faketweets-ml-an.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "import sqlalchemy \n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "\n",
    "import credentials"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "DB_USER = credentials.username()\n",
    "DB_PASSWORD = credentials.password()\n",
    "DB_SERVER = \"mydbinstance.cmgiyjltohsu.us-east-1.rds.amazonaws.com\"\n",
    "DATABASEURI = \"postgresql://\"+DB_USER+\":\"+DB_PASSWORD+\"@\"+DB_SERVER+\"/FakeTasticdb\"\n",
    "\n",
    "# Connect sql engine\n",
    "engine = sqlalchemy.create_engine(DATABASEURI)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### List of features:\n",
    "\n",
    "Non text:\n",
    "\n",
    "* hashtag\n",
    "* retweet count\n",
    "* retweet (yes/no)\n",
    "* retweet_count\n",
    "* user_verified\n",
    "* user_followers_count\n",
    "* user_statuses_count\n",
    "* user_friends_count\n",
    "* user_favorites_count\n",
    "* tweet_relative_age\n",
    "\n",
    "Text:\n",
    "\n",
    "* length\n",
    "* % capitalized\n",
    "* links\n",
    "* bad words\n",
    "* punctuation (!,?)\n",
    "* hashtags"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {},
   "outputs": [],
   "source": [
    "#conn = engine.connect()\n",
    "query = \"\"\"SELECT tweet_id, tweet_hashtag, retweet, retweet_count, tweet_text,\n",
    "                user_verified, user_followers_count, \n",
    "                user_statuses_count, user_friends_count,\n",
    "                user_favourites_count, tweet_relative_age, is_fake\n",
    "           FROM  tweets_detailed limit 50000\n",
    "           \"\"\"\n",
    "#alltweets = conn.execute(text(query))\n",
    "#conn.close()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "metadata": {},
   "outputs": [],
   "source": [
    "tweets = pd.read_sql_query(query, engine)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "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>tweet_id</th>\n",
       "      <th>tweet_hashtag</th>\n",
       "      <th>retweet</th>\n",
       "      <th>retweet_count</th>\n",
       "      <th>tweet_text</th>\n",
       "      <th>user_verified</th>\n",
       "      <th>user_followers_count</th>\n",
       "      <th>user_statuses_count</th>\n",
       "      <th>user_friends_count</th>\n",
       "      <th>user_favourites_count</th>\n",
       "      <th>tweet_relative_age</th>\n",
       "      <th>is_fake</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>544347456238931968</td>\n",
       "      <td>sydneysiege</td>\n",
       "      <td>False</td>\n",
       "      <td>0</td>\n",
       "      <td>@Uber_Sydney @mtats Thank you!</td>\n",
       "      <td>False</td>\n",
       "      <td>1295</td>\n",
       "      <td>7015</td>\n",
       "      <td>2001</td>\n",
       "      <td>43</td>\n",
       "      <td>5</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>544302519191863296</td>\n",
       "      <td>sydneysiege</td>\n",
       "      <td>False</td>\n",
       "      <td>0</td>\n",
       "      <td>@braduok there's 1-2 men holding 13 people hos...</td>\n",
       "      <td>False</td>\n",
       "      <td>5348</td>\n",
       "      <td>50787</td>\n",
       "      <td>2026</td>\n",
       "      <td>9176</td>\n",
       "      <td>10</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>544319371292983297</td>\n",
       "      <td>sydneysiege</td>\n",
       "      <td>False</td>\n",
       "      <td>0</td>\n",
       "      <td>@calsdiscostick they said it connects to the I...</td>\n",
       "      <td>False</td>\n",
       "      <td>390</td>\n",
       "      <td>2112</td>\n",
       "      <td>942</td>\n",
       "      <td>1205</td>\n",
       "      <td>3</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>544276351952244737</td>\n",
       "      <td>sydneysiege</td>\n",
       "      <td>False</td>\n",
       "      <td>1</td>\n",
       "      <td>@Luke_rative @australian true that but without...</td>\n",
       "      <td>False</td>\n",
       "      <td>59</td>\n",
       "      <td>999</td>\n",
       "      <td>214</td>\n",
       "      <td>663</td>\n",
       "      <td>10</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>544319916212367360</td>\n",
       "      <td>sydneysiege</td>\n",
       "      <td>False</td>\n",
       "      <td>0</td>\n",
       "      <td>@DeAnnaSabaoth Hope it helps :) have a good da...</td>\n",
       "      <td>False</td>\n",
       "      <td>556</td>\n",
       "      <td>12421</td>\n",
       "      <td>145</td>\n",
       "      <td>480</td>\n",
       "      <td>11</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             tweet_id tweet_hashtag  retweet  retweet_count  \\\n",
       "0  544347456238931968   sydneysiege    False              0   \n",
       "1  544302519191863296   sydneysiege    False              0   \n",
       "2  544319371292983297   sydneysiege    False              0   \n",
       "3  544276351952244737   sydneysiege    False              1   \n",
       "4  544319916212367360   sydneysiege    False              0   \n",
       "\n",
       "                                          tweet_text  user_verified  \\\n",
       "0                     @Uber_Sydney @mtats Thank you!          False   \n",
       "1  @braduok there's 1-2 men holding 13 people hos...          False   \n",
       "2  @calsdiscostick they said it connects to the I...          False   \n",
       "3  @Luke_rative @australian true that but without...          False   \n",
       "4  @DeAnnaSabaoth Hope it helps :) have a good da...          False   \n",
       "\n",
       "   user_followers_count  user_statuses_count  user_friends_count  \\\n",
       "0                  1295                 7015                2001   \n",
       "1                  5348                50787                2026   \n",
       "2                   390                 2112                 942   \n",
       "3                    59                  999                 214   \n",
       "4                   556                12421                 145   \n",
       "\n",
       "   user_favourites_count  tweet_relative_age  is_fake  \n",
       "0                     43                   5     True  \n",
       "1                   9176                  10     True  \n",
       "2                   1205                   3     True  \n",
       "3                    663                  10     True  \n",
       "4                    480                  11     True  "
      ]
     },
     "execution_count": 88,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tweets.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.38714"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(tweets[tweets['is_fake']==True]) / len(tweets)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1800x576 with 10 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "target_name = 'is_fake'\n",
    "continuous_vars = ['user_followers_count', 'user_statuses_count', 'user_friends_count','user_favourites_count','tweet_relative_age']\n",
    "fig, ax = plt.subplots(2,5,figsize=(25,8))\n",
    "for i, var in enumerate(continuous_vars):\n",
    "    aux = pd.DataFrame({var: tweets[var],\n",
    "              target_name: tweets[target_name]})\n",
    "    ax[0, i].hist(tweets[var], bins=20, density=True)\n",
    "    ax[0, i].set_title(var)\n",
    "    meds = aux.groupby(target_name).median().sort_values(var, ascending=False).index\n",
    "    sns.boxplot(x=target_name, y=var, \n",
    "                data=aux, \n",
    "                order=meds,\n",
    "                width=0.5,\n",
    "                ax=ax[1,i],\n",
    "                fliersize=1,\n",
    "                color='orange')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Baseline Model without text"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.compose import make_column_transformer\n",
    "from sklearn.pipeline import make_pipeline, Pipeline, FeatureUnion\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.model_selection import GridSearchCV\n",
    "from sklearn.preprocessing import OneHotEncoder, StandardScaler, MinMaxScaler, Normalizer\n",
    "from sklearn.metrics import (roc_curve, precision_recall_curve,\n",
    "                             recall_score, make_scorer)\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from sklearn.feature_extraction.text import CountVectorizer\n",
    "from sklearn.feature_extraction.text import TfidfVectorizer, TfidfTransformer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Split data train / test\n",
    "X = tweets[[x for x in tweets.columns if x not in ['tweet_id', 'is_fake', 'tweet_text']]]\n",
    "y = tweets['is_fake']\n",
    "\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=0)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Logistic Regression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/preprocessing/data.py:323: DataConversionWarning: Data with input dtype bool, int64 were all converted to float64 by MinMaxScaler.\n",
      "  return self.partial_fit(X, y)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/linear_model/logistic.py:433: FutureWarning: Default solver will be changed to 'lbfgs' in 0.22. Specify a solver to silence this warning.\n",
      "  FutureWarning)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "{'logisticregression__C': 1000.0, 'logisticregression__penalty': 'l1'}"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "scoring = {'roc_auc': 'roc_auc', \n",
    "           'average_precision': 'average_precision',\n",
    "           'macro_average_recall': make_scorer(recall_score, average='binary')}\n",
    "\n",
    "grid_param = [{'logisticregression__penalty': ['l1', 'l2'],\n",
    "               'logisticregression__C': np.logspace(-3, 3, 13)}]\n",
    "pipe = make_pipeline(MinMaxScaler(), LogisticRegression())\n",
    "grid_logistic = GridSearchCV(pipe, grid_param, cv=5, verbose=0,\n",
    "                             scoring=scoring, refit='roc_auc',\n",
    "                             n_jobs=-1, return_train_score=True)\n",
    "grid_logistic.fit(X_train, y_train)\n",
    "results_logistic = grid_logistic.cv_results_\n",
    "grid_logistic.best_params_"
   ]
  },
  {
   "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>penalty</th>\n",
       "      <th>C</th>\n",
       "      <th>macro_average_recall</th>\n",
       "      <th>roc_auc</th>\n",
       "      <th>average_precision</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>l1</td>\n",
       "      <td>1000</td>\n",
       "      <td>0.058613</td>\n",
       "      <td>0.530729</td>\n",
       "      <td>0.426999</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>l1</td>\n",
       "      <td>316.228</td>\n",
       "      <td>0.058613</td>\n",
       "      <td>0.530722</td>\n",
       "      <td>0.426986</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>l2</td>\n",
       "      <td>1000</td>\n",
       "      <td>0.058613</td>\n",
       "      <td>0.530714</td>\n",
       "      <td>0.426944</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>l1</td>\n",
       "      <td>100</td>\n",
       "      <td>0.058613</td>\n",
       "      <td>0.530709</td>\n",
       "      <td>0.426961</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>l2</td>\n",
       "      <td>316.228</td>\n",
       "      <td>0.058613</td>\n",
       "      <td>0.530684</td>\n",
       "      <td>0.426929</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   penalty        C  macro_average_recall   roc_auc  average_precision\n",
       "24      l1     1000              0.058613  0.530729           0.426999\n",
       "22      l1  316.228              0.058613  0.530722           0.426986\n",
       "25      l2     1000              0.058613  0.530714           0.426944\n",
       "20      l1      100              0.058613  0.530709           0.426961\n",
       "23      l2  316.228              0.058613  0.530684           0.426929"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results_logistic_df = pd.DataFrame()\n",
    "results_logistic_df['penalty'] = results_logistic['param_logisticregression__penalty'].data\n",
    "results_logistic_df['C']  = results_logistic['param_logisticregression__C'].data\n",
    "results_logistic_df['macro_average_recall']  = results_logistic['mean_test_macro_average_recall']\n",
    "results_logistic_df['roc_auc'] = results_logistic['mean_test_roc_auc']\n",
    "results_logistic_df['average_precision'] = results_logistic['mean_test_average_precision']\n",
    "results_logistic_df.sort_values(by='roc_auc', ascending=False).head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.5250684970383145"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "grid_logistic.score(X_test, y_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Random Forest"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/ensemble/forest.py:246: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "{'randomforestclassifier__max_leaf_nodes': 49}"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "grid_param = {'randomforestclassifier__max_leaf_nodes':range(5, 50)}\n",
    "pipe = make_pipeline(RandomForestClassifier(warm_start=True, random_state=0))\n",
    "grid_rf = GridSearchCV(pipe, grid_param, cv=5, verbose=0,\n",
    "                             n_jobs=-1, return_train_score=True,\n",
    "                             scoring=scoring,\n",
    "                             refit='roc_auc'\n",
    "                            )\n",
    "grid_rf.fit(X_train, y_train)\n",
    "results_rf = grid_rf.cv_results_\n",
    "grid_rf.best_params_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.6675498632864831"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "grid_rf.score(X_test, y_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Using text data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Baseline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of train tweets 37500\n",
      "Number of test tweets 12500\n",
      "% of class 1 train 0.3871733333333333\n",
      "% of class 1 test 0.38704\n"
     ]
    }
   ],
   "source": [
    "# Split data train / test\n",
    "X = tweets['tweet_text']\n",
    "y = tweets['is_fake']\n",
    "\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=0, shuffle=True)\n",
    "print(\"Number of train tweets\", len(X_train))\n",
    "print(\"Number of test tweets\",len(X_test))\n",
    "\n",
    "# check balances\n",
    "print(\"% of class 1 train\", len(y_train[y_train == 1]) / len(y_train))\n",
    "print(\"% of class 1 test\", len(y_test[y_test == 1]) / len(y_test))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "GridSearchCV(cv=5, error_score='raise-deprecating',\n",
       "       estimator=Pipeline(memory=None,\n",
       "     steps=[('countvectorizer', CountVectorizer(analyzer='word', binary=False, decode_error='strict',\n",
       "        dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\n",
       "        lowercase=True, max_df=1.0, max_features=None, min_df=1,\n",
       "        ngram_range=(1, 1), preprocessor=None, stop_words=None,\n",
       "  ...ty='l2', random_state=None, solver='liblinear',\n",
       "          tol=0.0001, verbose=0, warm_start=False))]),\n",
       "       fit_params=None, iid='warn', n_jobs=None,\n",
       "       param_grid={'logisticregression__C': array([ 0.1,  1. , 10. ])},\n",
       "       pre_dispatch='2*n_jobs', refit=True, return_train_score='warn',\n",
       "       scoring='roc_auc', verbose=0)"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.random.seed(2019)\n",
    "param_grid_baseline = {\"logisticregression__C\": np.logspace(-1,1,3)}\n",
    "pipe_baseline = make_pipeline(CountVectorizer(),\n",
    "                              LogisticRegression(solver='liblinear'))\n",
    "grid_baseline = GridSearchCV(pipe_baseline,\n",
    "                             param_grid=param_grid_baseline,\n",
    "                             cv=5, scoring=\"roc_auc\")\n",
    "grid_baseline.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'logisticregression__C': 10.0}"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "grid_baseline.best_params_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9654684465626349"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.mean(grid_baseline.cv_results_['mean_test_score'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9868683211273668"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "grid_baseline.score(X_test, y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot_coefficients(coefs, features):\n",
    "    fig,ax = plt.subplots(figsize=(17,7))\n",
    "    \n",
    "    coefs = pd.DataFrame(coefs, columns=['coef']).reset_index()\n",
    "    coefs['abs_coef'] = abs(coefs['coef'])\n",
    "    coefs['sign'] = coefs['coef'].map(np.sign)\n",
    "    coefs['features'] = coefs['index'].apply(lambda x: features[x])\n",
    "    top_coefs = coefs.sort_values(by='abs_coef', ascending=False).head(50)\n",
    "    top_coefs = top_coefs.sort_values(by='coef', ascending=False)\n",
    "    top_coefs.sort_values(by='coef', ascending=False).plot(kind='bar', \n",
    "                y='coef', x='features',\n",
    "                ax=ax,\n",
    "                rot=40,\n",
    "                legend=False,\n",
    "                sort_columns=True,\n",
    "                title='Top 20 features for Logistic Regression',\n",
    "                color=np.where(top_coefs[\"coef\"]> 0, 'b', 'g'))\n",
    "    labels = [l.get_text() for l in ax.get_xticklabels()]\n",
    "    ax = ax.set_xticklabels(labels, ha='right')\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1224x504 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "voc_word_baseline = grid_baseline.best_estimator_. \\\n",
    "                        named_steps['countvectorizer'].vocabulary_\n",
    "voc_index_baseline = {y:x for x,y in voc_word_baseline.items()}\n",
    "# Coefficientes\n",
    "coefs_baseline = grid_baseline.best_estimator_.named_steps['logisticregression'].coef_[0]\n",
    "plot_coefficients(coefs_baseline, voc_index_baseline)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Normalizer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "GridSearchCV(cv=5, error_score='raise-deprecating',\n",
       "       estimator=Pipeline(memory=None,\n",
       "     steps=[('countvectorizer', CountVectorizer(analyzer='word', binary=False, decode_error='strict',\n",
       "        dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\n",
       "        lowercase=True, max_df=1.0, max_features=None, min_df=1,\n",
       "        ngram_range=(1, 1), preprocessor=None, stop_words=None,\n",
       "  ...ty='l2', random_state=None, solver='liblinear',\n",
       "          tol=0.0001, verbose=0, warm_start=False))]),\n",
       "       fit_params=None, iid='warn', n_jobs=None,\n",
       "       param_grid={'logisticregression__C': array([ 0.1,  1. , 10. ]), 'normalizer': [None, Normalizer(copy=True, norm='l2')]},\n",
       "       pre_dispatch='2*n_jobs', refit=True, return_train_score='warn',\n",
       "       scoring='roc_auc', verbose=0)"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.random.seed(2019)\n",
    "param_grid_normalizer = {\"logisticregression__C\": np.logspace(-1,1,3),\n",
    "                         \"normalizer\": [None, Normalizer()]}\n",
    "pipe_normalizer = make_pipeline(CountVectorizer(),\n",
    "                                Normalizer(),\n",
    "                                LogisticRegression(solver='liblinear'))\n",
    "grid_normalizer = GridSearchCV(pipe_normalizer,\n",
    "                               param_grid=param_grid_normalizer,\n",
    "                               cv=5, scoring=\"roc_auc\")\n",
    "grid_normalizer.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'logisticregression__C': 10.0, 'normalizer': None}"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "grid_normalizer.best_params_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9868683211273668"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "grid_normalizer.score(X_test, y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.941761712005135"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.mean(grid_normalizer.cv_results_['mean_test_score'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1224x504 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "voc_word_normalizer = grid_normalizer.best_estimator_. \\\n",
    "                named_steps['countvectorizer'].vocabulary_\n",
    "voc_index_normalizer = {y:x for x,y in voc_word_normalizer.items()}\n",
    "# Coefficientes\n",
    "coefs_normalizer = grid_normalizer.best_estimator_. \\\n",
    "                        named_steps['logisticregression'].coef_[0]\n",
    "plot_coefficients(coefs_normalizer, voc_index_normalizer)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Char"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "GridSearchCV(cv=5, error_score='raise-deprecating',\n",
       "       estimator=Pipeline(memory=None,\n",
       "     steps=[('countvectorizer', CountVectorizer(analyzer='char_wb', binary=False, decode_error='strict',\n",
       "        dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\n",
       "        lowercase=True, max_df=1.0, max_features=None, min_df=4,\n",
       "        ngram_range=(1, 1), preprocessor=None, stop_words=None,...ty='l2', random_state=None, solver='liblinear',\n",
       "          tol=0.0001, verbose=0, warm_start=False))]),\n",
       "       fit_params=None, iid='warn', n_jobs=None,\n",
       "       param_grid={'logisticregression__C': [10], 'normalizer': [None, Normalizer(copy=True, norm='l2')], 'countvectorizer__ngram_range': [(2, 4), (3, 6)]},\n",
       "       pre_dispatch='2*n_jobs', refit=True, return_train_score='warn',\n",
       "       scoring='roc_auc', verbose=0)"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.random.seed(2019)\n",
    "param_grid_char = {\"logisticregression__C\": [10],#np.logspace(-1,1,3),\n",
    "                   \"normalizer\": [None, Normalizer()],\n",
    "                   \"countvectorizer__ngram_range\": [(2, 4), (3, 6)]}\n",
    "pipe_char = make_pipeline(CountVectorizer(analyzer=\"char_wb\", min_df=4),\n",
    "                          Normalizer(),\n",
    "                          TfidfTransformer(),\n",
    "                          LogisticRegression(solver='liblinear'))\n",
    "grid_char = GridSearchCV(pipe_char, param_grid=param_grid_char, cv=5, scoring=\"roc_auc\")\n",
    "grid_char.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.976384738510051"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.mean(grid_char.cv_results_['mean_test_score'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9848476706366946"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "grid_char.score(X_test, y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1224x504 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "voc_word_char = grid_char.best_estimator_. \\\n",
    "                    named_steps['countvectorizer'].vocabulary_\n",
    "voc_index_char = {y:x for x,y in voc_word_char.items()}\n",
    "# Coefficientes\n",
    "coefs_char = grid_char.best_estimator_.named_steps['logisticregression'].coef_[0]\n",
    "plot_coefficients(coefs_char, voc_index_char)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Adding text features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[nltk_data] Downloading package punkt to\n",
      "[nltk_data]     /Users/andreanavarrete/nltk_data...\n",
      "[nltk_data]   Package punkt is already up-to-date!\n"
     ]
    }
   ],
   "source": [
    "import re\n",
    "import nltk\n",
    "nltk.download('punkt')\n",
    "from nltk.tokenize import sent_tokenize\n",
    "\n",
    "def num_hashtags(body):\n",
    "    r = re.findall(r\"(?=(# ))\", body)\n",
    "    return len(r)\n",
    "\n",
    "def num_tags(body):\n",
    "    r = re.findall(r\"(?=(@ ))\", body)\n",
    "    return len(r)\n",
    "\n",
    "def get_num_sentences(body):\n",
    "    return len(sent_tokenize(body))\n",
    "    \n",
    "def find_html(body):\n",
    "    string_lst = ['https://', 'www.', 'http://', 'html']\n",
    "    r = re.findall(r\"(?=(\"+'|'.join(string_lst)+r\"))\", body)\n",
    "    if len(r) > 0:\n",
    "        return 1\n",
    "    else:\n",
    "        return 0\n",
    "\n",
    "def percentage_capital(body):\n",
    "    cap = sum(1 for c in body if c.isupper())\n",
    "    return cap/ sum(1 for c in test if c.isalpha())\n",
    "    \n",
    "\n",
    "def num_exclamation(body):\n",
    "    r = re.findall(r\"(?=(!))\", body)\n",
    "    return len(r)\n",
    "\n",
    "def more_than_three(body):\n",
    "    r = re.findall(r'((\\w)\\2{3,})', body)\n",
    "    if len(r) > 0:\n",
    "        return 1\n",
    "    else:\n",
    "        return 0\n",
    "\n",
    "def num_questions(body):\n",
    "    r = re.findall(r\"(?=(\\? ))\", body)\n",
    "    return len(r)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "39758    @DAJSamarasinghe @FamousRob @nicchristensen Ye...\n",
       "18457    @dougom Conservatives are profoundly sick of R...\n",
       "33239    @ramblesofpower It's not \"islamophobia\" it's \"...\n",
       "35006    @KyaYaarKuchBi @MariamVeiszadeh \\nI am sorry, ...\n",
       "30982    @DailyMailAU @joeleigh41 \\n the BBC translated...\n",
       "Name: tweet_text, dtype: object"
      ]
     },
     "execution_count": 63,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 93,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Length\n",
    "tweets['tweet_length'] = tweets['tweet_text'].map(len)\n",
    "# Number of sentences\n",
    "tweets['num_sentences'] =  tweets['tweet_text'].map(get_num_sentences)\n",
    "# Number of hastags\n",
    "tweets['num_hashtags'] = tweets['tweet_text'].map(num_hashtags)\n",
    "# Number of taggs\n",
    "tweets['num_tags'] = tweets['tweet_text'].map(num_tags)\n",
    "# find_html\n",
    "tweets['with_html'] = tweets['tweet_text'].map(find_html)\n",
    "# Percentage Capitalize\n",
    "tweets['percentage_capital'] = tweets['tweet_text'].map(percentage_capital)\n",
    "# Number of excalamation marks\n",
    "tweets['num_exclamation'] = tweets['tweet_text'].map(num_exclamation)\n",
    "# Num questions\n",
    "tweets['num_questions'] = tweets['tweet_text'].map(num_questions)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of train tweets 37500\n",
      "Number of test tweets 12500\n",
      "% of class 1 train 0.3871733333333333\n",
      "% of class 1 test 0.38704\n"
     ]
    }
   ],
   "source": [
    "# Features\n",
    "text_features = ['tweet_text', 'tweet_length', 'num_sentences', \n",
    "                 'num_hashtags', 'num_tags', 'with_html',\n",
    "                 'percentage_capital', 'num_exclamation', 'num_questions']\n",
    "\n",
    "# Split data train / test\n",
    "X = tweets[text_features]\n",
    "y = tweets['is_fake']\n",
    "\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=0, shuffle=True)\n",
    "print(\"Number of train tweets\", len(X_train))\n",
    "print(\"Number of test tweets\",len(X_test))\n",
    "\n",
    "# check balances\n",
    "print(\"% of class 1 train\", len(y_train[y_train == 1]) / len(y_train))\n",
    "print(\"% of class 1 test\", len(y_test[y_test == 1]) / len(y_test))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "GridSearchCV(cv=5, error_score='raise-deprecating',\n",
       "       estimator=Pipeline(memory=None,\n",
       "     steps=[('pre', ColumnTransformer(n_jobs=None, remainder='passthrough', sparse_threshold=0.3,\n",
       "         transformer_weights=None,\n",
       "         transformers=[('tfidfvectorizer', TfidfVectorizer(analyzer='char_wb', binary=False, decode_error='strict',\n",
       "        dtype=<class 'numpy.float64'>, encoding='utf-8',...ty='l2', random_state=None, solver='liblinear',\n",
       "          tol=0.0001, verbose=0, warm_start=False))]),\n",
       "       fit_params=None, iid='warn', n_jobs=None,\n",
       "       param_grid={'clf__C': array([ 0.1,  1. , 10. ]), 'pre__tfidfvectorizer__ngram_range': [(2, 4), (3, 6)]},\n",
       "       pre_dispatch='2*n_jobs', refit=True, return_train_score='warn',\n",
       "       scoring='roc_auc', verbose=0)"
      ]
     },
     "execution_count": 71,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.random.seed(2019)\n",
    "param_grid_features = {\"clf__C\": np.logspace(-1,1,3),\n",
    "                       \"pre__tfidfvectorizer__ngram_range\": [(2, 4), (3, 6)]}\n",
    "pipe_vec = make_pipeline(TfidfVectorizer())\n",
    "\n",
    "preprocess = make_column_transformer(\n",
    "            (TfidfVectorizer(analyzer=\"char_wb\", min_df=4), 'tweet_text'),\n",
    "                                     remainder='passthrough')\n",
    "pipe_features = Pipeline([('pre', preprocess),\n",
    "                 ('clf', LogisticRegression(solver='liblinear'))])\n",
    "grid_features = GridSearchCV(pipe_features, \n",
    "                             param_grid=param_grid_features, \n",
    "                             cv=5, scoring=\"roc_auc\")\n",
    "grid_features.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'clf__C': 10.0, 'pre__tfidfvectorizer__ngram_range': (3, 6)}"
      ]
     },
     "execution_count": 72,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "grid_features.best_params_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9402093490299347"
      ]
     },
     "execution_count": 73,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.mean(grid_features.cv_results_['mean_test_score'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9846527895352085"
      ]
     },
     "execution_count": 75,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "grid_features.score(X_test, y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1224x504 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "voc_char_features = grid_features.best_estimator_.named_steps['pre']. \\\n",
    "                        named_transformers_['tfidfvectorizer'].vocabulary_.copy()\n",
    "extra_features = {x:i+len(voc_char_features) \n",
    "                  for i,x in enumerate([x for x in text_features if x!= 'tweet_text'])}\n",
    "voc_char_features.update(extra_features)\n",
    "voc_index_features = {y:x for x,y in voc_char_features.items()}\n",
    "# Coefficientes\n",
    "coefs_features = grid_features.best_estimator_.named_steps['clf'].coef_[0]\n",
    "plot_coefficients(coefs_features, voc_index_features)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[5, 15, 25, 35, 45]"
      ]
     },
     "execution_count": 81,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "list(range(5, 50, 10))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/ensemble/forest.py:246: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/ensemble/forest.py:246: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/ensemble/forest.py:246: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/ensemble/forest.py:246: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/ensemble/forest.py:246: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/ensemble/forest.py:246: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/ensemble/forest.py:246: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/ensemble/forest.py:246: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/ensemble/forest.py:246: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/ensemble/forest.py:246: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/ensemble/forest.py:246: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/ensemble/forest.py:246: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/ensemble/forest.py:246: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/ensemble/forest.py:246: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/ensemble/forest.py:246: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/ensemble/forest.py:246: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/ensemble/forest.py:246: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/ensemble/forest.py:246: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/ensemble/forest.py:246: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/ensemble/forest.py:246: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/ensemble/forest.py:246: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/ensemble/forest.py:246: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/ensemble/forest.py:246: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/ensemble/forest.py:246: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/ensemble/forest.py:246: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/ensemble/forest.py:246: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/ensemble/forest.py:246: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/ensemble/forest.py:246: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/ensemble/forest.py:246: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/ensemble/forest.py:246: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/ensemble/forest.py:246: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/ensemble/forest.py:246: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/ensemble/forest.py:246: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/ensemble/forest.py:246: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/ensemble/forest.py:246: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/ensemble/forest.py:246: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/ensemble/forest.py:246: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/ensemble/forest.py:246: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/ensemble/forest.py:246: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/ensemble/forest.py:246: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/ensemble/forest.py:246: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/ensemble/forest.py:246: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/ensemble/forest.py:246: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/ensemble/forest.py:246: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/ensemble/forest.py:246: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/ensemble/forest.py:246: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/ensemble/forest.py:246: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/ensemble/forest.py:246: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/ensemble/forest.py:246: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/ensemble/forest.py:246: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/ensemble/forest.py:246: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "GridSearchCV(cv=5, error_score='raise-deprecating',\n",
       "       estimator=Pipeline(memory=None,\n",
       "     steps=[('pre', ColumnTransformer(n_jobs=None, remainder='passthrough', sparse_threshold=0.3,\n",
       "         transformer_weights=None,\n",
       "         transformers=[('tfidfvectorizer', TfidfVectorizer(analyzer='char_wb', binary=False, decode_error='strict',\n",
       "        dtype=<class 'numpy.float64'>, encoding='utf-8',...tors='warn', n_jobs=None,\n",
       "            oob_score=False, random_state=0, verbose=0, warm_start=True))]),\n",
       "       fit_params=None, iid='warn', n_jobs=None,\n",
       "       param_grid={'clf__max_leaf_nodes': range(5, 50, 10), 'pre__tfidfvectorizer__ngram_range': [(2, 4), (3, 6)]},\n",
       "       pre_dispatch='2*n_jobs', refit=True, return_train_score='warn',\n",
       "       scoring='roc_auc', verbose=0)"
      ]
     },
     "execution_count": 82,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.random.seed(2019)\n",
    "param_grid_features = {'clf__max_leaf_nodes':range(5, 50, 10),\n",
    "                       \"pre__tfidfvectorizer__ngram_range\": [(2, 4), (3, 6)]}\n",
    "\n",
    "preprocess = make_column_transformer(\n",
    "            (TfidfVectorizer(analyzer=\"char_wb\", min_df=4), 'tweet_text'),\n",
    "                                     remainder='passthrough')\n",
    "pipe_features = Pipeline([('pre', preprocess),\n",
    "                 ('clf', RandomForestClassifier(warm_start=True, random_state=0))])\n",
    "grid_features = GridSearchCV(pipe_features, \n",
    "                             param_grid=param_grid_features, \n",
    "                             cv=5, scoring=\"roc_auc\")\n",
    "grid_features.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'clf__max_leaf_nodes': 45, 'pre__tfidfvectorizer__ngram_range': (3, 6)}"
      ]
     },
     "execution_count": 83,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "grid_features.best_params_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7678059511679496"
      ]
     },
     "execution_count": 84,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.mean(grid_features.cv_results_['mean_test_score'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.8186840286736354"
      ]
     },
     "execution_count": 85,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "grid_features.score(X_test, y_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## All Features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 94,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of train tweets 37500\n",
      "Number of test tweets 12500\n",
      "% of class 1 train 0.3871733333333333\n",
      "% of class 1 test 0.38704\n"
     ]
    }
   ],
   "source": [
    "text_features = ['tweet_text', 'tweet_length', 'num_sentences', \n",
    "                 'num_hashtags', 'num_tags', 'with_html',\n",
    "                 'percentage_capital', 'num_exclamation', 'num_questions']\n",
    "\n",
    "tweet_features = ['tweet_hashtag', 'retweet', 'retweet_count',\n",
    "                  'user_verified', 'user_followers_count', \n",
    "                  'user_statuses_count', 'user_friends_count',\n",
    "                  'user_favourites_count', 'tweet_relative_age']\n",
    "\n",
    "all_features = text_features + tweet_features\n",
    "# Split data train / test\n",
    "X = tweets[all_features]\n",
    "y = tweets['is_fake']\n",
    "\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=0, shuffle=True)\n",
    "print(\"Number of train tweets\", len(X_train))\n",
    "print(\"Number of test tweets\",len(X_test))\n",
    "\n",
    "# check balances\n",
    "print(\"% of class 1 train\", len(y_train[y_train == 1]) / len(y_train))\n",
    "print(\"% of class 1 test\", len(y_test[y_test == 1]) / len(y_test))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 97,
   "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>tweet_text</th>\n",
       "      <th>tweet_length</th>\n",
       "      <th>num_sentences</th>\n",
       "      <th>num_hashtags</th>\n",
       "      <th>num_tags</th>\n",
       "      <th>with_html</th>\n",
       "      <th>percentage_capital</th>\n",
       "      <th>num_exclamation</th>\n",
       "      <th>num_questions</th>\n",
       "      <th>tweet_hashtag</th>\n",
       "      <th>retweet</th>\n",
       "      <th>retweet_count</th>\n",
       "      <th>user_verified</th>\n",
       "      <th>user_followers_count</th>\n",
       "      <th>user_statuses_count</th>\n",
       "      <th>user_friends_count</th>\n",
       "      <th>user_favourites_count</th>\n",
       "      <th>tweet_relative_age</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>39758</th>\n",
       "      <td>@DAJSamarasinghe @FamousRob @nicchristensen Ye...</td>\n",
       "      <td>81</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.142857</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>sydneysiege</td>\n",
       "      <td>False</td>\n",
       "      <td>1</td>\n",
       "      <td>False</td>\n",
       "      <td>792</td>\n",
       "      <td>23494</td>\n",
       "      <td>1055</td>\n",
       "      <td>1239</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18457</th>\n",
       "      <td>@dougom Conservatives are profoundly sick of R...</td>\n",
       "      <td>77</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.031746</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>sydneysiege</td>\n",
       "      <td>False</td>\n",
       "      <td>0</td>\n",
       "      <td>False</td>\n",
       "      <td>16019</td>\n",
       "      <td>251219</td>\n",
       "      <td>2656</td>\n",
       "      <td>84337</td>\n",
       "      <td>11</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>33239</th>\n",
       "      <td>@ramblesofpower It's not \"islamophobia\" it's \"...</td>\n",
       "      <td>126</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.031746</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>sydneysiege</td>\n",
       "      <td>False</td>\n",
       "      <td>1</td>\n",
       "      <td>False</td>\n",
       "      <td>3157</td>\n",
       "      <td>19659</td>\n",
       "      <td>2995</td>\n",
       "      <td>8230</td>\n",
       "      <td>11</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>35006</th>\n",
       "      <td>@KyaYaarKuchBi @MariamVeiszadeh \\nI am sorry, ...</td>\n",
       "      <td>127</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.126984</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>sydneysiege</td>\n",
       "      <td>False</td>\n",
       "      <td>1</td>\n",
       "      <td>False</td>\n",
       "      <td>192</td>\n",
       "      <td>6652</td>\n",
       "      <td>479</td>\n",
       "      <td>122</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30982</th>\n",
       "      <td>@DailyMailAU @joeleigh41 \\n the BBC translated...</td>\n",
       "      <td>118</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.317460</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>sydneysiege</td>\n",
       "      <td>False</td>\n",
       "      <td>0</td>\n",
       "      <td>False</td>\n",
       "      <td>1588</td>\n",
       "      <td>55787</td>\n",
       "      <td>1818</td>\n",
       "      <td>34</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                              tweet_text  tweet_length  \\\n",
       "39758  @DAJSamarasinghe @FamousRob @nicchristensen Ye...            81   \n",
       "18457  @dougom Conservatives are profoundly sick of R...            77   \n",
       "33239  @ramblesofpower It's not \"islamophobia\" it's \"...           126   \n",
       "35006  @KyaYaarKuchBi @MariamVeiszadeh \\nI am sorry, ...           127   \n",
       "30982  @DailyMailAU @joeleigh41 \\n the BBC translated...           118   \n",
       "\n",
       "       num_sentences  num_hashtags  num_tags  with_html  percentage_capital  \\\n",
       "39758              2             0         0          0            0.142857   \n",
       "18457              1             0         0          0            0.031746   \n",
       "33239              2             0         0          0            0.031746   \n",
       "35006              1             0         0          0            0.126984   \n",
       "30982              2             0         0          0            0.317460   \n",
       "\n",
       "       num_exclamation  num_questions tweet_hashtag  retweet  retweet_count  \\\n",
       "39758                0              0   sydneysiege    False              1   \n",
       "18457                0              0   sydneysiege    False              0   \n",
       "33239                0              0   sydneysiege    False              1   \n",
       "35006                0              0   sydneysiege    False              1   \n",
       "30982                1              0   sydneysiege    False              0   \n",
       "\n",
       "       user_verified  user_followers_count  user_statuses_count  \\\n",
       "39758          False                   792                23494   \n",
       "18457          False                 16019               251219   \n",
       "33239          False                  3157                19659   \n",
       "35006          False                   192                 6652   \n",
       "30982          False                  1588                55787   \n",
       "\n",
       "       user_friends_count  user_favourites_count  tweet_relative_age  \n",
       "39758                1055                   1239                   1  \n",
       "18457                2656                  84337                  11  \n",
       "33239                2995                   8230                  11  \n",
       "35006                 479                    122                   5  \n",
       "30982                1818                     34                   0  "
      ]
     },
     "execution_count": 97,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 101,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/preprocessing/data.py:625: DataConversionWarning: Data with input dtype int64, float64 were all converted to float64 by StandardScaler.\n",
      "  return self.partial_fit(X, y)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/base.py:465: DataConversionWarning: Data with input dtype int64, float64 were all converted to float64 by StandardScaler.\n",
      "  return self.fit(X, y, **fit_params).transform(X)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/svm/base.py:931: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/pipeline.py:605: DataConversionWarning: Data with input dtype int64, float64 were all converted to float64 by StandardScaler.\n",
      "  res = transformer.transform(X)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/pipeline.py:605: DataConversionWarning: Data with input dtype int64, float64 were all converted to float64 by StandardScaler.\n",
      "  res = transformer.transform(X)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/preprocessing/data.py:625: DataConversionWarning: Data with input dtype int64, float64 were all converted to float64 by StandardScaler.\n",
      "  return self.partial_fit(X, y)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/base.py:465: DataConversionWarning: Data with input dtype int64, float64 were all converted to float64 by StandardScaler.\n",
      "  return self.fit(X, y, **fit_params).transform(X)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/svm/base.py:931: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/pipeline.py:605: DataConversionWarning: Data with input dtype int64, float64 were all converted to float64 by StandardScaler.\n",
      "  res = transformer.transform(X)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/pipeline.py:605: DataConversionWarning: Data with input dtype int64, float64 were all converted to float64 by StandardScaler.\n",
      "  res = transformer.transform(X)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/preprocessing/data.py:625: DataConversionWarning: Data with input dtype int64, float64 were all converted to float64 by StandardScaler.\n",
      "  return self.partial_fit(X, y)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/base.py:465: DataConversionWarning: Data with input dtype int64, float64 were all converted to float64 by StandardScaler.\n",
      "  return self.fit(X, y, **fit_params).transform(X)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/pipeline.py:605: DataConversionWarning: Data with input dtype int64, float64 were all converted to float64 by StandardScaler.\n",
      "  res = transformer.transform(X)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/pipeline.py:605: DataConversionWarning: Data with input dtype int64, float64 were all converted to float64 by StandardScaler.\n",
      "  res = transformer.transform(X)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/preprocessing/data.py:625: DataConversionWarning: Data with input dtype int64, float64 were all converted to float64 by StandardScaler.\n",
      "  return self.partial_fit(X, y)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/base.py:465: DataConversionWarning: Data with input dtype int64, float64 were all converted to float64 by StandardScaler.\n",
      "  return self.fit(X, y, **fit_params).transform(X)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/svm/base.py:931: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/pipeline.py:605: DataConversionWarning: Data with input dtype int64, float64 were all converted to float64 by StandardScaler.\n",
      "  res = transformer.transform(X)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/pipeline.py:605: DataConversionWarning: Data with input dtype int64, float64 were all converted to float64 by StandardScaler.\n",
      "  res = transformer.transform(X)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/preprocessing/data.py:625: DataConversionWarning: Data with input dtype int64, float64 were all converted to float64 by StandardScaler.\n",
      "  return self.partial_fit(X, y)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/base.py:465: DataConversionWarning: Data with input dtype int64, float64 were all converted to float64 by StandardScaler.\n",
      "  return self.fit(X, y, **fit_params).transform(X)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/svm/base.py:931: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/pipeline.py:605: DataConversionWarning: Data with input dtype int64, float64 were all converted to float64 by StandardScaler.\n",
      "  res = transformer.transform(X)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/pipeline.py:605: DataConversionWarning: Data with input dtype int64, float64 were all converted to float64 by StandardScaler.\n",
      "  res = transformer.transform(X)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/preprocessing/data.py:625: DataConversionWarning: Data with input dtype int64, float64 were all converted to float64 by StandardScaler.\n",
      "  return self.partial_fit(X, y)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/base.py:465: DataConversionWarning: Data with input dtype int64, float64 were all converted to float64 by StandardScaler.\n",
      "  return self.fit(X, y, **fit_params).transform(X)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/pipeline.py:605: DataConversionWarning: Data with input dtype int64, float64 were all converted to float64 by StandardScaler.\n",
      "  res = transformer.transform(X)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/pipeline.py:605: DataConversionWarning: Data with input dtype int64, float64 were all converted to float64 by StandardScaler.\n",
      "  res = transformer.transform(X)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/preprocessing/data.py:625: DataConversionWarning: Data with input dtype int64, float64 were all converted to float64 by StandardScaler.\n",
      "  return self.partial_fit(X, y)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/base.py:465: DataConversionWarning: Data with input dtype int64, float64 were all converted to float64 by StandardScaler.\n",
      "  return self.fit(X, y, **fit_params).transform(X)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/pipeline.py:605: DataConversionWarning: Data with input dtype int64, float64 were all converted to float64 by StandardScaler.\n",
      "  res = transformer.transform(X)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/pipeline.py:605: DataConversionWarning: Data with input dtype int64, float64 were all converted to float64 by StandardScaler.\n",
      "  res = transformer.transform(X)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/preprocessing/data.py:625: DataConversionWarning: Data with input dtype int64, float64 were all converted to float64 by StandardScaler.\n",
      "  return self.partial_fit(X, y)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/base.py:465: DataConversionWarning: Data with input dtype int64, float64 were all converted to float64 by StandardScaler.\n",
      "  return self.fit(X, y, **fit_params).transform(X)\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/svm/base.py:931: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/pipeline.py:605: DataConversionWarning: Data with input dtype int64, float64 were all converted to float64 by StandardScaler.\n",
      "  res = transformer.transform(X)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/pipeline.py:605: DataConversionWarning: Data with input dtype int64, float64 were all converted to float64 by StandardScaler.\n",
      "  res = transformer.transform(X)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/preprocessing/data.py:625: DataConversionWarning: Data with input dtype int64, float64 were all converted to float64 by StandardScaler.\n",
      "  return self.partial_fit(X, y)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/base.py:465: DataConversionWarning: Data with input dtype int64, float64 were all converted to float64 by StandardScaler.\n",
      "  return self.fit(X, y, **fit_params).transform(X)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/svm/base.py:931: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/pipeline.py:605: DataConversionWarning: Data with input dtype int64, float64 were all converted to float64 by StandardScaler.\n",
      "  res = transformer.transform(X)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/pipeline.py:605: DataConversionWarning: Data with input dtype int64, float64 were all converted to float64 by StandardScaler.\n",
      "  res = transformer.transform(X)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/preprocessing/data.py:625: DataConversionWarning: Data with input dtype int64, float64 were all converted to float64 by StandardScaler.\n",
      "  return self.partial_fit(X, y)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/base.py:465: DataConversionWarning: Data with input dtype int64, float64 were all converted to float64 by StandardScaler.\n",
      "  return self.fit(X, y, **fit_params).transform(X)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/svm/base.py:931: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/pipeline.py:605: DataConversionWarning: Data with input dtype int64, float64 were all converted to float64 by StandardScaler.\n",
      "  res = transformer.transform(X)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/pipeline.py:605: DataConversionWarning: Data with input dtype int64, float64 were all converted to float64 by StandardScaler.\n",
      "  res = transformer.transform(X)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/preprocessing/data.py:625: DataConversionWarning: Data with input dtype int64, float64 were all converted to float64 by StandardScaler.\n",
      "  return self.partial_fit(X, y)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/base.py:465: DataConversionWarning: Data with input dtype int64, float64 were all converted to float64 by StandardScaler.\n",
      "  return self.fit(X, y, **fit_params).transform(X)\n",
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/svm/base.py:931: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "GridSearchCV(cv=5, error_score='raise-deprecating',\n",
       "       estimator=Pipeline(memory=None,\n",
       "     steps=[('pre', ColumnTransformer(n_jobs=None, remainder='passthrough', sparse_threshold=0.3,\n",
       "         transformer_weights=None,\n",
       "         transformers=[('tfidfvectorizer', TfidfVectorizer(analyzer='char_wb', binary=False, decode_error='strict',\n",
       "        dtype=<class 'numpy.float64'>, encoding='utf-8',...ty='l2', random_state=None, solver='liblinear',\n",
       "          tol=0.0001, verbose=0, warm_start=False))]),\n",
       "       fit_params=None, iid='warn', n_jobs=None,\n",
       "       param_grid={'clf__C': [10], 'pre__tfidfvectorizer__ngram_range': [(2, 4), (3, 6)]},\n",
       "       pre_dispatch='2*n_jobs', refit=True, return_train_score='warn',\n",
       "       scoring='roc_auc', verbose=0)"
      ]
     },
     "execution_count": 101,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.random.seed(2019)\n",
    "\n",
    "numeric_cols = X_train.select_dtypes(include=[np.number]).columns.tolist()\n",
    "param_grid_features = {\"clf__C\": [10], #np.logspace(-1,1,3),\n",
    "                       \"pre__tfidfvectorizer__ngram_range\": [(2, 4), (3, 6)]}\n",
    "pipe_vec = make_pipeline(TfidfVectorizer())\n",
    "\n",
    "preprocess = make_column_transformer(\n",
    "            (TfidfVectorizer(analyzer=\"char_wb\", min_df=4), 'tweet_text'),\n",
    "            (StandardScaler(), numeric_cols),\n",
    "            (OneHotEncoder(handle_unknown='ignore'), ['tweet_hashtag']),\n",
    "                                     remainder='passthrough')\n",
    "pipe_features = Pipeline([('pre', preprocess),\n",
    "                 ('clf', LogisticRegression(solver='liblinear'))])\n",
    "grid_features = GridSearchCV(pipe_features, \n",
    "                             param_grid=param_grid_features, \n",
    "                             cv=5, scoring=\"roc_auc\")\n",
    "grid_features.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 102,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'clf__C': 10, 'pre__tfidfvectorizer__ngram_range': (3, 6)}"
      ]
     },
     "execution_count": 102,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "grid_features.best_params_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 103,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9697471331269765"
      ]
     },
     "execution_count": 103,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.mean(grid_features.cv_results_['mean_test_score'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/andreanavarrete/.pyenv/versions/3.6.0/lib/python3.6/site-packages/sklearn/pipeline.py:605: DataConversionWarning: Data with input dtype int64, float64 were all converted to float64 by StandardScaler.\n",
      "  res = transformer.transform(X)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "0.9743199097374619"
      ]
     },
     "execution_count": 104,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "grid_features.score(X_test, y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 128,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1224x504 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "voc_char_features = grid_features.best_estimator_.named_steps['pre']. \\\n",
    "                        named_transformers_['tfidfvectorizer'].vocabulary_.copy()\n",
    "numeric_features = {x:i+len(voc_char_features) \n",
    "                  for i,x in enumerate(numeric_cols)}\n",
    "voc_char_features.update(numeric_features)\n",
    "\n",
    "onehot = list(grid_features.best_estimator_.named_steps['pre']. \\\n",
    "                named_transformers_['onehotencoder'].get_feature_names())\n",
    "onehot_features = {x:i+len(voc_char_features) \n",
    "                  for i,x in enumerate(onehot)}\n",
    "voc_char_features.update(onehot_features)\n",
    "\n",
    "extra_features = {x:i+len(voc_char_features) \n",
    "                  for i,x in enumerate([x for x in all_features if x not in ['tweet_text', 'tweet_hashtag'] + numeric_cols])}\n",
    "voc_char_features.update(extra_features)\n",
    "\n",
    "voc_index_features = {y:x for x,y in voc_char_features.items()}\n",
    "# Coefficientes\n",
    "coefs_features = grid_features.best_estimator_.named_steps['clf'].coef_[0]\n",
    "plot_coefficients(coefs_features, voc_index_features)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "adaCourse",
   "language": "python",
   "name": "jupyterkern"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}