GaussianNB_model.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "import time\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "from IPython.display import display\n",
    "from sklearn import metrics\n",
    "from sklearn.feature_extraction.text import CountVectorizer, TfidfVectorizer\n",
    "from sklearn.model_selection import GridSearchCV, train_test_split\n",
    "from sklearn.naive_bayes import GaussianNB\n",
    "from sklearn.pipeline import Pipeline\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "\n",
    "t = time.time()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 0 - Création des fonctions utiles"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "def sentiment140(stop_words, lemmatization, negation, nb_data):\n",
    "    \"\"\"Fonction permettant d'ouvrir un échantillon de la base Sentiment140 préprocessée avec la combinaison \n",
    "    de booléens (stop_words, lemmatization, negation)\"\"\"\n",
    "    \n",
    "    file = \"train\"\n",
    "    if stop_words:\n",
    "        file += \"_stop\"\n",
    "    if lemmatization:\n",
    "        file += \"_lemm\"\n",
    "    if negation:\n",
    "        file += \"_neg\"\n",
    "    return pd.read_pickle(os.path.join(\"data\", \"Sentiment140\", file + \".bz2\")).sample(nb_data, random_state=1234)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>sentiment</th>\n",
       "      <th>id</th>\n",
       "      <th>date</th>\n",
       "      <th>query_string</th>\n",
       "      <th>user</th>\n",
       "      <th>text</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>599410</th>\n",
       "      <td>0</td>\n",
       "      <td>2220663113</td>\n",
       "      <td>Thu Jun 18 03:39:49 PDT 2009</td>\n",
       "      <td>NO_QUERY</td>\n",
       "      <td>gemgembam</td>\n",
       "      <td>open itunes will not open naooo the silence is...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>549856</th>\n",
       "      <td>0</td>\n",
       "      <td>2202759103</td>\n",
       "      <td>Tue Jun 16 22:21:43 PDT 2009</td>\n",
       "      <td>NO_QUERY</td>\n",
       "      <td>twistofcolour</td>\n",
       "      <td>the sun is already gone for the day</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1363048</th>\n",
       "      <td>4</td>\n",
       "      <td>2049441111</td>\n",
       "      <td>Fri Jun 05 16:53:10 PDT 2009</td>\n",
       "      <td>NO_QUERY</td>\n",
       "      <td>Niico1e</td>\n",
       "      <td>hiya this is my st tweet i am so tired though ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>558776</th>\n",
       "      <td>0</td>\n",
       "      <td>2204830292</td>\n",
       "      <td>Wed Jun 17 03:14:38 PDT 2009</td>\n",
       "      <td>NO_QUERY</td>\n",
       "      <td>MeriEsteban</td>\n",
       "      <td>i have the bac today at p m i do not know if i...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1582398</th>\n",
       "      <td>4</td>\n",
       "      <td>2190306594</td>\n",
       "      <td>Tue Jun 16 02:21:03 PDT 2009</td>\n",
       "      <td>NO_QUERY</td>\n",
       "      <td>MeghanMagnolia</td>\n",
       "      <td>as far as i remember you were not half bad ine...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1353343</th>\n",
       "      <td>4</td>\n",
       "      <td>2046706853</td>\n",
       "      <td>Fri Jun 05 12:34:26 PDT 2009</td>\n",
       "      <td>NO_QUERY</td>\n",
       "      <td>WahooPooh</td>\n",
       "      <td>as tempting as that is i am just waiting on my...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>586354</th>\n",
       "      <td>0</td>\n",
       "      <td>2215937155</td>\n",
       "      <td>Wed Jun 17 18:49:10 PDT 2009</td>\n",
       "      <td>NO_QUERY</td>\n",
       "      <td>amandarie</td>\n",
       "      <td>cool how i have to be stuck at home i want to ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>860355</th>\n",
       "      <td>4</td>\n",
       "      <td>1676343245</td>\n",
       "      <td>Fri May 01 21:46:07 PDT 2009</td>\n",
       "      <td>NO_QUERY</td>\n",
       "      <td>MatchesMalone</td>\n",
       "      <td>oh and followfriday my writing partner because...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>521668</th>\n",
       "      <td>0</td>\n",
       "      <td>2192486836</td>\n",
       "      <td>Tue Jun 16 07:06:35 PDT 2009</td>\n",
       "      <td>NO_QUERY</td>\n",
       "      <td>hockeyfan45214</td>\n",
       "      <td>you would think if i continue to block someone...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>243888</th>\n",
       "      <td>0</td>\n",
       "      <td>1981647618</td>\n",
       "      <td>Sun May 31 09:35:54 PDT 2009</td>\n",
       "      <td>NO_QUERY</td>\n",
       "      <td>ilovemytroops</td>\n",
       "      <td>nevaeh buchanan missing friends of her mother ...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         sentiment          id                          date query_string  \\\n",
       "599410           0  2220663113  Thu Jun 18 03:39:49 PDT 2009     NO_QUERY   \n",
       "549856           0  2202759103  Tue Jun 16 22:21:43 PDT 2009     NO_QUERY   \n",
       "1363048          4  2049441111  Fri Jun 05 16:53:10 PDT 2009     NO_QUERY   \n",
       "558776           0  2204830292  Wed Jun 17 03:14:38 PDT 2009     NO_QUERY   \n",
       "1582398          4  2190306594  Tue Jun 16 02:21:03 PDT 2009     NO_QUERY   \n",
       "1353343          4  2046706853  Fri Jun 05 12:34:26 PDT 2009     NO_QUERY   \n",
       "586354           0  2215937155  Wed Jun 17 18:49:10 PDT 2009     NO_QUERY   \n",
       "860355           4  1676343245  Fri May 01 21:46:07 PDT 2009     NO_QUERY   \n",
       "521668           0  2192486836  Tue Jun 16 07:06:35 PDT 2009     NO_QUERY   \n",
       "243888           0  1981647618  Sun May 31 09:35:54 PDT 2009     NO_QUERY   \n",
       "\n",
       "                   user                                               text  \n",
       "599410        gemgembam  open itunes will not open naooo the silence is...  \n",
       "549856    twistofcolour                the sun is already gone for the day  \n",
       "1363048         Niico1e  hiya this is my st tweet i am so tired though ...  \n",
       "558776      MeriEsteban  i have the bac today at p m i do not know if i...  \n",
       "1582398  MeghanMagnolia  as far as i remember you were not half bad ine...  \n",
       "1353343       WahooPooh  as tempting as that is i am just waiting on my...  \n",
       "586354        amandarie  cool how i have to be stuck at home i want to ...  \n",
       "860355    MatchesMalone  oh and followfriday my writing partner because...  \n",
       "521668   hockeyfan45214  you would think if i continue to block someone...  \n",
       "243888    ilovemytroops  nevaeh buchanan missing friends of her mother ...  "
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sentiment140(False, False, False, 10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "class Modelisation():\n",
    "    \"\"\"Classe permettant d'entrainer un modèle et de l'évaluer\"\"\"\n",
    "    \n",
    "    def __init__(self, X, y, vectorizer, model, scaling=True):\n",
    "        X = vectorizer.fit_transform(X)\n",
    "\n",
    "        if scaling:\n",
    "            scaler = StandardScaler(with_mean=False)\n",
    "            X = scaler.fit_transform(X)\n",
    "        \n",
    "        if isinstance(model, GaussianNB):\n",
    "            X = X.toarray()\n",
    "\n",
    "        X_train, X_test, y_train, y_test = train_test_split(X, y, train_size=0.80, random_state=1234)\n",
    "\n",
    "        model.fit(X_train, y_train)\n",
    "        y_pred = model.predict(X_test)\n",
    "\n",
    "        sc_accuracy = metrics.accuracy_score(y_pred, y_test)    \n",
    "        sc_balanced_accuracy = metrics.balanced_accuracy_score(y_pred, y_test)\n",
    "        sc_roc_auc = metrics.roc_auc_score(y_test, model.predict_proba(X_test)[:, 1])\n",
    "        probs = model.predict_proba(X_test)[:, 1]\n",
    "\n",
    "        metrics_score = {'accuracy': sc_accuracy, 'balanced_accuracy': sc_balanced_accuracy, 'roc_auc': sc_roc_auc}\n",
    "\n",
    "        self.X_train, self.X_test, self.y_train, self.y_test = X_train, X_test, y_train, y_test\n",
    "        self.model = model\n",
    "        self.probs = probs\n",
    "        self.metrics_score = metrics_score\n",
    "    \n",
    "    def get_data(self):\n",
    "        return self.X_train, self.X_test, self.y_train, self.y_test\n",
    "    \n",
    "    def show_conf_matrix(self):\n",
    "        metrics.plot_confusion_matrix(self.model, self.X_test, self.y_test, cmap='Blues')\n",
    "        plt.show()\n",
    "\n",
    "        print(f\"Accuracy : {self.metrics_score['accuracy']:.4f}\")\n",
    "        print(f\"Balanced accuracy : {self.metrics_score['balanced_accuracy']:.4f}\")\n",
    "        print(f\"ROC AUC : {self.metrics_score['roc_auc']:.4f}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "def comparaison(X, y, N, model, scaling=True, show_graph=True):\n",
    "    \"\"\"\n",
    "    Création d'un DataFrame qui contient les performances des différents couples (vectorizer, model) étudiés\n",
    "    Les performances regardées sont l'aire sous la courbe ROC, la précision et la \"balanced accuracy\"\n",
    "    N correspond au nombre de features que vont utiliser les vectorizers\n",
    "    \"\"\"\n",
    "    \n",
    "    table = []\n",
    "    \n",
    "    m = Modelisation(X, y, CountVectorizer(max_features = N), model)\n",
    "    table.append(m.metrics_score)\n",
    "\n",
    "    m = Modelisation(X, y, TfidfVectorizer(max_features = N), model)\n",
    "    table.append(m.metrics_score)\n",
    "\n",
    "    m = Modelisation(X, y, CountVectorizer(max_features = N, ngram_range=(1,2)), model)\n",
    "    table.append(m.metrics_score)\n",
    "\n",
    "    m = Modelisation(X, y, TfidfVectorizer(max_features = N, ngram_range=(1,2)), model)\n",
    "    table.append(m.metrics_score)\n",
    "    \n",
    "    \n",
    "    df = pd.DataFrame(table, index=[f\"CountVectorizer {N}\", f\"TfidfVectorizer {N}\", f\"CountVectorizer {N} ngram 2\", f\"TfidfVectorizer {N} ngram 2\"])\n",
    "    \n",
    "    return df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "def influence_preprocessing(N, model, show_graph=False, show_table=False):\n",
    "    \"\"\"Création d'un DataFrame qui regroupe les performances des couples (vectorizer, model) sur les différents \n",
    "    \"états de preprocessing\" de la base Sentiment140, en fonction de la combinaison (stop_words, lemmatization, negation) \"\"\"\n",
    "    \n",
    "    tables = []\n",
    "    for stop_words in [False, True]:\n",
    "        for lemmatization in [False, True]:\n",
    "            for negation in [False, True]:\n",
    "                print(f\"stop_words: {stop_words}, lemmatization: {lemmatization}, negation: {negation}\")\n",
    "                df_temp = sentiment140(stop_words, lemmatization, negation, NB_DATA)\n",
    "                X = df_temp.text.to_list()\n",
    "                y = df_temp.sentiment.to_list()            \n",
    "                df_metrics = comparaison(X, y, N, model, show_graph=show_graph)\n",
    "                if show_table:\n",
    "                    display(df_metrics)\n",
    "                df_metrics['stop_words'] = stop_words\n",
    "                df_metrics['lemmatization'] = lemmatization\n",
    "                df_metrics['negation'] = negation\n",
    "                tables.append(df_metrics)\n",
    "    return pd.concat(tables)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "def influence_features(X, y, max_features_list, model, scaling=True):\n",
    "    \"\"\"Comparaison des performances des couples (vectorizer, model) en faisant varier le nombre de features\"\"\"\n",
    "    \n",
    "    table = []\n",
    "    for max_features in max_features_list:\n",
    "        m = Modelisation(X, y, CountVectorizer(max_features=max_features), model)\n",
    "        metrics_score = m.metrics_score\n",
    "        metrics_score['features'] = max_features\n",
    "        metrics_score['method'] = \"CountVectorizer\"\n",
    "        table.append(metrics_score)\n",
    "\n",
    "        m = Modelisation(X, y, TfidfVectorizer(max_features=max_features), model)\n",
    "        metrics_score = m.metrics_score\n",
    "        metrics_score['features'] = max_features\n",
    "        metrics_score['method'] = \"TfidfVectorizer\"\n",
    "        table.append(metrics_score)\n",
    "\n",
    "        m = Modelisation(X, y, CountVectorizer(max_features=max_features, ngram_range=(1, 2)), model)\n",
    "        metrics_score = m.metrics_score\n",
    "        metrics_score['features'] = max_features\n",
    "        metrics_score['method'] = \"CountVectorizer ngram 2\"\n",
    "        table.append(metrics_score)\n",
    "\n",
    "        m = Modelisation(X, y, TfidfVectorizer(max_features=max_features, ngram_range=(1, 2)), model)\n",
    "        metrics_score = m.metrics_score\n",
    "        metrics_score['features'] = max_features\n",
    "        metrics_score['method'] = \"TfidfVectorizer ngram 2\"\n",
    "        table.append(metrics_score)\n",
    "\n",
    "    df = pd.DataFrame(table)\n",
    "\n",
    "    methods = list(set(df.method))\n",
    "    scores = list(df.columns[:-2])\n",
    "    fig, axes = plt.subplots(1, len(scores), figsize=(20, 6))\n",
    "    for i_score in range(len(scores)):\n",
    "        for j_method in range(len(methods)):\n",
    "            df[df.method == methods[j_method]].plot(x='features', y=scores[i_score], label=methods[j_method], ax=axes[i_score])\n",
    "        axes[i_score].set(xlabel='Nombre de features')\n",
    "        axes[i_score].legend()\n",
    "        axes[i_score].set_title(scores[i_score])\n",
    "    plt.show()\n",
    "    return df"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 1 - Sélection des meilleurs paramètres du modèle Gaussian Naive Bayes"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "On commence par choisir le nombre de lignes que l'on extrait de la base Sentiment140, sur laquelle on va entraîner et évaluer notre modèle."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "NB_DATA = 50000"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 1.1 - Comparaison des méthodes de preprocessing et de \"feature extraction\" (vectorizer)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "stop_words: False, lemmatization: False, negation: False\n",
      "stop_words: False, lemmatization: False, negation: True\n",
      "stop_words: False, lemmatization: True, negation: False\n",
      "stop_words: False, lemmatization: True, negation: True\n",
      "stop_words: True, lemmatization: False, negation: False\n",
      "stop_words: True, lemmatization: False, negation: True\n",
      "stop_words: True, lemmatization: True, negation: False\n",
      "stop_words: True, lemmatization: True, negation: True\n",
      "Wall time: 4min 52s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "df_metrics = influence_preprocessing(2500, GaussianNB())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>accuracy</th>\n",
       "      <th>balanced_accuracy</th>\n",
       "      <th>roc_auc</th>\n",
       "      <th>stop_words</th>\n",
       "      <th>lemmatization</th>\n",
       "      <th>negation</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2500 ngram 2</th>\n",
       "      <td>0.7325</td>\n",
       "      <td>0.732387</td>\n",
       "      <td>0.772568</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2500 ngram 2</th>\n",
       "      <td>0.7298</td>\n",
       "      <td>0.729708</td>\n",
       "      <td>0.771818</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2500 ngram 2</th>\n",
       "      <td>0.7319</td>\n",
       "      <td>0.731856</td>\n",
       "      <td>0.768116</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2500 ngram 2</th>\n",
       "      <td>0.7291</td>\n",
       "      <td>0.729122</td>\n",
       "      <td>0.766167</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2500 ngram 2</th>\n",
       "      <td>0.7247</td>\n",
       "      <td>0.725874</td>\n",
       "      <td>0.755659</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2500 ngram 2</th>\n",
       "      <td>0.7247</td>\n",
       "      <td>0.725352</td>\n",
       "      <td>0.754344</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2500 ngram 2</th>\n",
       "      <td>0.7241</td>\n",
       "      <td>0.725575</td>\n",
       "      <td>0.751736</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2500 ngram 2</th>\n",
       "      <td>0.7237</td>\n",
       "      <td>0.724857</td>\n",
       "      <td>0.751676</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2500 ngram 2</th>\n",
       "      <td>0.7030</td>\n",
       "      <td>0.703174</td>\n",
       "      <td>0.733989</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2500 ngram 2</th>\n",
       "      <td>0.7058</td>\n",
       "      <td>0.706936</td>\n",
       "      <td>0.729322</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2500 ngram 2</th>\n",
       "      <td>0.7012</td>\n",
       "      <td>0.702338</td>\n",
       "      <td>0.727278</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2500 ngram 2</th>\n",
       "      <td>0.7010</td>\n",
       "      <td>0.702112</td>\n",
       "      <td>0.721867</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2500 ngram 2</th>\n",
       "      <td>0.6990</td>\n",
       "      <td>0.700200</td>\n",
       "      <td>0.721122</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2500</th>\n",
       "      <td>0.6953</td>\n",
       "      <td>0.699267</td>\n",
       "      <td>0.715957</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2500 ngram 2</th>\n",
       "      <td>0.6942</td>\n",
       "      <td>0.698566</td>\n",
       "      <td>0.714059</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2500</th>\n",
       "      <td>0.6950</td>\n",
       "      <td>0.698409</td>\n",
       "      <td>0.713616</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2500 ngram 2</th>\n",
       "      <td>0.6895</td>\n",
       "      <td>0.693454</td>\n",
       "      <td>0.710520</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2500</th>\n",
       "      <td>0.6930</td>\n",
       "      <td>0.697896</td>\n",
       "      <td>0.709352</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2500</th>\n",
       "      <td>0.6885</td>\n",
       "      <td>0.691590</td>\n",
       "      <td>0.704389</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2500 ngram 2</th>\n",
       "      <td>0.6834</td>\n",
       "      <td>0.688448</td>\n",
       "      <td>0.703298</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2500</th>\n",
       "      <td>0.6792</td>\n",
       "      <td>0.683262</td>\n",
       "      <td>0.702315</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2500</th>\n",
       "      <td>0.6792</td>\n",
       "      <td>0.680431</td>\n",
       "      <td>0.701486</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2500</th>\n",
       "      <td>0.6829</td>\n",
       "      <td>0.691391</td>\n",
       "      <td>0.699121</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2500</th>\n",
       "      <td>0.6754</td>\n",
       "      <td>0.680024</td>\n",
       "      <td>0.692833</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2500</th>\n",
       "      <td>0.6746</td>\n",
       "      <td>0.685422</td>\n",
       "      <td>0.692313</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2500</th>\n",
       "      <td>0.6710</td>\n",
       "      <td>0.673124</td>\n",
       "      <td>0.688084</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2500</th>\n",
       "      <td>0.6686</td>\n",
       "      <td>0.674506</td>\n",
       "      <td>0.685616</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2500</th>\n",
       "      <td>0.6678</td>\n",
       "      <td>0.676039</td>\n",
       "      <td>0.682113</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2500</th>\n",
       "      <td>0.6631</td>\n",
       "      <td>0.675652</td>\n",
       "      <td>0.678810</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2500</th>\n",
       "      <td>0.6573</td>\n",
       "      <td>0.669733</td>\n",
       "      <td>0.672506</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2500</th>\n",
       "      <td>0.6478</td>\n",
       "      <td>0.655066</td>\n",
       "      <td>0.661374</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2500</th>\n",
       "      <td>0.6423</td>\n",
       "      <td>0.655489</td>\n",
       "      <td>0.657447</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                              accuracy  balanced_accuracy   roc_auc  \\\n",
       "TfidfVectorizer 2500 ngram 2    0.7325           0.732387  0.772568   \n",
       "TfidfVectorizer 2500 ngram 2    0.7298           0.729708  0.771818   \n",
       "TfidfVectorizer 2500 ngram 2    0.7319           0.731856  0.768116   \n",
       "TfidfVectorizer 2500 ngram 2    0.7291           0.729122  0.766167   \n",
       "CountVectorizer 2500 ngram 2    0.7247           0.725874  0.755659   \n",
       "CountVectorizer 2500 ngram 2    0.7247           0.725352  0.754344   \n",
       "CountVectorizer 2500 ngram 2    0.7241           0.725575  0.751736   \n",
       "CountVectorizer 2500 ngram 2    0.7237           0.724857  0.751676   \n",
       "TfidfVectorizer 2500 ngram 2    0.7030           0.703174  0.733989   \n",
       "TfidfVectorizer 2500 ngram 2    0.7058           0.706936  0.729322   \n",
       "CountVectorizer 2500 ngram 2    0.7012           0.702338  0.727278   \n",
       "TfidfVectorizer 2500 ngram 2    0.7010           0.702112  0.721867   \n",
       "TfidfVectorizer 2500 ngram 2    0.6990           0.700200  0.721122   \n",
       "TfidfVectorizer 2500            0.6953           0.699267  0.715957   \n",
       "CountVectorizer 2500 ngram 2    0.6942           0.698566  0.714059   \n",
       "TfidfVectorizer 2500            0.6950           0.698409  0.713616   \n",
       "CountVectorizer 2500 ngram 2    0.6895           0.693454  0.710520   \n",
       "TfidfVectorizer 2500            0.6930           0.697896  0.709352   \n",
       "TfidfVectorizer 2500            0.6885           0.691590  0.704389   \n",
       "CountVectorizer 2500 ngram 2    0.6834           0.688448  0.703298   \n",
       "TfidfVectorizer 2500            0.6792           0.683262  0.702315   \n",
       "TfidfVectorizer 2500            0.6792           0.680431  0.701486   \n",
       "CountVectorizer 2500            0.6829           0.691391  0.699121   \n",
       "TfidfVectorizer 2500            0.6754           0.680024  0.692833   \n",
       "CountVectorizer 2500            0.6746           0.685422  0.692313   \n",
       "TfidfVectorizer 2500            0.6710           0.673124  0.688084   \n",
       "CountVectorizer 2500            0.6686           0.674506  0.685616   \n",
       "CountVectorizer 2500            0.6678           0.676039  0.682113   \n",
       "CountVectorizer 2500            0.6631           0.675652  0.678810   \n",
       "CountVectorizer 2500            0.6573           0.669733  0.672506   \n",
       "CountVectorizer 2500            0.6478           0.655066  0.661374   \n",
       "CountVectorizer 2500            0.6423           0.655489  0.657447   \n",
       "\n",
       "                              stop_words  lemmatization  negation  \n",
       "TfidfVectorizer 2500 ngram 2       False           True     False  \n",
       "TfidfVectorizer 2500 ngram 2       False          False     False  \n",
       "TfidfVectorizer 2500 ngram 2       False          False      True  \n",
       "TfidfVectorizer 2500 ngram 2       False           True      True  \n",
       "CountVectorizer 2500 ngram 2       False          False      True  \n",
       "CountVectorizer 2500 ngram 2       False          False     False  \n",
       "CountVectorizer 2500 ngram 2       False           True      True  \n",
       "CountVectorizer 2500 ngram 2       False           True     False  \n",
       "TfidfVectorizer 2500 ngram 2        True          False      True  \n",
       "TfidfVectorizer 2500 ngram 2        True          False     False  \n",
       "CountVectorizer 2500 ngram 2        True          False      True  \n",
       "TfidfVectorizer 2500 ngram 2        True           True     False  \n",
       "TfidfVectorizer 2500 ngram 2        True           True      True  \n",
       "TfidfVectorizer 2500               False          False      True  \n",
       "CountVectorizer 2500 ngram 2        True          False     False  \n",
       "TfidfVectorizer 2500                True          False      True  \n",
       "CountVectorizer 2500 ngram 2        True           True      True  \n",
       "TfidfVectorizer 2500               False           True      True  \n",
       "TfidfVectorizer 2500                True           True      True  \n",
       "CountVectorizer 2500 ngram 2        True           True     False  \n",
       "TfidfVectorizer 2500               False          False     False  \n",
       "TfidfVectorizer 2500                True          False     False  \n",
       "CountVectorizer 2500                True          False      True  \n",
       "TfidfVectorizer 2500               False           True     False  \n",
       "CountVectorizer 2500               False          False      True  \n",
       "TfidfVectorizer 2500                True           True     False  \n",
       "CountVectorizer 2500                True          False     False  \n",
       "CountVectorizer 2500                True           True      True  \n",
       "CountVectorizer 2500               False           True      True  \n",
       "CountVectorizer 2500               False          False     False  \n",
       "CountVectorizer 2500                True           True     False  \n",
       "CountVectorizer 2500               False           True     False  "
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_metrics.sort_values(by=\"roc_auc\", ascending=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>accuracy</th>\n",
       "      <th>balanced_accuracy</th>\n",
       "      <th>roc_auc</th>\n",
       "      <th>stop_words</th>\n",
       "      <th>lemmatization</th>\n",
       "      <th>negation</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2500 ngram 2</th>\n",
       "      <td>0.7325</td>\n",
       "      <td>0.732387</td>\n",
       "      <td>0.772568</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2500 ngram 2</th>\n",
       "      <td>0.7319</td>\n",
       "      <td>0.731856</td>\n",
       "      <td>0.768116</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2500 ngram 2</th>\n",
       "      <td>0.7298</td>\n",
       "      <td>0.729708</td>\n",
       "      <td>0.771818</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2500 ngram 2</th>\n",
       "      <td>0.7291</td>\n",
       "      <td>0.729122</td>\n",
       "      <td>0.766167</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2500 ngram 2</th>\n",
       "      <td>0.7247</td>\n",
       "      <td>0.725874</td>\n",
       "      <td>0.755659</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2500 ngram 2</th>\n",
       "      <td>0.7241</td>\n",
       "      <td>0.725575</td>\n",
       "      <td>0.751736</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2500 ngram 2</th>\n",
       "      <td>0.7247</td>\n",
       "      <td>0.725352</td>\n",
       "      <td>0.754344</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2500 ngram 2</th>\n",
       "      <td>0.7237</td>\n",
       "      <td>0.724857</td>\n",
       "      <td>0.751676</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2500 ngram 2</th>\n",
       "      <td>0.7058</td>\n",
       "      <td>0.706936</td>\n",
       "      <td>0.729322</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2500 ngram 2</th>\n",
       "      <td>0.7030</td>\n",
       "      <td>0.703174</td>\n",
       "      <td>0.733989</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2500 ngram 2</th>\n",
       "      <td>0.7012</td>\n",
       "      <td>0.702338</td>\n",
       "      <td>0.727278</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2500 ngram 2</th>\n",
       "      <td>0.7010</td>\n",
       "      <td>0.702112</td>\n",
       "      <td>0.721867</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2500 ngram 2</th>\n",
       "      <td>0.6990</td>\n",
       "      <td>0.700200</td>\n",
       "      <td>0.721122</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2500</th>\n",
       "      <td>0.6953</td>\n",
       "      <td>0.699267</td>\n",
       "      <td>0.715957</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2500 ngram 2</th>\n",
       "      <td>0.6942</td>\n",
       "      <td>0.698566</td>\n",
       "      <td>0.714059</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2500</th>\n",
       "      <td>0.6950</td>\n",
       "      <td>0.698409</td>\n",
       "      <td>0.713616</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2500</th>\n",
       "      <td>0.6930</td>\n",
       "      <td>0.697896</td>\n",
       "      <td>0.709352</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2500 ngram 2</th>\n",
       "      <td>0.6895</td>\n",
       "      <td>0.693454</td>\n",
       "      <td>0.710520</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2500</th>\n",
       "      <td>0.6885</td>\n",
       "      <td>0.691590</td>\n",
       "      <td>0.704389</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2500</th>\n",
       "      <td>0.6829</td>\n",
       "      <td>0.691391</td>\n",
       "      <td>0.699121</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2500 ngram 2</th>\n",
       "      <td>0.6834</td>\n",
       "      <td>0.688448</td>\n",
       "      <td>0.703298</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2500</th>\n",
       "      <td>0.6746</td>\n",
       "      <td>0.685422</td>\n",
       "      <td>0.692313</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2500</th>\n",
       "      <td>0.6792</td>\n",
       "      <td>0.683262</td>\n",
       "      <td>0.702315</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2500</th>\n",
       "      <td>0.6792</td>\n",
       "      <td>0.680431</td>\n",
       "      <td>0.701486</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2500</th>\n",
       "      <td>0.6754</td>\n",
       "      <td>0.680024</td>\n",
       "      <td>0.692833</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2500</th>\n",
       "      <td>0.6678</td>\n",
       "      <td>0.676039</td>\n",
       "      <td>0.682113</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2500</th>\n",
       "      <td>0.6631</td>\n",
       "      <td>0.675652</td>\n",
       "      <td>0.678810</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2500</th>\n",
       "      <td>0.6686</td>\n",
       "      <td>0.674506</td>\n",
       "      <td>0.685616</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2500</th>\n",
       "      <td>0.6710</td>\n",
       "      <td>0.673124</td>\n",
       "      <td>0.688084</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2500</th>\n",
       "      <td>0.6573</td>\n",
       "      <td>0.669733</td>\n",
       "      <td>0.672506</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2500</th>\n",
       "      <td>0.6423</td>\n",
       "      <td>0.655489</td>\n",
       "      <td>0.657447</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2500</th>\n",
       "      <td>0.6478</td>\n",
       "      <td>0.655066</td>\n",
       "      <td>0.661374</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                              accuracy  balanced_accuracy   roc_auc  \\\n",
       "TfidfVectorizer 2500 ngram 2    0.7325           0.732387  0.772568   \n",
       "TfidfVectorizer 2500 ngram 2    0.7319           0.731856  0.768116   \n",
       "TfidfVectorizer 2500 ngram 2    0.7298           0.729708  0.771818   \n",
       "TfidfVectorizer 2500 ngram 2    0.7291           0.729122  0.766167   \n",
       "CountVectorizer 2500 ngram 2    0.7247           0.725874  0.755659   \n",
       "CountVectorizer 2500 ngram 2    0.7241           0.725575  0.751736   \n",
       "CountVectorizer 2500 ngram 2    0.7247           0.725352  0.754344   \n",
       "CountVectorizer 2500 ngram 2    0.7237           0.724857  0.751676   \n",
       "TfidfVectorizer 2500 ngram 2    0.7058           0.706936  0.729322   \n",
       "TfidfVectorizer 2500 ngram 2    0.7030           0.703174  0.733989   \n",
       "CountVectorizer 2500 ngram 2    0.7012           0.702338  0.727278   \n",
       "TfidfVectorizer 2500 ngram 2    0.7010           0.702112  0.721867   \n",
       "TfidfVectorizer 2500 ngram 2    0.6990           0.700200  0.721122   \n",
       "TfidfVectorizer 2500            0.6953           0.699267  0.715957   \n",
       "CountVectorizer 2500 ngram 2    0.6942           0.698566  0.714059   \n",
       "TfidfVectorizer 2500            0.6950           0.698409  0.713616   \n",
       "TfidfVectorizer 2500            0.6930           0.697896  0.709352   \n",
       "CountVectorizer 2500 ngram 2    0.6895           0.693454  0.710520   \n",
       "TfidfVectorizer 2500            0.6885           0.691590  0.704389   \n",
       "CountVectorizer 2500            0.6829           0.691391  0.699121   \n",
       "CountVectorizer 2500 ngram 2    0.6834           0.688448  0.703298   \n",
       "CountVectorizer 2500            0.6746           0.685422  0.692313   \n",
       "TfidfVectorizer 2500            0.6792           0.683262  0.702315   \n",
       "TfidfVectorizer 2500            0.6792           0.680431  0.701486   \n",
       "TfidfVectorizer 2500            0.6754           0.680024  0.692833   \n",
       "CountVectorizer 2500            0.6678           0.676039  0.682113   \n",
       "CountVectorizer 2500            0.6631           0.675652  0.678810   \n",
       "CountVectorizer 2500            0.6686           0.674506  0.685616   \n",
       "TfidfVectorizer 2500            0.6710           0.673124  0.688084   \n",
       "CountVectorizer 2500            0.6573           0.669733  0.672506   \n",
       "CountVectorizer 2500            0.6423           0.655489  0.657447   \n",
       "CountVectorizer 2500            0.6478           0.655066  0.661374   \n",
       "\n",
       "                              stop_words  lemmatization  negation  \n",
       "TfidfVectorizer 2500 ngram 2       False           True     False  \n",
       "TfidfVectorizer 2500 ngram 2       False          False      True  \n",
       "TfidfVectorizer 2500 ngram 2       False          False     False  \n",
       "TfidfVectorizer 2500 ngram 2       False           True      True  \n",
       "CountVectorizer 2500 ngram 2       False          False      True  \n",
       "CountVectorizer 2500 ngram 2       False           True      True  \n",
       "CountVectorizer 2500 ngram 2       False          False     False  \n",
       "CountVectorizer 2500 ngram 2       False           True     False  \n",
       "TfidfVectorizer 2500 ngram 2        True          False     False  \n",
       "TfidfVectorizer 2500 ngram 2        True          False      True  \n",
       "CountVectorizer 2500 ngram 2        True          False      True  \n",
       "TfidfVectorizer 2500 ngram 2        True           True     False  \n",
       "TfidfVectorizer 2500 ngram 2        True           True      True  \n",
       "TfidfVectorizer 2500               False          False      True  \n",
       "CountVectorizer 2500 ngram 2        True          False     False  \n",
       "TfidfVectorizer 2500                True          False      True  \n",
       "TfidfVectorizer 2500               False           True      True  \n",
       "CountVectorizer 2500 ngram 2        True           True      True  \n",
       "TfidfVectorizer 2500                True           True      True  \n",
       "CountVectorizer 2500                True          False      True  \n",
       "CountVectorizer 2500 ngram 2        True           True     False  \n",
       "CountVectorizer 2500               False          False      True  \n",
       "TfidfVectorizer 2500               False          False     False  \n",
       "TfidfVectorizer 2500                True          False     False  \n",
       "TfidfVectorizer 2500               False           True     False  \n",
       "CountVectorizer 2500                True           True      True  \n",
       "CountVectorizer 2500               False           True      True  \n",
       "CountVectorizer 2500                True          False     False  \n",
       "TfidfVectorizer 2500                True           True     False  \n",
       "CountVectorizer 2500               False          False     False  \n",
       "CountVectorizer 2500               False           True     False  \n",
       "CountVectorizer 2500                True           True     False  "
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_metrics.sort_values(by=\"balanced_accuracy\", ascending=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>accuracy</th>\n",
       "      <th>balanced_accuracy</th>\n",
       "      <th>roc_auc</th>\n",
       "      <th>stop_words</th>\n",
       "      <th>lemmatization</th>\n",
       "      <th>negation</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2500 ngram 2</th>\n",
       "      <td>0.7325</td>\n",
       "      <td>0.732387</td>\n",
       "      <td>0.772568</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2500 ngram 2</th>\n",
       "      <td>0.7298</td>\n",
       "      <td>0.729708</td>\n",
       "      <td>0.771818</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2500 ngram 2</th>\n",
       "      <td>0.7319</td>\n",
       "      <td>0.731856</td>\n",
       "      <td>0.768116</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2500 ngram 2</th>\n",
       "      <td>0.7291</td>\n",
       "      <td>0.729122</td>\n",
       "      <td>0.766167</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2500 ngram 2</th>\n",
       "      <td>0.7247</td>\n",
       "      <td>0.725874</td>\n",
       "      <td>0.755659</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2500 ngram 2</th>\n",
       "      <td>0.7247</td>\n",
       "      <td>0.725352</td>\n",
       "      <td>0.754344</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2500 ngram 2</th>\n",
       "      <td>0.7241</td>\n",
       "      <td>0.725575</td>\n",
       "      <td>0.751736</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2500 ngram 2</th>\n",
       "      <td>0.7237</td>\n",
       "      <td>0.724857</td>\n",
       "      <td>0.751676</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2500 ngram 2</th>\n",
       "      <td>0.7030</td>\n",
       "      <td>0.703174</td>\n",
       "      <td>0.733989</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2500 ngram 2</th>\n",
       "      <td>0.7058</td>\n",
       "      <td>0.706936</td>\n",
       "      <td>0.729322</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2500 ngram 2</th>\n",
       "      <td>0.7012</td>\n",
       "      <td>0.702338</td>\n",
       "      <td>0.727278</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2500 ngram 2</th>\n",
       "      <td>0.7010</td>\n",
       "      <td>0.702112</td>\n",
       "      <td>0.721867</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2500 ngram 2</th>\n",
       "      <td>0.6990</td>\n",
       "      <td>0.700200</td>\n",
       "      <td>0.721122</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2500</th>\n",
       "      <td>0.6953</td>\n",
       "      <td>0.699267</td>\n",
       "      <td>0.715957</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2500 ngram 2</th>\n",
       "      <td>0.6942</td>\n",
       "      <td>0.698566</td>\n",
       "      <td>0.714059</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2500</th>\n",
       "      <td>0.6950</td>\n",
       "      <td>0.698409</td>\n",
       "      <td>0.713616</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2500 ngram 2</th>\n",
       "      <td>0.6895</td>\n",
       "      <td>0.693454</td>\n",
       "      <td>0.710520</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2500</th>\n",
       "      <td>0.6930</td>\n",
       "      <td>0.697896</td>\n",
       "      <td>0.709352</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2500</th>\n",
       "      <td>0.6885</td>\n",
       "      <td>0.691590</td>\n",
       "      <td>0.704389</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2500 ngram 2</th>\n",
       "      <td>0.6834</td>\n",
       "      <td>0.688448</td>\n",
       "      <td>0.703298</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2500</th>\n",
       "      <td>0.6792</td>\n",
       "      <td>0.683262</td>\n",
       "      <td>0.702315</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2500</th>\n",
       "      <td>0.6792</td>\n",
       "      <td>0.680431</td>\n",
       "      <td>0.701486</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2500</th>\n",
       "      <td>0.6829</td>\n",
       "      <td>0.691391</td>\n",
       "      <td>0.699121</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2500</th>\n",
       "      <td>0.6754</td>\n",
       "      <td>0.680024</td>\n",
       "      <td>0.692833</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2500</th>\n",
       "      <td>0.6746</td>\n",
       "      <td>0.685422</td>\n",
       "      <td>0.692313</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2500</th>\n",
       "      <td>0.6710</td>\n",
       "      <td>0.673124</td>\n",
       "      <td>0.688084</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2500</th>\n",
       "      <td>0.6686</td>\n",
       "      <td>0.674506</td>\n",
       "      <td>0.685616</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2500</th>\n",
       "      <td>0.6678</td>\n",
       "      <td>0.676039</td>\n",
       "      <td>0.682113</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2500</th>\n",
       "      <td>0.6631</td>\n",
       "      <td>0.675652</td>\n",
       "      <td>0.678810</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2500</th>\n",
       "      <td>0.6573</td>\n",
       "      <td>0.669733</td>\n",
       "      <td>0.672506</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2500</th>\n",
       "      <td>0.6478</td>\n",
       "      <td>0.655066</td>\n",
       "      <td>0.661374</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2500</th>\n",
       "      <td>0.6423</td>\n",
       "      <td>0.655489</td>\n",
       "      <td>0.657447</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                              accuracy  balanced_accuracy   roc_auc  \\\n",
       "TfidfVectorizer 2500 ngram 2    0.7325           0.732387  0.772568   \n",
       "TfidfVectorizer 2500 ngram 2    0.7298           0.729708  0.771818   \n",
       "TfidfVectorizer 2500 ngram 2    0.7319           0.731856  0.768116   \n",
       "TfidfVectorizer 2500 ngram 2    0.7291           0.729122  0.766167   \n",
       "CountVectorizer 2500 ngram 2    0.7247           0.725874  0.755659   \n",
       "CountVectorizer 2500 ngram 2    0.7247           0.725352  0.754344   \n",
       "CountVectorizer 2500 ngram 2    0.7241           0.725575  0.751736   \n",
       "CountVectorizer 2500 ngram 2    0.7237           0.724857  0.751676   \n",
       "TfidfVectorizer 2500 ngram 2    0.7030           0.703174  0.733989   \n",
       "TfidfVectorizer 2500 ngram 2    0.7058           0.706936  0.729322   \n",
       "CountVectorizer 2500 ngram 2    0.7012           0.702338  0.727278   \n",
       "TfidfVectorizer 2500 ngram 2    0.7010           0.702112  0.721867   \n",
       "TfidfVectorizer 2500 ngram 2    0.6990           0.700200  0.721122   \n",
       "TfidfVectorizer 2500            0.6953           0.699267  0.715957   \n",
       "CountVectorizer 2500 ngram 2    0.6942           0.698566  0.714059   \n",
       "TfidfVectorizer 2500            0.6950           0.698409  0.713616   \n",
       "CountVectorizer 2500 ngram 2    0.6895           0.693454  0.710520   \n",
       "TfidfVectorizer 2500            0.6930           0.697896  0.709352   \n",
       "TfidfVectorizer 2500            0.6885           0.691590  0.704389   \n",
       "CountVectorizer 2500 ngram 2    0.6834           0.688448  0.703298   \n",
       "TfidfVectorizer 2500            0.6792           0.683262  0.702315   \n",
       "TfidfVectorizer 2500            0.6792           0.680431  0.701486   \n",
       "CountVectorizer 2500            0.6829           0.691391  0.699121   \n",
       "TfidfVectorizer 2500            0.6754           0.680024  0.692833   \n",
       "CountVectorizer 2500            0.6746           0.685422  0.692313   \n",
       "TfidfVectorizer 2500            0.6710           0.673124  0.688084   \n",
       "CountVectorizer 2500            0.6686           0.674506  0.685616   \n",
       "CountVectorizer 2500            0.6678           0.676039  0.682113   \n",
       "CountVectorizer 2500            0.6631           0.675652  0.678810   \n",
       "CountVectorizer 2500            0.6573           0.669733  0.672506   \n",
       "CountVectorizer 2500            0.6478           0.655066  0.661374   \n",
       "CountVectorizer 2500            0.6423           0.655489  0.657447   \n",
       "\n",
       "                              stop_words  lemmatization  negation  \n",
       "TfidfVectorizer 2500 ngram 2       False           True     False  \n",
       "TfidfVectorizer 2500 ngram 2       False          False     False  \n",
       "TfidfVectorizer 2500 ngram 2       False          False      True  \n",
       "TfidfVectorizer 2500 ngram 2       False           True      True  \n",
       "CountVectorizer 2500 ngram 2       False          False      True  \n",
       "CountVectorizer 2500 ngram 2       False          False     False  \n",
       "CountVectorizer 2500 ngram 2       False           True      True  \n",
       "CountVectorizer 2500 ngram 2       False           True     False  \n",
       "TfidfVectorizer 2500 ngram 2        True          False      True  \n",
       "TfidfVectorizer 2500 ngram 2        True          False     False  \n",
       "CountVectorizer 2500 ngram 2        True          False      True  \n",
       "TfidfVectorizer 2500 ngram 2        True           True     False  \n",
       "TfidfVectorizer 2500 ngram 2        True           True      True  \n",
       "TfidfVectorizer 2500               False          False      True  \n",
       "CountVectorizer 2500 ngram 2        True          False     False  \n",
       "TfidfVectorizer 2500                True          False      True  \n",
       "CountVectorizer 2500 ngram 2        True           True      True  \n",
       "TfidfVectorizer 2500               False           True      True  \n",
       "TfidfVectorizer 2500                True           True      True  \n",
       "CountVectorizer 2500 ngram 2        True           True     False  \n",
       "TfidfVectorizer 2500               False          False     False  \n",
       "TfidfVectorizer 2500                True          False     False  \n",
       "CountVectorizer 2500                True          False      True  \n",
       "TfidfVectorizer 2500               False           True     False  \n",
       "CountVectorizer 2500               False          False      True  \n",
       "TfidfVectorizer 2500                True           True     False  \n",
       "CountVectorizer 2500                True          False     False  \n",
       "CountVectorizer 2500                True           True      True  \n",
       "CountVectorizer 2500               False           True      True  \n",
       "CountVectorizer 2500               False          False     False  \n",
       "CountVectorizer 2500                True           True     False  \n",
       "CountVectorizer 2500               False           True     False  "
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_metrics.sort_values(by=\"roc_auc\", ascending=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Pour N=500, la combinaison optimale semble être stop_words=False, lemmatization=True, negation=False\n",
    "\n",
    "Cependant, N=500 est un nombre de features assez faible. Il est donc pertinent de regarder si de nouvelles tendances se dégagent lorsque l'on fait augmenter le nombre de features N.\n",
    "\n",
    "Pour N=1000, la combinaison optimale est de nouveau stop_words=False, lemmatization=True, negation=False\n",
    "\n",
    "Pour N=2500, la combinaison optimale est encore stop_words=False, lemmatization=True, negation=False\n",
    "\n",
    "Finalement, pour la suite, nous retiendrons la combinaison stop_words=False, lemmatization=True, negation=False\n",
    "\n",
    "Enfin, peu importe le nombre de features, un des vectorizers est systématiquement plus performant que les autres : il s'agit du TfidfVectorizer avec ngram. Pour la combinaison qui nous intéresse (stop_words=False, lemmatization=True, negation=False), ce vectorizer est toujours le plus performant, et les écarts avec les autres vectorizers semblent s'accroitre lorsque le nombre de features augmente."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Paramètres retenus : "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Wall time: 11.8 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "NB_DATA = 50000\n",
    "STOP_WORDS = False\n",
    "LEMMATIZATION = True\n",
    "NEGATION = False\n",
    "\n",
    "df = sentiment140(STOP_WORDS, LEMMATIZATION, NEGATION, NB_DATA)\n",
    "X = df.text.to_list()\n",
    "y = df.sentiment.to_list()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 1.2 - Comparaison par rapport au nombre de features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1440x432 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Wall time: 2min 21s\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>accuracy</th>\n",
       "      <th>balanced_accuracy</th>\n",
       "      <th>roc_auc</th>\n",
       "      <th>features</th>\n",
       "      <th>method</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.6737</td>\n",
       "      <td>0.682104</td>\n",
       "      <td>0.748858</td>\n",
       "      <td>199</td>\n",
       "      <td>CountVectorizer</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.6814</td>\n",
       "      <td>0.683652</td>\n",
       "      <td>0.753252</td>\n",
       "      <td>199</td>\n",
       "      <td>TfidfVectorizer</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.6782</td>\n",
       "      <td>0.680015</td>\n",
       "      <td>0.743068</td>\n",
       "      <td>199</td>\n",
       "      <td>CountVectorizer ngram 2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.6833</td>\n",
       "      <td>0.683756</td>\n",
       "      <td>0.750600</td>\n",
       "      <td>199</td>\n",
       "      <td>TfidfVectorizer ngram 2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.6861</td>\n",
       "      <td>0.693823</td>\n",
       "      <td>0.753880</td>\n",
       "      <td>380</td>\n",
       "      <td>CountVectorizer</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0.6973</td>\n",
       "      <td>0.698716</td>\n",
       "      <td>0.755570</td>\n",
       "      <td>380</td>\n",
       "      <td>TfidfVectorizer</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>0.6981</td>\n",
       "      <td>0.702867</td>\n",
       "      <td>0.763950</td>\n",
       "      <td>380</td>\n",
       "      <td>CountVectorizer ngram 2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>0.7054</td>\n",
       "      <td>0.705699</td>\n",
       "      <td>0.768651</td>\n",
       "      <td>380</td>\n",
       "      <td>TfidfVectorizer ngram 2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>0.6898</td>\n",
       "      <td>0.695185</td>\n",
       "      <td>0.742559</td>\n",
       "      <td>724</td>\n",
       "      <td>CountVectorizer</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>0.6985</td>\n",
       "      <td>0.699149</td>\n",
       "      <td>0.746575</td>\n",
       "      <td>724</td>\n",
       "      <td>TfidfVectorizer</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>0.7025</td>\n",
       "      <td>0.709703</td>\n",
       "      <td>0.748986</td>\n",
       "      <td>724</td>\n",
       "      <td>CountVectorizer ngram 2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>0.7155</td>\n",
       "      <td>0.717018</td>\n",
       "      <td>0.764239</td>\n",
       "      <td>724</td>\n",
       "      <td>TfidfVectorizer ngram 2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>0.6668</td>\n",
       "      <td>0.679152</td>\n",
       "      <td>0.694169</td>\n",
       "      <td>1380</td>\n",
       "      <td>CountVectorizer</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>0.6882</td>\n",
       "      <td>0.691515</td>\n",
       "      <td>0.719675</td>\n",
       "      <td>1380</td>\n",
       "      <td>TfidfVectorizer</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>0.7137</td>\n",
       "      <td>0.719013</td>\n",
       "      <td>0.748096</td>\n",
       "      <td>1380</td>\n",
       "      <td>CountVectorizer ngram 2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>0.7229</td>\n",
       "      <td>0.723949</td>\n",
       "      <td>0.764360</td>\n",
       "      <td>1380</td>\n",
       "      <td>TfidfVectorizer ngram 2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>0.6426</td>\n",
       "      <td>0.654575</td>\n",
       "      <td>0.657286</td>\n",
       "      <td>2630</td>\n",
       "      <td>CountVectorizer</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>0.6750</td>\n",
       "      <td>0.678714</td>\n",
       "      <td>0.694001</td>\n",
       "      <td>2630</td>\n",
       "      <td>TfidfVectorizer</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>0.7251</td>\n",
       "      <td>0.726288</td>\n",
       "      <td>0.752442</td>\n",
       "      <td>2630</td>\n",
       "      <td>CountVectorizer ngram 2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>0.7341</td>\n",
       "      <td>0.733987</td>\n",
       "      <td>0.770720</td>\n",
       "      <td>2630</td>\n",
       "      <td>TfidfVectorizer ngram 2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>0.6363</td>\n",
       "      <td>0.644762</td>\n",
       "      <td>0.647135</td>\n",
       "      <td>5011</td>\n",
       "      <td>CountVectorizer</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>0.6671</td>\n",
       "      <td>0.667734</td>\n",
       "      <td>0.679939</td>\n",
       "      <td>5011</td>\n",
       "      <td>TfidfVectorizer</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>0.7065</td>\n",
       "      <td>0.710176</td>\n",
       "      <td>0.724754</td>\n",
       "      <td>5011</td>\n",
       "      <td>CountVectorizer ngram 2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>0.7192</td>\n",
       "      <td>0.719947</td>\n",
       "      <td>0.740764</td>\n",
       "      <td>5011</td>\n",
       "      <td>TfidfVectorizer ngram 2</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    accuracy  balanced_accuracy   roc_auc  features                   method\n",
       "0     0.6737           0.682104  0.748858       199          CountVectorizer\n",
       "1     0.6814           0.683652  0.753252       199          TfidfVectorizer\n",
       "2     0.6782           0.680015  0.743068       199  CountVectorizer ngram 2\n",
       "3     0.6833           0.683756  0.750600       199  TfidfVectorizer ngram 2\n",
       "4     0.6861           0.693823  0.753880       380          CountVectorizer\n",
       "5     0.6973           0.698716  0.755570       380          TfidfVectorizer\n",
       "6     0.6981           0.702867  0.763950       380  CountVectorizer ngram 2\n",
       "7     0.7054           0.705699  0.768651       380  TfidfVectorizer ngram 2\n",
       "8     0.6898           0.695185  0.742559       724          CountVectorizer\n",
       "9     0.6985           0.699149  0.746575       724          TfidfVectorizer\n",
       "10    0.7025           0.709703  0.748986       724  CountVectorizer ngram 2\n",
       "11    0.7155           0.717018  0.764239       724  TfidfVectorizer ngram 2\n",
       "12    0.6668           0.679152  0.694169      1380          CountVectorizer\n",
       "13    0.6882           0.691515  0.719675      1380          TfidfVectorizer\n",
       "14    0.7137           0.719013  0.748096      1380  CountVectorizer ngram 2\n",
       "15    0.7229           0.723949  0.764360      1380  TfidfVectorizer ngram 2\n",
       "16    0.6426           0.654575  0.657286      2630          CountVectorizer\n",
       "17    0.6750           0.678714  0.694001      2630          TfidfVectorizer\n",
       "18    0.7251           0.726288  0.752442      2630  CountVectorizer ngram 2\n",
       "19    0.7341           0.733987  0.770720      2630  TfidfVectorizer ngram 2\n",
       "20    0.6363           0.644762  0.647135      5011          CountVectorizer\n",
       "21    0.6671           0.667734  0.679939      5011          TfidfVectorizer\n",
       "22    0.7065           0.710176  0.724754      5011  CountVectorizer ngram 2\n",
       "23    0.7192           0.719947  0.740764      5011  TfidfVectorizer ngram 2"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "%%time\n",
    "max_features_list = np.logspace(2.3, 3.7, 6, endpoint=True, dtype=int)\n",
    "influence_features(X, y, max_features_list, GaussianNB())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Comme annoncé précédemment, le vectorizer TfidfVectorizer avec ngram est le plus performant. C'est donc ce dernier que nous utiliserons pour l'estimation finale."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 2 - Estimation finale du modèle"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "On utilise les paramètres choisis au vu des résultats des études précédentes."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "D'après les courbes précédemment obtenues, le nombre optimal de features vaut environ 2740. Une recherche dichotomique manuelle permet de trouver rapidement que le nombre optimal de features vaut 2725 dans notre cas."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy : 0.7341\n",
      "Balanced accuracy : 0.7340\n",
      "ROC AUC : 0.7707\n",
      "Wall time: 9.26 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "m = Modelisation(X, y, TfidfVectorizer(max_features=2725, ngram_range=(1,2)), GaussianNB())\n",
    "m.show_conf_matrix()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "La valeur 0 correspond à un tweet véhiculant un sentiment négatif, tandis que la valeur 4 correspond à un tweet véhiculant un sentiment positif.\n",
    "\n",
    "Le modèle entrainé sur 40000 tweets, et testé sur 10000 tweets affiche des résultats plutôt corrects, avec une précision (rapport du nombre de prédictions correctes sur le nombre de prédictions total) d'environ 73,4%."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy : 0.7224\n",
      "Balanced accuracy : 0.7231\n",
      "ROC AUC : 0.7505\n",
      "Wall time: 8.5 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "m = Modelisation(X, y, CountVectorizer(max_features=2725, ngram_range=(1,2)), GaussianNB())\n",
    "m.show_conf_matrix()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Comme annoncé, la méthode de feature extraction CountVectorizer donne de moins bons résultats."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Dans cette étude, nous avons considéré un échantillon de 50000 tweets de la base Sentiment140 préprocessée. Ce choix vient du fait que la base de tweets webscrapés que nous souhaitons analyser fait approximativement cette taille."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Annexe : amélioration des résultats avec un échantillon plus gros"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "NB_DATA = 100000"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "stop_words: False, lemmatization: False, negation: False\n",
      "stop_words: False, lemmatization: False, negation: True\n",
      "stop_words: False, lemmatization: True, negation: False\n",
      "stop_words: False, lemmatization: True, negation: True\n",
      "stop_words: True, lemmatization: False, negation: False\n",
      "stop_words: True, lemmatization: False, negation: True\n",
      "stop_words: True, lemmatization: True, negation: False\n",
      "stop_words: True, lemmatization: True, negation: True\n",
      "Wall time: 8min 7s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "df_metrics = influence_preprocessing(2000, GaussianNB())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>accuracy</th>\n",
       "      <th>balanced_accuracy</th>\n",
       "      <th>roc_auc</th>\n",
       "      <th>stop_words</th>\n",
       "      <th>lemmatization</th>\n",
       "      <th>negation</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2000 ngram 2</th>\n",
       "      <td>0.73805</td>\n",
       "      <td>0.738581</td>\n",
       "      <td>0.773829</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2000 ngram 2</th>\n",
       "      <td>0.73720</td>\n",
       "      <td>0.737657</td>\n",
       "      <td>0.773383</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2000 ngram 2</th>\n",
       "      <td>0.73635</td>\n",
       "      <td>0.736751</td>\n",
       "      <td>0.773048</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2000 ngram 2</th>\n",
       "      <td>0.73610</td>\n",
       "      <td>0.736278</td>\n",
       "      <td>0.772363</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2000 ngram 2</th>\n",
       "      <td>0.72170</td>\n",
       "      <td>0.723348</td>\n",
       "      <td>0.767132</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2000 ngram 2</th>\n",
       "      <td>0.72205</td>\n",
       "      <td>0.722420</td>\n",
       "      <td>0.765311</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2000 ngram 2</th>\n",
       "      <td>0.71875</td>\n",
       "      <td>0.720667</td>\n",
       "      <td>0.765309</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2000 ngram 2</th>\n",
       "      <td>0.72715</td>\n",
       "      <td>0.730021</td>\n",
       "      <td>0.758326</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2000 ngram 2</th>\n",
       "      <td>0.71995</td>\n",
       "      <td>0.719984</td>\n",
       "      <td>0.757416</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2000 ngram 2</th>\n",
       "      <td>0.71600</td>\n",
       "      <td>0.716485</td>\n",
       "      <td>0.757198</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2000 ngram 2</th>\n",
       "      <td>0.72645</td>\n",
       "      <td>0.729470</td>\n",
       "      <td>0.756285</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2000 ngram 2</th>\n",
       "      <td>0.72090</td>\n",
       "      <td>0.720905</td>\n",
       "      <td>0.755996</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2000 ngram 2</th>\n",
       "      <td>0.72315</td>\n",
       "      <td>0.725928</td>\n",
       "      <td>0.753964</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2000 ngram 2</th>\n",
       "      <td>0.72230</td>\n",
       "      <td>0.725862</td>\n",
       "      <td>0.753962</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2000 ngram 2</th>\n",
       "      <td>0.71920</td>\n",
       "      <td>0.719388</td>\n",
       "      <td>0.751633</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2000</th>\n",
       "      <td>0.71895</td>\n",
       "      <td>0.721594</td>\n",
       "      <td>0.746729</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2000 ngram 2</th>\n",
       "      <td>0.71125</td>\n",
       "      <td>0.712134</td>\n",
       "      <td>0.742069</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2000</th>\n",
       "      <td>0.71440</td>\n",
       "      <td>0.716604</td>\n",
       "      <td>0.739914</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2000</th>\n",
       "      <td>0.71505</td>\n",
       "      <td>0.717840</td>\n",
       "      <td>0.739411</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2000</th>\n",
       "      <td>0.71105</td>\n",
       "      <td>0.712249</td>\n",
       "      <td>0.737528</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2000</th>\n",
       "      <td>0.70650</td>\n",
       "      <td>0.707856</td>\n",
       "      <td>0.734732</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2000</th>\n",
       "      <td>0.70115</td>\n",
       "      <td>0.702397</td>\n",
       "      <td>0.730455</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2000</th>\n",
       "      <td>0.70220</td>\n",
       "      <td>0.709383</td>\n",
       "      <td>0.725096</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2000</th>\n",
       "      <td>0.69825</td>\n",
       "      <td>0.699306</td>\n",
       "      <td>0.724933</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2000</th>\n",
       "      <td>0.69610</td>\n",
       "      <td>0.696459</td>\n",
       "      <td>0.724100</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2000</th>\n",
       "      <td>0.69695</td>\n",
       "      <td>0.704563</td>\n",
       "      <td>0.721051</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2000</th>\n",
       "      <td>0.69480</td>\n",
       "      <td>0.699482</td>\n",
       "      <td>0.720257</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2000</th>\n",
       "      <td>0.69675</td>\n",
       "      <td>0.703664</td>\n",
       "      <td>0.716269</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2000</th>\n",
       "      <td>0.68495</td>\n",
       "      <td>0.690938</td>\n",
       "      <td>0.711606</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2000</th>\n",
       "      <td>0.68240</td>\n",
       "      <td>0.688267</td>\n",
       "      <td>0.706842</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2000</th>\n",
       "      <td>0.67870</td>\n",
       "      <td>0.682932</td>\n",
       "      <td>0.706235</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2000</th>\n",
       "      <td>0.67485</td>\n",
       "      <td>0.681997</td>\n",
       "      <td>0.698296</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                              accuracy  balanced_accuracy   roc_auc  \\\n",
       "TfidfVectorizer 2000 ngram 2   0.73805           0.738581  0.773829   \n",
       "TfidfVectorizer 2000 ngram 2   0.73720           0.737657  0.773383   \n",
       "TfidfVectorizer 2000 ngram 2   0.73635           0.736751  0.773048   \n",
       "TfidfVectorizer 2000 ngram 2   0.73610           0.736278  0.772363   \n",
       "TfidfVectorizer 2000 ngram 2   0.72170           0.723348  0.767132   \n",
       "CountVectorizer 2000 ngram 2   0.72205           0.722420  0.765311   \n",
       "TfidfVectorizer 2000 ngram 2   0.71875           0.720667  0.765309   \n",
       "CountVectorizer 2000 ngram 2   0.72715           0.730021  0.758326   \n",
       "TfidfVectorizer 2000 ngram 2   0.71995           0.719984  0.757416   \n",
       "TfidfVectorizer 2000 ngram 2   0.71600           0.716485  0.757198   \n",
       "CountVectorizer 2000 ngram 2   0.72645           0.729470  0.756285   \n",
       "CountVectorizer 2000 ngram 2   0.72090           0.720905  0.755996   \n",
       "CountVectorizer 2000 ngram 2   0.72315           0.725928  0.753964   \n",
       "CountVectorizer 2000 ngram 2   0.72230           0.725862  0.753962   \n",
       "CountVectorizer 2000 ngram 2   0.71920           0.719388  0.751633   \n",
       "TfidfVectorizer 2000           0.71895           0.721594  0.746729   \n",
       "CountVectorizer 2000 ngram 2   0.71125           0.712134  0.742069   \n",
       "TfidfVectorizer 2000           0.71440           0.716604  0.739914   \n",
       "TfidfVectorizer 2000           0.71505           0.717840  0.739411   \n",
       "TfidfVectorizer 2000           0.71105           0.712249  0.737528   \n",
       "TfidfVectorizer 2000           0.70650           0.707856  0.734732   \n",
       "TfidfVectorizer 2000           0.70115           0.702397  0.730455   \n",
       "CountVectorizer 2000           0.70220           0.709383  0.725096   \n",
       "TfidfVectorizer 2000           0.69825           0.699306  0.724933   \n",
       "TfidfVectorizer 2000           0.69610           0.696459  0.724100   \n",
       "CountVectorizer 2000           0.69695           0.704563  0.721051   \n",
       "CountVectorizer 2000           0.69480           0.699482  0.720257   \n",
       "CountVectorizer 2000           0.69675           0.703664  0.716269   \n",
       "CountVectorizer 2000           0.68495           0.690938  0.711606   \n",
       "CountVectorizer 2000           0.68240           0.688267  0.706842   \n",
       "CountVectorizer 2000           0.67870           0.682932  0.706235   \n",
       "CountVectorizer 2000           0.67485           0.681997  0.698296   \n",
       "\n",
       "                              stop_words  lemmatization  negation  \n",
       "TfidfVectorizer 2000 ngram 2       False          False      True  \n",
       "TfidfVectorizer 2000 ngram 2       False          False     False  \n",
       "TfidfVectorizer 2000 ngram 2       False           True      True  \n",
       "TfidfVectorizer 2000 ngram 2       False           True     False  \n",
       "TfidfVectorizer 2000 ngram 2        True          False     False  \n",
       "CountVectorizer 2000 ngram 2        True          False      True  \n",
       "TfidfVectorizer 2000 ngram 2        True          False      True  \n",
       "CountVectorizer 2000 ngram 2       False          False      True  \n",
       "TfidfVectorizer 2000 ngram 2        True           True     False  \n",
       "TfidfVectorizer 2000 ngram 2        True           True      True  \n",
       "CountVectorizer 2000 ngram 2       False           True      True  \n",
       "CountVectorizer 2000 ngram 2        True          False     False  \n",
       "CountVectorizer 2000 ngram 2       False           True     False  \n",
       "CountVectorizer 2000 ngram 2       False          False     False  \n",
       "CountVectorizer 2000 ngram 2        True           True      True  \n",
       "TfidfVectorizer 2000               False          False      True  \n",
       "CountVectorizer 2000 ngram 2        True           True     False  \n",
       "TfidfVectorizer 2000               False           True      True  \n",
       "TfidfVectorizer 2000                True          False      True  \n",
       "TfidfVectorizer 2000                True           True      True  \n",
       "TfidfVectorizer 2000               False          False     False  \n",
       "TfidfVectorizer 2000                True          False     False  \n",
       "CountVectorizer 2000               False          False      True  \n",
       "TfidfVectorizer 2000               False           True     False  \n",
       "TfidfVectorizer 2000                True           True     False  \n",
       "CountVectorizer 2000                True          False      True  \n",
       "CountVectorizer 2000                True           True      True  \n",
       "CountVectorizer 2000               False           True      True  \n",
       "CountVectorizer 2000               False          False     False  \n",
       "CountVectorizer 2000                True          False     False  \n",
       "CountVectorizer 2000                True           True     False  \n",
       "CountVectorizer 2000               False           True     False  "
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_metrics.sort_values(by=\"roc_auc\", ascending=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>accuracy</th>\n",
       "      <th>balanced_accuracy</th>\n",
       "      <th>roc_auc</th>\n",
       "      <th>stop_words</th>\n",
       "      <th>lemmatization</th>\n",
       "      <th>negation</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2000 ngram 2</th>\n",
       "      <td>0.73805</td>\n",
       "      <td>0.738581</td>\n",
       "      <td>0.773829</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2000 ngram 2</th>\n",
       "      <td>0.73720</td>\n",
       "      <td>0.737657</td>\n",
       "      <td>0.773383</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2000 ngram 2</th>\n",
       "      <td>0.73635</td>\n",
       "      <td>0.736751</td>\n",
       "      <td>0.773048</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2000 ngram 2</th>\n",
       "      <td>0.73610</td>\n",
       "      <td>0.736278</td>\n",
       "      <td>0.772363</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2000 ngram 2</th>\n",
       "      <td>0.72715</td>\n",
       "      <td>0.730021</td>\n",
       "      <td>0.758326</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2000 ngram 2</th>\n",
       "      <td>0.72645</td>\n",
       "      <td>0.729470</td>\n",
       "      <td>0.756285</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2000 ngram 2</th>\n",
       "      <td>0.72315</td>\n",
       "      <td>0.725928</td>\n",
       "      <td>0.753964</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2000 ngram 2</th>\n",
       "      <td>0.72230</td>\n",
       "      <td>0.725862</td>\n",
       "      <td>0.753962</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2000 ngram 2</th>\n",
       "      <td>0.72170</td>\n",
       "      <td>0.723348</td>\n",
       "      <td>0.767132</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2000 ngram 2</th>\n",
       "      <td>0.72205</td>\n",
       "      <td>0.722420</td>\n",
       "      <td>0.765311</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2000</th>\n",
       "      <td>0.71895</td>\n",
       "      <td>0.721594</td>\n",
       "      <td>0.746729</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2000 ngram 2</th>\n",
       "      <td>0.72090</td>\n",
       "      <td>0.720905</td>\n",
       "      <td>0.755996</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2000 ngram 2</th>\n",
       "      <td>0.71875</td>\n",
       "      <td>0.720667</td>\n",
       "      <td>0.765309</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2000 ngram 2</th>\n",
       "      <td>0.71995</td>\n",
       "      <td>0.719984</td>\n",
       "      <td>0.757416</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2000 ngram 2</th>\n",
       "      <td>0.71920</td>\n",
       "      <td>0.719388</td>\n",
       "      <td>0.751633</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2000</th>\n",
       "      <td>0.71505</td>\n",
       "      <td>0.717840</td>\n",
       "      <td>0.739411</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2000</th>\n",
       "      <td>0.71440</td>\n",
       "      <td>0.716604</td>\n",
       "      <td>0.739914</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2000 ngram 2</th>\n",
       "      <td>0.71600</td>\n",
       "      <td>0.716485</td>\n",
       "      <td>0.757198</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2000</th>\n",
       "      <td>0.71105</td>\n",
       "      <td>0.712249</td>\n",
       "      <td>0.737528</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2000 ngram 2</th>\n",
       "      <td>0.71125</td>\n",
       "      <td>0.712134</td>\n",
       "      <td>0.742069</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2000</th>\n",
       "      <td>0.70220</td>\n",
       "      <td>0.709383</td>\n",
       "      <td>0.725096</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2000</th>\n",
       "      <td>0.70650</td>\n",
       "      <td>0.707856</td>\n",
       "      <td>0.734732</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2000</th>\n",
       "      <td>0.69695</td>\n",
       "      <td>0.704563</td>\n",
       "      <td>0.721051</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2000</th>\n",
       "      <td>0.69675</td>\n",
       "      <td>0.703664</td>\n",
       "      <td>0.716269</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2000</th>\n",
       "      <td>0.70115</td>\n",
       "      <td>0.702397</td>\n",
       "      <td>0.730455</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2000</th>\n",
       "      <td>0.69480</td>\n",
       "      <td>0.699482</td>\n",
       "      <td>0.720257</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2000</th>\n",
       "      <td>0.69825</td>\n",
       "      <td>0.699306</td>\n",
       "      <td>0.724933</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TfidfVectorizer 2000</th>\n",
       "      <td>0.69610</td>\n",
       "      <td>0.696459</td>\n",
       "      <td>0.724100</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2000</th>\n",
       "      <td>0.68495</td>\n",
       "      <td>0.690938</td>\n",
       "      <td>0.711606</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2000</th>\n",
       "      <td>0.68240</td>\n",
       "      <td>0.688267</td>\n",
       "      <td>0.706842</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2000</th>\n",
       "      <td>0.67870</td>\n",
       "      <td>0.682932</td>\n",
       "      <td>0.706235</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CountVectorizer 2000</th>\n",
       "      <td>0.67485</td>\n",
       "      <td>0.681997</td>\n",
       "      <td>0.698296</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                              accuracy  balanced_accuracy   roc_auc  \\\n",
       "TfidfVectorizer 2000 ngram 2   0.73805           0.738581  0.773829   \n",
       "TfidfVectorizer 2000 ngram 2   0.73720           0.737657  0.773383   \n",
       "TfidfVectorizer 2000 ngram 2   0.73635           0.736751  0.773048   \n",
       "TfidfVectorizer 2000 ngram 2   0.73610           0.736278  0.772363   \n",
       "CountVectorizer 2000 ngram 2   0.72715           0.730021  0.758326   \n",
       "CountVectorizer 2000 ngram 2   0.72645           0.729470  0.756285   \n",
       "CountVectorizer 2000 ngram 2   0.72315           0.725928  0.753964   \n",
       "CountVectorizer 2000 ngram 2   0.72230           0.725862  0.753962   \n",
       "TfidfVectorizer 2000 ngram 2   0.72170           0.723348  0.767132   \n",
       "CountVectorizer 2000 ngram 2   0.72205           0.722420  0.765311   \n",
       "TfidfVectorizer 2000           0.71895           0.721594  0.746729   \n",
       "CountVectorizer 2000 ngram 2   0.72090           0.720905  0.755996   \n",
       "TfidfVectorizer 2000 ngram 2   0.71875           0.720667  0.765309   \n",
       "TfidfVectorizer 2000 ngram 2   0.71995           0.719984  0.757416   \n",
       "CountVectorizer 2000 ngram 2   0.71920           0.719388  0.751633   \n",
       "TfidfVectorizer 2000           0.71505           0.717840  0.739411   \n",
       "TfidfVectorizer 2000           0.71440           0.716604  0.739914   \n",
       "TfidfVectorizer 2000 ngram 2   0.71600           0.716485  0.757198   \n",
       "TfidfVectorizer 2000           0.71105           0.712249  0.737528   \n",
       "CountVectorizer 2000 ngram 2   0.71125           0.712134  0.742069   \n",
       "CountVectorizer 2000           0.70220           0.709383  0.725096   \n",
       "TfidfVectorizer 2000           0.70650           0.707856  0.734732   \n",
       "CountVectorizer 2000           0.69695           0.704563  0.721051   \n",
       "CountVectorizer 2000           0.69675           0.703664  0.716269   \n",
       "TfidfVectorizer 2000           0.70115           0.702397  0.730455   \n",
       "CountVectorizer 2000           0.69480           0.699482  0.720257   \n",
       "TfidfVectorizer 2000           0.69825           0.699306  0.724933   \n",
       "TfidfVectorizer 2000           0.69610           0.696459  0.724100   \n",
       "CountVectorizer 2000           0.68495           0.690938  0.711606   \n",
       "CountVectorizer 2000           0.68240           0.688267  0.706842   \n",
       "CountVectorizer 2000           0.67870           0.682932  0.706235   \n",
       "CountVectorizer 2000           0.67485           0.681997  0.698296   \n",
       "\n",
       "                              stop_words  lemmatization  negation  \n",
       "TfidfVectorizer 2000 ngram 2       False          False      True  \n",
       "TfidfVectorizer 2000 ngram 2       False          False     False  \n",
       "TfidfVectorizer 2000 ngram 2       False           True      True  \n",
       "TfidfVectorizer 2000 ngram 2       False           True     False  \n",
       "CountVectorizer 2000 ngram 2       False          False      True  \n",
       "CountVectorizer 2000 ngram 2       False           True      True  \n",
       "CountVectorizer 2000 ngram 2       False           True     False  \n",
       "CountVectorizer 2000 ngram 2       False          False     False  \n",
       "TfidfVectorizer 2000 ngram 2        True          False     False  \n",
       "CountVectorizer 2000 ngram 2        True          False      True  \n",
       "TfidfVectorizer 2000               False          False      True  \n",
       "CountVectorizer 2000 ngram 2        True          False     False  \n",
       "TfidfVectorizer 2000 ngram 2        True          False      True  \n",
       "TfidfVectorizer 2000 ngram 2        True           True     False  \n",
       "CountVectorizer 2000 ngram 2        True           True      True  \n",
       "TfidfVectorizer 2000                True          False      True  \n",
       "TfidfVectorizer 2000               False           True      True  \n",
       "TfidfVectorizer 2000 ngram 2        True           True      True  \n",
       "TfidfVectorizer 2000                True           True      True  \n",
       "CountVectorizer 2000 ngram 2        True           True     False  \n",
       "CountVectorizer 2000               False          False      True  \n",
       "TfidfVectorizer 2000               False          False     False  \n",
       "CountVectorizer 2000                True          False      True  \n",
       "CountVectorizer 2000               False           True      True  \n",
       "TfidfVectorizer 2000                True          False     False  \n",
       "CountVectorizer 2000                True           True      True  \n",
       "TfidfVectorizer 2000               False           True     False  \n",
       "TfidfVectorizer 2000                True           True     False  \n",
       "CountVectorizer 2000               False          False     False  \n",
       "CountVectorizer 2000                True          False     False  \n",
       "CountVectorizer 2000                True           True     False  \n",
       "CountVectorizer 2000               False           True     False  "
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_metrics.sort_values(by=\"balanced_accuracy\", ascending=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Cette fois, la combinaison stop_words=False, lemmatization=False, negation=True, associée au TfidfVectorizer avec ngram, semble donner les meilleurs résultats."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Paramètres retenus : "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Wall time: 16.1 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "NB_DATA = 100000\n",
    "STOP_WORDS = False\n",
    "LEMMATIZATION = False\n",
    "NEGATION = True\n",
    "\n",
    "df = sentiment140(STOP_WORDS, LEMMATIZATION, NEGATION, NB_DATA)\n",
    "X = df.text.to_list()\n",
    "y = df.sentiment.to_list()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1440x432 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Wall time: 6min 12s\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>accuracy</th>\n",
       "      <th>balanced_accuracy</th>\n",
       "      <th>roc_auc</th>\n",
       "      <th>features</th>\n",
       "      <th>method</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.66360</td>\n",
       "      <td>0.667945</td>\n",
       "      <td>0.736107</td>\n",
       "      <td>199</td>\n",
       "      <td>CountVectorizer</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.67455</td>\n",
       "      <td>0.674853</td>\n",
       "      <td>0.737347</td>\n",
       "      <td>199</td>\n",
       "      <td>TfidfVectorizer</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.65870</td>\n",
       "      <td>0.663736</td>\n",
       "      <td>0.732661</td>\n",
       "      <td>199</td>\n",
       "      <td>CountVectorizer ngram 2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.66865</td>\n",
       "      <td>0.669262</td>\n",
       "      <td>0.736251</td>\n",
       "      <td>199</td>\n",
       "      <td>TfidfVectorizer ngram 2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.68915</td>\n",
       "      <td>0.696169</td>\n",
       "      <td>0.758449</td>\n",
       "      <td>380</td>\n",
       "      <td>CountVectorizer</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0.69605</td>\n",
       "      <td>0.697111</td>\n",
       "      <td>0.757812</td>\n",
       "      <td>380</td>\n",
       "      <td>TfidfVectorizer</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>0.68820</td>\n",
       "      <td>0.692462</td>\n",
       "      <td>0.755577</td>\n",
       "      <td>380</td>\n",
       "      <td>CountVectorizer ngram 2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>0.69460</td>\n",
       "      <td>0.694890</td>\n",
       "      <td>0.755500</td>\n",
       "      <td>380</td>\n",
       "      <td>TfidfVectorizer ngram 2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>0.70335</td>\n",
       "      <td>0.711145</td>\n",
       "      <td>0.748982</td>\n",
       "      <td>724</td>\n",
       "      <td>CountVectorizer</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>0.71385</td>\n",
       "      <td>0.716911</td>\n",
       "      <td>0.758378</td>\n",
       "      <td>724</td>\n",
       "      <td>TfidfVectorizer</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>0.70680</td>\n",
       "      <td>0.713446</td>\n",
       "      <td>0.750281</td>\n",
       "      <td>724</td>\n",
       "      <td>CountVectorizer ngram 2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>0.71510</td>\n",
       "      <td>0.716713</td>\n",
       "      <td>0.764420</td>\n",
       "      <td>724</td>\n",
       "      <td>TfidfVectorizer ngram 2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>0.69995</td>\n",
       "      <td>0.708015</td>\n",
       "      <td>0.727490</td>\n",
       "      <td>1380</td>\n",
       "      <td>CountVectorizer</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>0.71660</td>\n",
       "      <td>0.720354</td>\n",
       "      <td>0.746105</td>\n",
       "      <td>1380</td>\n",
       "      <td>TfidfVectorizer</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>0.72060</td>\n",
       "      <td>0.726203</td>\n",
       "      <td>0.754464</td>\n",
       "      <td>1380</td>\n",
       "      <td>CountVectorizer ngram 2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>0.73285</td>\n",
       "      <td>0.734463</td>\n",
       "      <td>0.769734</td>\n",
       "      <td>1380</td>\n",
       "      <td>TfidfVectorizer ngram 2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>0.69995</td>\n",
       "      <td>0.706729</td>\n",
       "      <td>0.719361</td>\n",
       "      <td>2630</td>\n",
       "      <td>CountVectorizer</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>0.71925</td>\n",
       "      <td>0.721025</td>\n",
       "      <td>0.743271</td>\n",
       "      <td>2630</td>\n",
       "      <td>TfidfVectorizer</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>0.73115</td>\n",
       "      <td>0.731634</td>\n",
       "      <td>0.765217</td>\n",
       "      <td>2630</td>\n",
       "      <td>CountVectorizer ngram 2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>0.74120</td>\n",
       "      <td>0.741218</td>\n",
       "      <td>0.778165</td>\n",
       "      <td>2630</td>\n",
       "      <td>TfidfVectorizer ngram 2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>0.67080</td>\n",
       "      <td>0.681487</td>\n",
       "      <td>0.679073</td>\n",
       "      <td>5011</td>\n",
       "      <td>CountVectorizer</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>0.69440</td>\n",
       "      <td>0.698795</td>\n",
       "      <td>0.705585</td>\n",
       "      <td>5011</td>\n",
       "      <td>TfidfVectorizer</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>0.73165</td>\n",
       "      <td>0.735055</td>\n",
       "      <td>0.752605</td>\n",
       "      <td>5011</td>\n",
       "      <td>CountVectorizer ngram 2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>0.74295</td>\n",
       "      <td>0.743579</td>\n",
       "      <td>0.767737</td>\n",
       "      <td>5011</td>\n",
       "      <td>TfidfVectorizer ngram 2</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    accuracy  balanced_accuracy   roc_auc  features                   method\n",
       "0    0.66360           0.667945  0.736107       199          CountVectorizer\n",
       "1    0.67455           0.674853  0.737347       199          TfidfVectorizer\n",
       "2    0.65870           0.663736  0.732661       199  CountVectorizer ngram 2\n",
       "3    0.66865           0.669262  0.736251       199  TfidfVectorizer ngram 2\n",
       "4    0.68915           0.696169  0.758449       380          CountVectorizer\n",
       "5    0.69605           0.697111  0.757812       380          TfidfVectorizer\n",
       "6    0.68820           0.692462  0.755577       380  CountVectorizer ngram 2\n",
       "7    0.69460           0.694890  0.755500       380  TfidfVectorizer ngram 2\n",
       "8    0.70335           0.711145  0.748982       724          CountVectorizer\n",
       "9    0.71385           0.716911  0.758378       724          TfidfVectorizer\n",
       "10   0.70680           0.713446  0.750281       724  CountVectorizer ngram 2\n",
       "11   0.71510           0.716713  0.764420       724  TfidfVectorizer ngram 2\n",
       "12   0.69995           0.708015  0.727490      1380          CountVectorizer\n",
       "13   0.71660           0.720354  0.746105      1380          TfidfVectorizer\n",
       "14   0.72060           0.726203  0.754464      1380  CountVectorizer ngram 2\n",
       "15   0.73285           0.734463  0.769734      1380  TfidfVectorizer ngram 2\n",
       "16   0.69995           0.706729  0.719361      2630          CountVectorizer\n",
       "17   0.71925           0.721025  0.743271      2630          TfidfVectorizer\n",
       "18   0.73115           0.731634  0.765217      2630  CountVectorizer ngram 2\n",
       "19   0.74120           0.741218  0.778165      2630  TfidfVectorizer ngram 2\n",
       "20   0.67080           0.681487  0.679073      5011          CountVectorizer\n",
       "21   0.69440           0.698795  0.705585      5011          TfidfVectorizer\n",
       "22   0.73165           0.735055  0.752605      5011  CountVectorizer ngram 2\n",
       "23   0.74295           0.743579  0.767737      5011  TfidfVectorizer ngram 2"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "%%time\n",
    "max_features_list = np.logspace(2.3, 3.7, 6, endpoint=True, dtype=int)\n",
    "influence_features(X, y, max_features_list, GaussianNB())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Comme précédemment, le TfidfVectorizer avec ngram donne les meilleurs résultats. D'après les courbes ci-dessus, le nombre optimal de features qui ne dégrade pas l'aire sous la courbe ROC est environ 2750, nombre que l'on peut affiner à 2726."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy : 0.7419\n",
      "Balanced accuracy : 0.7420\n",
      "ROC AUC : 0.7771\n",
      "Wall time: 23.7 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "m = Modelisation(X, y, TfidfVectorizer(max_features=2726, ngram_range=(1,2)), GaussianNB())\n",
    "m.show_conf_matrix()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Finalement, le modèle ainsi entrainé affiche une précision d'environ 74,2%. Ces résultats sont donc légèrement meilleurs que ceux obtenus avec l'échantillon de 50000 tweets de la première étude. \n",
    "\n",
    "Toutefois, les résultats des deux études sont globalement assez similaires, ce qui laisse penser que les résultats de la première étude sont plutôt bons, malgré la faible taille de l'échantillon considéré."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Temps d'exécution total : 00:23:00\n"
     ]
    }
   ],
   "source": [
    "print(f\"Temps d'exécution total : {time.strftime('%H:%M:%S', time.gmtime(time.time()-t))}\")"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}