2.42530003e-01  3.35299999e-01 -3.55340004e-01\n",
      "  2.53580004e-01  3.89060006e-02  2.43139997e-01 -2.86960006e-02]\n"
     ]
    }
   ],
   "source": [
    "### pour representer un mot par son vecteur \n",
    "def funct_view_embedding(word):\n",
    "    print(f\"For the word \\\"{word}\\\", its respective vector representation in a 300 Dimensional Space is:\")\n",
    "    print(\"\\n\",embedding_matrix[word_index[f\"{word}\"]])\n",
    "\n",
    "funct_view_embedding(\"love\")  ## un petit exemple"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "id": "3ed894b3-0e7c-4362-9ed9-a7c9182a1652",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(84758, 300)"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "embedding_matrix.shape\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "id": "92c817db-e87c-4258-970c-3cdf3164a79e",
   "metadata": {},
   "outputs": [],
   "source": [
    "## Creation de la couche d'embedding\n",
    "embedding_layer = Embedding(vocab_size,embedding_dim,\n",
    "                                        weights=[embedding_matrix],\n",
    "                                          input_length=MAX_SEQUENCE_LENGTH, ### length of longest sentence\n",
    "                                          trainable=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ab64bdfb-842e-49f8-9949-4e2d68b229f6",
   "metadata": {},
   "source": [
    "<h1 style = 'color : blue; font-family: cambria; text-align: center'>IV- CONSTRUCTION DU MODELE</h1>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "id": "38b2b54c-6fe9-48fc-9979-b9e9d2c4a214",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"model\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " input_1 (InputLayer)        [(None, 30)]              0         \n",
      "                                                                 \n",
      " embedding (Embedding)       (None, 30, 300)           25427400  \n",
      "                                                                 \n",
      " spatial_dropout1d (SpatialD  (None, 30, 300)          0         \n",
      " ropout1D)                                                       \n",
      "                                                                 \n",
      " conv1d (Conv1D)             (None, 26, 64)            96064     \n",
      "                                                                 \n",
      " bidirectional (Bidirectiona  (None, 128)              66048     \n",
      " l)                                                              \n",
      "                                                                 \n",
      " dense (Dense)               (None, 512)               66048     \n",
      "                                                                 \n",
      " dropout (Dropout)           (None, 512)               0         \n",
      "                                                                 \n",
      " dense_1 (Dense)             (None, 512)               262656    \n",
      "                                                                 \n",
      " dense_2 (Dense)             (None, 1)                 513       \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 25,918,729\n",
      "Trainable params: 491,329\n",
      "Non-trainable params: 25,427,400\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "## Learning Rate:\n",
    "lr = 1e-2\n",
    "\n",
    "## Creation des couches\n",
    "sequence_input = Input(shape=(MAX_SEQUENCE_LENGTH,), dtype='int32')\n",
    "embedding_sequences = embedding_layer(sequence_input)\n",
    "x = SpatialDropout1D(0.2)(embedding_sequences)\n",
    "x = Conv1D(64, 5, activation='relu')(x)\n",
    "x = Bidirectional(LSTM(64, dropout=0.2, recurrent_dropout=0.2))(x)\n",
    "x = Dense(512, activation='relu')(x)\n",
    "x = Dropout(0.5)(x)\n",
    "x = Dense(512, activation='relu')(x)\n",
    "outputs = Dense(1, activation='sigmoid')(x)\n",
    "model = tf.keras.Model(sequence_input, outputs) \n",
    "\n",
    "model.compile(optimizer=Adam(learning_rate=lr), loss='binary_crossentropy',\n",
    "              metrics=['accuracy'])\n",
    "\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "id": "0bd539f6-047d-49c9-aabb-4f05e2ae52f8",
   "metadata": {},
   "outputs": [],
   "source": [
    "## politique de controle de learning rate:\n",
    "early_stopping = EarlyStopping(monitor='val_loss',patience=5,mode=\"auto\")\n",
    "\n",
    "## Checkpoints :\n",
    "checkpoint_path = 'model_checkpoints/'\n",
    "checkpoint = ModelCheckpoint(\n",
    "    filepath=checkpoint_path,\n",
    "    save_freq='epoch',\n",
    "    save_weights_only=True,\n",
    "    verbose=1\n",
    ")\n",
    "\n",
    "## Automatic learning rate update:\n",
    "reduce_lr = ReduceLROnPlateau(\n",
    "    monitor='val_loss', \n",
    "    factor=0.1,   \n",
    "    patience=5, \n",
    "    min_lr=0.001,\n",
    "    verbose=1\n",
    ")\n",
    "\n",
    "## Learning Rate Scheduler:\n",
    "def lr_decay(epoch, lr):\n",
    "    if epoch != 0 and epoch % 5 == 0:\n",
    "        return lr * 0.2\n",
    "    return lr\n",
    "\n",
    "lrs=LearningRateScheduler(lr_decay, verbose=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "id": "3b084f04-68bd-4a1b-bee2-a305d299a40e",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:From C:\\Users\\Acern\\AppData\\Local\\Temp\\ipykernel_6508\\2293344113.py:1: is_gpu_available (from tensorflow.python.framework.test_util) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Use `tf.config.list_physical_devices('GPU')` instead.\n",
      "Training on CPU...\n"
     ]
    }
   ],
   "source": [
    "print(\"Training on GPU...\") if tf.test.is_gpu_available() else print(\"Training on CPU...\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "id": "0c893d9a-6d22-49c1-88f4-463f11435497",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 1: LearningRateScheduler setting learning rate to 0.009999999776482582.\n",
      "Epoch 1/20\n",
      "80/80 [==============================] - 55s 640ms/step - loss: 0.6375 - accuracy: 0.6607 - val_loss: 0.5491 - val_accuracy: 0.7190 - lr: 0.0100\n",
      "\n",
      "Epoch 2: LearningRateScheduler setting learning rate to 0.009999999776482582.\n",
      "Epoch 2/20\n",
      "80/80 [==============================] - 51s 640ms/step - loss: 0.5489 - accuracy: 0.7211 - val_loss: 0.5335 - val_accuracy: 0.7304 - lr: 0.0100\n",
      "\n",
      "Epoch 3: LearningRateScheduler setting learning rate to 0.009999999776482582.\n",
      "Epoch 3/20\n",
      "80/80 [==============================] - 52s 644ms/step - loss: 0.5224 - accuracy: 0.7406 - val_loss: 0.5314 - val_accuracy: 0.7261 - lr: 0.0100\n",
      "\n",
      "Epoch 4: LearningRateScheduler setting learning rate to 0.009999999776482582.\n",
      "Epoch 4/20\n",
      "80/80 [==============================] - 52s 652ms/step - loss: 0.5052 - accuracy: 0.7512 - val_loss: 0.5306 - val_accuracy: 0.7344 - lr: 0.0100\n",
      "\n",
      "Epoch 5: LearningRateScheduler setting learning rate to 0.009999999776482582.\n",
      "Epoch 5/20\n",
      "80/80 [==============================] - 52s 651ms/step - loss: 0.4906 - accuracy: 0.7624 - val_loss: 0.5299 - val_accuracy: 0.7355 - lr: 0.0100\n",
      "\n",
      "Epoch 6: LearningRateScheduler setting learning rate to 0.0019999999552965165.\n",
      "Epoch 6/20\n",
      "80/80 [==============================] - 56s 707ms/step - loss: 0.4554 - accuracy: 0.7826 - val_loss: 0.5369 - val_accuracy: 0.7374 - lr: 0.0020\n",
      "\n",
      "Epoch 7: LearningRateScheduler setting learning rate to 0.001999999862164259.\n",
      "Epoch 7/20\n",
      "80/80 [==============================] - 58s 723ms/step - loss: 0.4401 - accuracy: 0.7941 - val_loss: 0.5541 - val_accuracy: 0.7339 - lr: 0.0020\n",
      "\n",
      "Epoch 8: LearningRateScheduler setting learning rate to 0.001999999862164259.\n",
      "Epoch 8/20\n",
      "80/80 [==============================] - 59s 741ms/step - loss: 0.4305 - accuracy: 0.7990 - val_loss: 0.5519 - val_accuracy: 0.7314 - lr: 0.0020\n",
      "\n",
      "Epoch 9: LearningRateScheduler setting learning rate to 0.001999999862164259.\n",
      "Epoch 9/20\n",
      "80/80 [==============================] - 57s 710ms/step - loss: 0.4240 - accuracy: 0.8031 - val_loss: 0.5539 - val_accuracy: 0.7322 - lr: 0.0020\n",
      "\n",
      "Epoch 10: LearningRateScheduler setting learning rate to 0.001999999862164259.\n",
      "Epoch 10/20\n",
      "80/80 [==============================] - 58s 720ms/step - loss: 0.4167 - accuracy: 0.8066 - val_loss: 0.5598 - val_accuracy: 0.7296 - lr: 0.0020\n"
     ]
    }
   ],
   "source": [
    "batch_size = 1000\n",
    "epochs = 20\n",
    "history = model.fit(x_train, y_train, batch_size=batch_size, epochs=epochs,\n",
    "                    validation_data=(x_test, y_test), callbacks=[early_stopping,lrs])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "id": "7b2ccc17-ae8b-43c5-bdcd-f5084d20e128",
   "metadata": {},
   "outputs": [],
   "source": [
    "model.save('save/model_avanced_lstm/model_biLstm.h5')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b342c874-e69c-4237-8fe7-403d3a6a4c6d",
   "metadata": {},
   "source": [
    "<h1 style = 'color : blue; font-family: cambria; text-align: center'>V- PREDICTION ET VISUALISATION DU RESULTAT</h1>"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "37f9dc35-b55f-4a93-80a4-f0c61d83a8c8",
   "metadata": {},
   "source": [
    "<h3 style = 'color : orange; font-family: cambria'>V.1- Prediction </h3>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "id": "91c60f3c-1803-4bad-9b98-1333354815c2",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x17767d0ff40>"
      ]
     },
     "execution_count": 49,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1500x700 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig,ax=plt.subplots(1,2,figsize=(15,7))\n",
    "ax[0].plot(history.history['accuracy'],c= 'b',label=\"train_accuracy\")\n",
    "ax[0].plot(history.history['val_accuracy'],c= 'r',label=\"val_accuracy\")\n",
    "ax[0].set_xlabel(\"Number of Epochs\")\n",
    "ax[0].set_ylabel(\"Accuracy\")\n",
    "ax[0].legend()\n",
    "\n",
    "ax[1].plot(history.history['loss'],c= 'b',label=\"train_loss\")\n",
    "ax[1].plot(history.history['val_loss'],c= 'r',label=\"val_loss\")\n",
    "ax[1].set_xlabel(\"Number of Epochs\")\n",
    "ax[1].set_ylabel(\"Loss\")\n",
    "ax[1].legend()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "id": "425cfa9d-4f5c-4196-9ae3-0060f1d35fb8",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "20/20 [==============================] - 3s 104ms/step\n"
     ]
    }
   ],
   "source": [
    "def decode_sentiment(score):\n",
    "    return \"Positive\" if score>0.5 else \"Negative\"\n",
    "\n",
    "\n",
    "scores = model.predict(x_test, verbose=1, batch_size=1000)\n",
    "y_pred_1d = [decode_sentiment(score) for score in scores]"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5801d866-5f91-4b10-a2d4-00fef81f38aa",
   "metadata": {},
   "source": [
    "<h3 style = 'color : orange; font-family: cambria'>V.2- Visualisation des resultats</h3>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "id": "d51eba3b-b917-428d-8356-ecbcbb7b8594",
   "metadata": {},
   "outputs": [],
   "source": [
    "### Affichage de la matrice de confusion\n",
    "\n",
    "def plot_confusion_matrix(cm, classes,\n",
    "                          title='Confusion matrix',\n",
    "                          cmap=plt.cm.bwr):\n",
    "    \"\"\"\n",
    "    Cette fonction trace et affiche la matrice de confusin pour ce model.\n",
    "    \"\"\"\n",
    "\n",
    "    cm = cm.astype('float') / cm.sum(axis=1)[:, np.newaxis]\n",
    "\n",
    "    plt.imshow(cm, interpolation='nearest', cmap=cmap)\n",
    "    plt.title(title, fontsize=20)\n",
    "    plt.colorbar()\n",
    "    tick_marks = np.arange(len(classes))\n",
    "    plt.xticks(tick_marks, classes, fontsize=13)\n",
    "    plt.yticks(tick_marks, classes, fontsize=13)\n",
    "\n",
    "    fmt = '.2f'\n",
    "    thresh = cm.max() / 2.\n",
    "    for i, j in itertools.product(range(cm.shape[0]), range(cm.shape[1])):\n",
    "        plt.text(j, i, format(cm[i, j], fmt),\n",
    "                 horizontalalignment=\"center\",\n",
    "                 color=\"white\" if cm[i, j] > thresh else \"black\")\n",
    "\n",
    "    plt.ylabel('True label', fontsize=17)\n",
    "    plt.xlabel('Predicted label', fontsize=17)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "id": "20bb9817-7540-4274-b7eb-876cb2c9a847",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 600x600 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "cnf_matrix = confusion_matrix(test_data.label.to_list(), y_pred_1d)\n",
    "plt.figure(figsize=(6,6))\n",
    "plot_confusion_matrix(cnf_matrix, classes=test_data.label.unique(), title=\"Confusion matrix\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "id": "ba5697db-acbe-4d37-9924-1991f880c17e",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "    Negative       0.73      0.74      0.73     10024\n",
      "    Positive       0.73      0.72      0.73      9976\n",
      "\n",
      "    accuracy                           0.73     20000\n",
      "   macro avg       0.73      0.73      0.73     20000\n",
      "weighted avg       0.73      0.73      0.73     20000\n",
      "\n"
     ]
    }
   ],
   "source": [
    "#Tableau des precisions du model\n",
    "print(classification_report(list(test_data.label), y_pred_1d))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2c7a49a6-e1c9-4726-98bd-a3fdcb2b2833",
   "metadata": {},
   "source": [
    "<h3 style = 'color : orange; font-family: cambria; text-align: center;'>CONCLUSION</h3>\n",
    "<p style = 'border-bottom: 2px solid red'>Le Notebook est une forme experimentale d'analyse et d'entrainement d'un modèle d'analyse de données.Le modèle  entraîné a une precision de 73%<br/>\n",
    "A la fin de l'entraînement du modèle final (sur le datasets final), il faudra  penser au prametrage des hyperparamètres avec keras Tuner.<br/></p>\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f9453f6e",
   "metadata": {},
   "source": []
  }
 ],
 "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": "3.9.13"
  },
  "vscode": {
   "interpreter": {
    "hash": "bf22bc0a685d21803e24b22b4e95183ec0fe5ee48047b11b36126386770f7c81"
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}