4-ML_models.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Import Modules"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "6VXPfRp-tARR",
    "outputId": "6c27acaf-2c6a-467f-fd7a-68efce98e30b"
   },
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import re\n",
    "import time\n",
    "import warnings\n",
    "import sqlite3\n",
    "from sqlalchemy import create_engine # database connection\n",
    "import csv\n",
    "import os\n",
    "warnings.filterwarnings(\"ignore\")\n",
    "import datetime as dt\n",
    "import numpy as np\n",
    "from nltk.corpus import stopwords\n",
    "from sklearn.decomposition import TruncatedSVD\n",
    "from sklearn.preprocessing import normalize\n",
    "from sklearn.feature_extraction.text import CountVectorizer\n",
    "from sklearn.manifold import TSNE\n",
    "import seaborn as sns\n",
    "from sklearn.neighbors import KNeighborsClassifier\n",
    "from sklearn.metrics import confusion_matrix\n",
    "from sklearn.metrics.classification import accuracy_score, log_loss\n",
    "from sklearn.feature_extraction.text import TfidfVectorizer\n",
    "from collections import Counter\n",
    "from scipy.sparse import hstack\n",
    "from sklearn.multiclass import OneVsRestClassifier\n",
    "from sklearn.svm import SVC\n",
    "from sklearn.model_selection import StratifiedKFold \n",
    "from collections import Counter, defaultdict\n",
    "from sklearn.calibration import CalibratedClassifierCV\n",
    "from sklearn.naive_bayes import MultinomialNB\n",
    "from sklearn.naive_bayes import GaussianNB\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.model_selection import GridSearchCV\n",
    "import math\n",
    "from sklearn.metrics import normalized_mutual_info_score\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from tqdm import tqdm_notebook as tqdm\n",
    "\n",
    "\n",
    "from sklearn.model_selection import cross_val_score\n",
    "from sklearn.linear_model import SGDClassifier\n",
    "from mlxtend.classifier import StackingClassifier\n",
    "\n",
    "from sklearn import model_selection\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.metrics import precision_recall_curve, auc, roc_curve"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "ZihvUPvHtARd"
   },
   "source": [
    "# Machine Learning Models"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "CtN9VBPutARf"
   },
   "source": [
    "## Reading data from file and storing into sql table"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "owBQdjY1tARh"
   },
   "outputs": [],
   "source": [
    "#Creating db file from csv\n",
    "if not os.path.isfile('train.db'):\n",
    "    disk_engine = create_engine('sqlite:///train.db')\n",
    "    start = dt.datetime.now()\n",
    "    chunksize = 180000\n",
    "    j = 0\n",
    "    index_start = 1\n",
    "    for df in pd.read_csv('final_features.csv', names=['Unnamed: 0','id','is_duplicate','cwc_min','cwc_max','csc_min','csc_max','ctc_min','ctc_max','last_word_eq','first_word_eq','abs_len_diff','mean_len','token_set_ratio','token_sort_ratio','fuzz_ratio','fuzz_partial_ratio','longest_substr_ratio','freq_qid1','freq_qid2','q1len','q2len','q1_n_words','q2_n_words','word_Common','word_Total','word_share','freq_q1+q2','freq_q1-q2','0_x','1_x','2_x','3_x','4_x','5_x','6_x','7_x','8_x','9_x','10_x','11_x','12_x','13_x','14_x','15_x','16_x','17_x','18_x','19_x','20_x','21_x','22_x','23_x','24_x','25_x','26_x','27_x','28_x','29_x','30_x','31_x','32_x','33_x','34_x','35_x','36_x','37_x','38_x','39_x','40_x','41_x','42_x','43_x','44_x','45_x','46_x','47_x','48_x','49_x','50_x','51_x','52_x','53_x','54_x','55_x','56_x','57_x','58_x','59_x','60_x','61_x','62_x','63_x','64_x','65_x','66_x','67_x','68_x','69_x','70_x','71_x','72_x','73_x','74_x','75_x','76_x','77_x','78_x','79_x','80_x','81_x','82_x','83_x','84_x','85_x','86_x','87_x','88_x','89_x','90_x','91_x','92_x','93_x','94_x','95_x','96_x','97_x','98_x','99_x','100_x','101_x','102_x','103_x','104_x','105_x','106_x','107_x','108_x','109_x','110_x','111_x','112_x','113_x','114_x','115_x','116_x','117_x','118_x','119_x','120_x','121_x','122_x','123_x','124_x','125_x','126_x','127_x','128_x','129_x','130_x','131_x','132_x','133_x','134_x','135_x','136_x','137_x','138_x','139_x','140_x','141_x','142_x','143_x','144_x','145_x','146_x','147_x','148_x','149_x','150_x','151_x','152_x','153_x','154_x','155_x','156_x','157_x','158_x','159_x','160_x','161_x','162_x','163_x','164_x','165_x','166_x','167_x','168_x','169_x','170_x','171_x','172_x','173_x','174_x','175_x','176_x','177_x','178_x','179_x','180_x','181_x','182_x','183_x','184_x','185_x','186_x','187_x','188_x','189_x','190_x','191_x','192_x','193_x','194_x','195_x','196_x','197_x','198_x','199_x','200_x','201_x','202_x','203_x','204_x','205_x','206_x','207_x','208_x','209_x','210_x','211_x','212_x','213_x','214_x','215_x','216_x','217_x','218_x','219_x','220_x','221_x','222_x','223_x','224_x','225_x','226_x','227_x','228_x','229_x','230_x','231_x','232_x','233_x','234_x','235_x','236_x','237_x','238_x','239_x','240_x','241_x','242_x','243_x','244_x','245_x','246_x','247_x','248_x','249_x','250_x','251_x','252_x','253_x','254_x','255_x','256_x','257_x','258_x','259_x','260_x','261_x','262_x','263_x','264_x','265_x','266_x','267_x','268_x','269_x','270_x','271_x','272_x','273_x','274_x','275_x','276_x','277_x','278_x','279_x','280_x','281_x','282_x','283_x','284_x','285_x','286_x','287_x','288_x','289_x','290_x','291_x','292_x','293_x','294_x','295_x','296_x','297_x','298_x','299_x','300_x','301_x','302_x','303_x','304_x','305_x','306_x','307_x','308_x','309_x','310_x','311_x','312_x','313_x','314_x','315_x','316_x','317_x','318_x','319_x','320_x','321_x','322_x','323_x','324_x','325_x','326_x','327_x','328_x','329_x','330_x','331_x','332_x','333_x','334_x','335_x','336_x','337_x','338_x','339_x','340_x','341_x','342_x','343_x','344_x','345_x','346_x','347_x','348_x','349_x','350_x','351_x','352_x','353_x','354_x','355_x','356_x','357_x','358_x','359_x','360_x','361_x','362_x','363_x','364_x','365_x','366_x','367_x','368_x','369_x','370_x','371_x','372_x','373_x','374_x','375_x','376_x','377_x','378_x','379_x','380_x','381_x','382_x','383_x','0_y','1_y','2_y','3_y','4_y','5_y','6_y','7_y','8_y','9_y','10_y','11_y','12_y','13_y','14_y','15_y','16_y','17_y','18_y','19_y','20_y','21_y','22_y','23_y','24_y','25_y','26_y','27_y','28_y','29_y','30_y','31_y','32_y','33_y','34_y','35_y','36_y','37_y','38_y','39_y','40_y','41_y','42_y','43_y','44_y','45_y','46_y','47_y','48_y','49_y','50_y','51_y','52_y','53_y','54_y','55_y','56_y','57_y','58_y','59_y','60_y','61_y','62_y','63_y','64_y','65_y','66_y','67_y','68_y','69_y','70_y','71_y','72_y','73_y','74_y','75_y','76_y','77_y','78_y','79_y','80_y','81_y','82_y','83_y','84_y','85_y','86_y','87_y','88_y','89_y','90_y','91_y','92_y','93_y','94_y','95_y','96_y','97_y','98_y','99_y','100_y','101_y','102_y','103_y','104_y','105_y','106_y','107_y','108_y','109_y','110_y','111_y','112_y','113_y','114_y','115_y','116_y','117_y','118_y','119_y','120_y','121_y','122_y','123_y','124_y','125_y','126_y','127_y','128_y','129_y','130_y','131_y','132_y','133_y','134_y','135_y','136_y','137_y','138_y','139_y','140_y','141_y','142_y','143_y','144_y','145_y','146_y','147_y','148_y','149_y','150_y','151_y','152_y','153_y','154_y','155_y','156_y','157_y','158_y','159_y','160_y','161_y','162_y','163_y','164_y','165_y','166_y','167_y','168_y','169_y','170_y','171_y','172_y','173_y','174_y','175_y','176_y','177_y','178_y','179_y','180_y','181_y','182_y','183_y','184_y','185_y','186_y','187_y','188_y','189_y','190_y','191_y','192_y','193_y','194_y','195_y','196_y','197_y','198_y','199_y','200_y','201_y','202_y','203_y','204_y','205_y','206_y','207_y','208_y','209_y','210_y','211_y','212_y','213_y','214_y','215_y','216_y','217_y','218_y','219_y','220_y','221_y','222_y','223_y','224_y','225_y','226_y','227_y','228_y','229_y','230_y','231_y','232_y','233_y','234_y','235_y','236_y','237_y','238_y','239_y','240_y','241_y','242_y','243_y','244_y','245_y','246_y','247_y','248_y','249_y','250_y','251_y','252_y','253_y','254_y','255_y','256_y','257_y','258_y','259_y','260_y','261_y','262_y','263_y','264_y','265_y','266_y','267_y','268_y','269_y','270_y','271_y','272_y','273_y','274_y','275_y','276_y','277_y','278_y','279_y','280_y','281_y','282_y','283_y','284_y','285_y','286_y','287_y','288_y','289_y','290_y','291_y','292_y','293_y','294_y','295_y','296_y','297_y','298_y','299_y','300_y','301_y','302_y','303_y','304_y','305_y','306_y','307_y','308_y','309_y','310_y','311_y','312_y','313_y','314_y','315_y','316_y','317_y','318_y','319_y','320_y','321_y','322_y','323_y','324_y','325_y','326_y','327_y','328_y','329_y','330_y','331_y','332_y','333_y','334_y','335_y','336_y','337_y','338_y','339_y','340_y','341_y','342_y','343_y','344_y','345_y','346_y','347_y','348_y','349_y','350_y','351_y','352_y','353_y','354_y','355_y','356_y','357_y','358_y','359_y','360_y','361_y','362_y','363_y','364_y','365_y','366_y','367_y','368_y','369_y','370_y','371_y','372_y','373_y','374_y','375_y','376_y','377_y','378_y','379_y','380_y','381_y','382_y','383_y'], chunksize=chunksize, iterator=True, encoding='utf-8', ):\n",
    "        df.index += index_start\n",
    "        j+=1\n",
    "        print('{} rows'.format(j*chunksize))\n",
    "        df.to_sql('data', disk_engine, if_exists='append')\n",
    "        index_start = df.index[-1] + 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "4hpD3aBktARn"
   },
   "outputs": [],
   "source": [
    "#http://www.sqlitetutorial.net/sqlite-python/create-tables/\n",
    "def create_connection(db_file):\n",
    "    \"\"\" create a database connection to the SQLite database\n",
    "        specified by db_file\n",
    "    :param db_file: database file\n",
    "    :return: Connection object or None\n",
    "    \"\"\"\n",
    "    try:\n",
    "        conn = sqlite3.connect(db_file)\n",
    "        return conn\n",
    "    except Error as e:\n",
    "        print(e)\n",
    " \n",
    "    return None\n",
    "\n",
    "\n",
    "def checkTableExists(dbcon):\n",
    "    cursr = dbcon.cursor()\n",
    "    str = \"select name from sqlite_master where type='table'\"\n",
    "    table_names = cursr.execute(str)\n",
    "    print(\"Tables in the databse:\")\n",
    "    tables =table_names.fetchall() \n",
    "    print(tables[0][0])\n",
    "    return(len(tables))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "nR8ZIUnttARs",
    "outputId": "810fb3fb-7da2-4b78-9e29-9edabbf68cf6"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Tables in the databse:\n",
      "data\n"
     ]
    }
   ],
   "source": [
    "read_db = 'train.db'\n",
    "conn_r = create_connection(read_db)\n",
    "checkTableExists(conn_r)\n",
    "conn_r.close()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "SZq5gaaztARy"
   },
   "outputs": [],
   "source": [
    "# try to sample data according to the computing power you have\n",
    "if os.path.isfile(read_db):\n",
    "    conn_r = create_connection(read_db)\n",
    "    if conn_r is not None:\n",
    "        # for selecting first 1M rows\n",
    "        # data = pd.read_sql_query(\"\"\"SELECT * FROM data LIMIT 100001;\"\"\", conn_r)\n",
    "        \n",
    "        # for selecting random points\n",
    "        data = pd.read_sql_query(\"SELECT * From data ORDER BY RANDOM() LIMIT 100001;\", conn_r)\n",
    "        conn_r.commit()\n",
    "        conn_r.close()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "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>index</th>\n",
       "      <th>Unnamed: 0</th>\n",
       "      <th>id</th>\n",
       "      <th>is_duplicate</th>\n",
       "      <th>cwc_min</th>\n",
       "      <th>cwc_max</th>\n",
       "      <th>csc_min</th>\n",
       "      <th>csc_max</th>\n",
       "      <th>ctc_min</th>\n",
       "      <th>ctc_max</th>\n",
       "      <th>...</th>\n",
       "      <th>374_y</th>\n",
       "      <th>375_y</th>\n",
       "      <th>376_y</th>\n",
       "      <th>377_y</th>\n",
       "      <th>378_y</th>\n",
       "      <th>379_y</th>\n",
       "      <th>380_y</th>\n",
       "      <th>381_y</th>\n",
       "      <th>382_y</th>\n",
       "      <th>383_y</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>176823</td>\n",
       "      <td>176821.0</td>\n",
       "      <td>176821</td>\n",
       "      <td>0</td>\n",
       "      <td>0.749981250468738</td>\n",
       "      <td>0.749981250468738</td>\n",
       "      <td>0.999966667777741</td>\n",
       "      <td>0.599988000239995</td>\n",
       "      <td>0.857130612419823</td>\n",
       "      <td>0.599994000059999</td>\n",
       "      <td>...</td>\n",
       "      <td>2.31778582185507</td>\n",
       "      <td>5.6331538669765</td>\n",
       "      <td>4.89348021522164</td>\n",
       "      <td>-5.17775326967239</td>\n",
       "      <td>9.4263166487217</td>\n",
       "      <td>3.13274863362312</td>\n",
       "      <td>-4.3573300242424</td>\n",
       "      <td>1.61109055578709</td>\n",
       "      <td>-0.855521187186241</td>\n",
       "      <td>10.871232509613</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>479148</td>\n",
       "      <td>299146.0</td>\n",
       "      <td>299146</td>\n",
       "      <td>0</td>\n",
       "      <td>0.39999600004</td>\n",
       "      <td>0.333330555578704</td>\n",
       "      <td>0.285710204139941</td>\n",
       "      <td>0.19999800002</td>\n",
       "      <td>0.315787811643097</td>\n",
       "      <td>0.272726033063486</td>\n",
       "      <td>...</td>\n",
       "      <td>16.3642287962139</td>\n",
       "      <td>53.3244080338627</td>\n",
       "      <td>29.211609069258</td>\n",
       "      <td>-17.5507791563869</td>\n",
       "      <td>35.5258968770504</td>\n",
       "      <td>15.8884926885366</td>\n",
       "      <td>-38.209486505948</td>\n",
       "      <td>4.21350291371346</td>\n",
       "      <td>9.04329504817724</td>\n",
       "      <td>-0.970366135239601</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>483704</td>\n",
       "      <td>303702.0</td>\n",
       "      <td>303702</td>\n",
       "      <td>0</td>\n",
       "      <td>0.999950002499875</td>\n",
       "      <td>0.66664444518516</td>\n",
       "      <td>0.799984000319994</td>\n",
       "      <td>0.799984000319994</td>\n",
       "      <td>0.857130612419823</td>\n",
       "      <td>0.749990625117186</td>\n",
       "      <td>...</td>\n",
       "      <td>7.57879930734634</td>\n",
       "      <td>3.4353878647089</td>\n",
       "      <td>0.5721076130867</td>\n",
       "      <td>-2.40631306171417</td>\n",
       "      <td>2.04250628873706</td>\n",
       "      <td>2.95989117026329</td>\n",
       "      <td>-7.26367550692521</td>\n",
       "      <td>6.86653059720993</td>\n",
       "      <td>5.38648246973753</td>\n",
       "      <td>3.31079116649926</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>49161</td>\n",
       "      <td>49159.0</td>\n",
       "      <td>49159</td>\n",
       "      <td>0</td>\n",
       "      <td>0.499987500312492</td>\n",
       "      <td>0.153844970423304</td>\n",
       "      <td>0.399992000159997</td>\n",
       "      <td>0.19999800002</td>\n",
       "      <td>0.444439506227709</td>\n",
       "      <td>0.153845562132453</td>\n",
       "      <td>...</td>\n",
       "      <td>39.4039672315121</td>\n",
       "      <td>41.0954083353281</td>\n",
       "      <td>-3.2861255928874</td>\n",
       "      <td>-18.175609048456</td>\n",
       "      <td>34.2586436569691</td>\n",
       "      <td>-2.49455424770713</td>\n",
       "      <td>-40.016661837697</td>\n",
       "      <td>37.2811732664704</td>\n",
       "      <td>42.6948520373553</td>\n",
       "      <td>5.89821659773588</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>369844</td>\n",
       "      <td>189842.0</td>\n",
       "      <td>189842</td>\n",
       "      <td>1</td>\n",
       "      <td>0.999980000399992</td>\n",
       "      <td>0.555549382784636</td>\n",
       "      <td>0.66664444518516</td>\n",
       "      <td>0.285710204139941</td>\n",
       "      <td>0.874989062636717</td>\n",
       "      <td>0.411762283751272</td>\n",
       "      <td>...</td>\n",
       "      <td>-10.4165788888931</td>\n",
       "      <td>0.388532944023609</td>\n",
       "      <td>9.99660426750779</td>\n",
       "      <td>0.138266384601593</td>\n",
       "      <td>1.02059417963028</td>\n",
       "      <td>1.12199047207832</td>\n",
       "      <td>-13.5841726064682</td>\n",
       "      <td>-5.08992633223534</td>\n",
       "      <td>7.334357842803</td>\n",
       "      <td>7.29962074756622</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 798 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "    index  Unnamed: 0      id is_duplicate            cwc_min  \\\n",
       "0  176823    176821.0  176821            0  0.749981250468738   \n",
       "1  479148    299146.0  299146            0      0.39999600004   \n",
       "2  483704    303702.0  303702            0  0.999950002499875   \n",
       "3   49161     49159.0   49159            0  0.499987500312492   \n",
       "4  369844    189842.0  189842            1  0.999980000399992   \n",
       "\n",
       "             cwc_max            csc_min            csc_max            ctc_min  \\\n",
       "0  0.749981250468738  0.999966667777741  0.599988000239995  0.857130612419823   \n",
       "1  0.333330555578704  0.285710204139941      0.19999800002  0.315787811643097   \n",
       "2   0.66664444518516  0.799984000319994  0.799984000319994  0.857130612419823   \n",
       "3  0.153844970423304  0.399992000159997      0.19999800002  0.444439506227709   \n",
       "4  0.555549382784636   0.66664444518516  0.285710204139941  0.874989062636717   \n",
       "\n",
       "             ctc_max  ...              374_y              375_y  \\\n",
       "0  0.599994000059999  ...   2.31778582185507    5.6331538669765   \n",
       "1  0.272726033063486  ...   16.3642287962139   53.3244080338627   \n",
       "2  0.749990625117186  ...   7.57879930734634    3.4353878647089   \n",
       "3  0.153845562132453  ...   39.4039672315121   41.0954083353281   \n",
       "4  0.411762283751272  ...  -10.4165788888931  0.388532944023609   \n",
       "\n",
       "              376_y              377_y             378_y              379_y  \\\n",
       "0  4.89348021522164  -5.17775326967239   9.4263166487217   3.13274863362312   \n",
       "1   29.211609069258  -17.5507791563869  35.5258968770504   15.8884926885366   \n",
       "2   0.5721076130867  -2.40631306171417  2.04250628873706   2.95989117026329   \n",
       "3  -3.2861255928874   -18.175609048456  34.2586436569691  -2.49455424770713   \n",
       "4  9.99660426750779  0.138266384601593  1.02059417963028   1.12199047207832   \n",
       "\n",
       "               380_y              381_y               382_y  \\\n",
       "0   -4.3573300242424   1.61109055578709  -0.855521187186241   \n",
       "1   -38.209486505948   4.21350291371346    9.04329504817724   \n",
       "2  -7.26367550692521   6.86653059720993    5.38648246973753   \n",
       "3   -40.016661837697   37.2811732664704    42.6948520373553   \n",
       "4  -13.5841726064682  -5.08992633223534      7.334357842803   \n",
       "\n",
       "                383_y  \n",
       "0     10.871232509613  \n",
       "1  -0.970366135239601  \n",
       "2    3.31079116649926  \n",
       "3    5.89821659773588  \n",
       "4    7.29962074756622  \n",
       "\n",
       "[5 rows x 798 columns]"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "ZkeBKktKtAR3"
   },
   "outputs": [],
   "source": [
    "# remove the first row \n",
    "data.drop(data.index[0], inplace=True)\n",
    "y_true = data['is_duplicate']\n",
    "data.drop(['Unnamed: 0', 'id','index','is_duplicate'], axis=1, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "QKSenpsmtAR9",
    "outputId": "81d890ce-df79-4402-9324-84817dbd5a7d"
   },
   "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>cwc_min</th>\n",
       "      <th>cwc_max</th>\n",
       "      <th>csc_min</th>\n",
       "      <th>csc_max</th>\n",
       "      <th>ctc_min</th>\n",
       "      <th>ctc_max</th>\n",
       "      <th>last_word_eq</th>\n",
       "      <th>first_word_eq</th>\n",
       "      <th>abs_len_diff</th>\n",
       "      <th>mean_len</th>\n",
       "      <th>...</th>\n",
       "      <th>374_y</th>\n",
       "      <th>375_y</th>\n",
       "      <th>376_y</th>\n",
       "      <th>377_y</th>\n",
       "      <th>378_y</th>\n",
       "      <th>379_y</th>\n",
       "      <th>380_y</th>\n",
       "      <th>381_y</th>\n",
       "      <th>382_y</th>\n",
       "      <th>383_y</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.39999600004</td>\n",
       "      <td>0.333330555578704</td>\n",
       "      <td>0.285710204139941</td>\n",
       "      <td>0.19999800002</td>\n",
       "      <td>0.315787811643097</td>\n",
       "      <td>0.272726033063486</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>20.5</td>\n",
       "      <td>...</td>\n",
       "      <td>16.3642287962139</td>\n",
       "      <td>53.3244080338627</td>\n",
       "      <td>29.211609069258</td>\n",
       "      <td>-17.5507791563869</td>\n",
       "      <td>35.5258968770504</td>\n",
       "      <td>15.8884926885366</td>\n",
       "      <td>-38.209486505948</td>\n",
       "      <td>4.21350291371346</td>\n",
       "      <td>9.04329504817724</td>\n",
       "      <td>-0.970366135239601</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.999950002499875</td>\n",
       "      <td>0.66664444518516</td>\n",
       "      <td>0.799984000319994</td>\n",
       "      <td>0.799984000319994</td>\n",
       "      <td>0.857130612419823</td>\n",
       "      <td>0.749990625117186</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>7.5</td>\n",
       "      <td>...</td>\n",
       "      <td>7.57879930734634</td>\n",
       "      <td>3.4353878647089</td>\n",
       "      <td>0.5721076130867</td>\n",
       "      <td>-2.40631306171417</td>\n",
       "      <td>2.04250628873706</td>\n",
       "      <td>2.95989117026329</td>\n",
       "      <td>-7.26367550692521</td>\n",
       "      <td>6.86653059720993</td>\n",
       "      <td>5.38648246973753</td>\n",
       "      <td>3.31079116649926</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.499987500312492</td>\n",
       "      <td>0.153844970423304</td>\n",
       "      <td>0.399992000159997</td>\n",
       "      <td>0.19999800002</td>\n",
       "      <td>0.444439506227709</td>\n",
       "      <td>0.153845562132453</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>17.0</td>\n",
       "      <td>17.5</td>\n",
       "      <td>...</td>\n",
       "      <td>39.4039672315121</td>\n",
       "      <td>41.0954083353281</td>\n",
       "      <td>-3.2861255928874</td>\n",
       "      <td>-18.175609048456</td>\n",
       "      <td>34.2586436569691</td>\n",
       "      <td>-2.49455424770713</td>\n",
       "      <td>-40.016661837697</td>\n",
       "      <td>37.2811732664704</td>\n",
       "      <td>42.6948520373553</td>\n",
       "      <td>5.89821659773588</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.999980000399992</td>\n",
       "      <td>0.555549382784636</td>\n",
       "      <td>0.66664444518516</td>\n",
       "      <td>0.285710204139941</td>\n",
       "      <td>0.874989062636717</td>\n",
       "      <td>0.411762283751272</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>12.5</td>\n",
       "      <td>...</td>\n",
       "      <td>-10.4165788888931</td>\n",
       "      <td>0.388532944023609</td>\n",
       "      <td>9.99660426750779</td>\n",
       "      <td>0.138266384601593</td>\n",
       "      <td>1.02059417963028</td>\n",
       "      <td>1.12199047207832</td>\n",
       "      <td>-13.5841726064682</td>\n",
       "      <td>-5.08992633223534</td>\n",
       "      <td>7.334357842803</td>\n",
       "      <td>7.29962074756622</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0.999950002499875</td>\n",
       "      <td>0.66664444518516</td>\n",
       "      <td>0.499987500312492</td>\n",
       "      <td>0.399992000159997</td>\n",
       "      <td>0.666655555740738</td>\n",
       "      <td>0.499993750078124</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>...</td>\n",
       "      <td>2.30695885419846</td>\n",
       "      <td>6.21115762321278</td>\n",
       "      <td>3.50843986868858</td>\n",
       "      <td>-4.74290472269058</td>\n",
       "      <td>4.17678442597389</td>\n",
       "      <td>6.0351145118475</td>\n",
       "      <td>-9.57282587885857</td>\n",
       "      <td>3.07779650390148</td>\n",
       "      <td>8.49933102726936</td>\n",
       "      <td>0.682479113340378</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 794 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "             cwc_min            cwc_max            csc_min            csc_max  \\\n",
       "1      0.39999600004  0.333330555578704  0.285710204139941      0.19999800002   \n",
       "2  0.999950002499875   0.66664444518516  0.799984000319994  0.799984000319994   \n",
       "3  0.499987500312492  0.153844970423304  0.399992000159997      0.19999800002   \n",
       "4  0.999980000399992  0.555549382784636   0.66664444518516  0.285710204139941   \n",
       "5  0.999950002499875   0.66664444518516  0.499987500312492  0.399992000159997   \n",
       "\n",
       "             ctc_min            ctc_max last_word_eq first_word_eq  \\\n",
       "1  0.315787811643097  0.272726033063486          0.0           1.0   \n",
       "2  0.857130612419823  0.749990625117186          1.0           1.0   \n",
       "3  0.444439506227709  0.153845562132453          0.0           0.0   \n",
       "4  0.874989062636717  0.411762283751272          0.0           1.0   \n",
       "5  0.666655555740738  0.499993750078124          0.0           1.0   \n",
       "\n",
       "  abs_len_diff mean_len  ...              374_y              375_y  \\\n",
       "1          3.0     20.5  ...   16.3642287962139   53.3244080338627   \n",
       "2          1.0      7.5  ...   7.57879930734634    3.4353878647089   \n",
       "3         17.0     17.5  ...   39.4039672315121   41.0954083353281   \n",
       "4          9.0     12.5  ...  -10.4165788888931  0.388532944023609   \n",
       "5          2.0      7.0  ...   2.30695885419846   6.21115762321278   \n",
       "\n",
       "              376_y              377_y             378_y              379_y  \\\n",
       "1   29.211609069258  -17.5507791563869  35.5258968770504   15.8884926885366   \n",
       "2   0.5721076130867  -2.40631306171417  2.04250628873706   2.95989117026329   \n",
       "3  -3.2861255928874   -18.175609048456  34.2586436569691  -2.49455424770713   \n",
       "4  9.99660426750779  0.138266384601593  1.02059417963028   1.12199047207832   \n",
       "5  3.50843986868858  -4.74290472269058  4.17678442597389    6.0351145118475   \n",
       "\n",
       "               380_y              381_y             382_y               383_y  \n",
       "1   -38.209486505948   4.21350291371346  9.04329504817724  -0.970366135239601  \n",
       "2  -7.26367550692521   6.86653059720993  5.38648246973753    3.31079116649926  \n",
       "3   -40.016661837697   37.2811732664704  42.6948520373553    5.89821659773588  \n",
       "4  -13.5841726064682  -5.08992633223534    7.334357842803    7.29962074756622  \n",
       "5  -9.57282587885857   3.07779650390148  8.49933102726936   0.682479113340378  \n",
       "\n",
       "[5 rows x 794 columns]"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "KaWHDzqUtASD"
   },
   "source": [
    "## Converting strings to numerics"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "iLV60gkptASD",
    "outputId": "f297e0f4-52d5-4ab4-8a43-f0ff82f63698"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Converting datas inside 794 columns to numerics...\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "7b459cb6b11541a185f6a68e3f2408b7",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, max=794), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "# after we read from sql table each entry was read it as a string\n",
    "# we convert all the features into numaric before we apply any model\n",
    "cols = list(data.columns)\n",
    "print(f\"Converting datas inside {len(list(data.columns))} columns to numerics...\")\n",
    "for i in tqdm(cols):\n",
    "    data[i] = data[i].apply(pd.to_numeric)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "_LpfQwc9tASJ"
   },
   "outputs": [],
   "source": [
    "# https://stackoverflow.com/questions/7368789/convert-all-strings-in-a-list-to-int\n",
    "y_true = list(map(int, y_true))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "CuMTqWGutASO"
   },
   "source": [
    "## Random train test split( 70:30)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "3Rat2obGtASP"
   },
   "outputs": [],
   "source": [
    "X_train,X_test, y_train, y_test = train_test_split(data, y_true, stratify=y_true, test_size=0.3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "1Iw9zCHqtASS",
    "outputId": "910b684b-0876-4dd8-e0d9-457846236833"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of data points in train data : (70000, 794)\n",
      "Number of data points in test data : (30000, 794)\n"
     ]
    }
   ],
   "source": [
    "print(\"Number of data points in train data :\",X_train.shape)\n",
    "print(\"Number of data points in test data :\",X_test.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "0oDV15LJtASY",
    "outputId": "70a1e4eb-3f31-4f1e-a53b-ad972978505d"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "---------- Distribution of output variable in train data ----------\n",
      "Class 0:  0.6313571428571428 Class 1:  0.36864285714285716\n",
      "---------- Distribution of output variable in test data ----------\n",
      "Class 0:  0.3686333333333333 Class 1:  0.3686333333333333\n"
     ]
    }
   ],
   "source": [
    "print(\"-\"*10, \"Distribution of output variable in train data\", \"-\"*10)\n",
    "train_distr = Counter(y_train)  #will return a dictionaly of howmany 1 and zero. eg {0: 34, 1:78}\n",
    "train_len = len(y_train)\n",
    "print(\"Class 0: \",int(train_distr[0])/train_len,\"Class 1: \", int(train_distr[1])/train_len)\n",
    "print(\"-\"*10, \"Distribution of output variable in test data\", \"-\"*10)\n",
    "test_distr = Counter(y_test)\n",
    "test_len = len(y_test)\n",
    "print(\"Class 0: \",int(test_distr[1])/test_len, \"Class 1: \",int(test_distr[1])/test_len)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "XfxcPT6jtASg"
   },
   "outputs": [],
   "source": [
    "# This function plots the confusion matrices given y_i, y_i_hat.\n",
    "def plot_confusion_matrix(test_y, predict_y):\n",
    "    C = confusion_matrix(test_y, predict_y)\n",
    "    # C = 9,9 matrix, each cell (i,j) represents number of points of class i are predicted class j\n",
    "    #[TN FP]\n",
    "    #[FN TP]   \n",
    "    \n",
    "    A =(((C.T)/(C.sum(axis=1))).T)\n",
    "    #precision = true positive / predicted positive\n",
    "    #divid each element of the confusion matrix with the sum of elements in that column\n",
    "    \n",
    "    # C = [[1, 2],\n",
    "    #     [3, 4]]\n",
    "    # C.T = [[1, 3],\n",
    "    #        [2, 4]]\n",
    "    # C.sum(axis = 1)  axis=0 corresonds to columns and axis=1 corresponds to rows in two diamensional array\n",
    "    # C.sum(axix =1) = [[3, 7]]\n",
    "    # ((C.T)/(C.sum(axis=1))) = [[1/3, 3/7]\n",
    "    #                           [2/3, 4/7]]\n",
    "\n",
    "    # ((C.T)/(C.sum(axis=1))).T = [[1/3, 2/3]\n",
    "    #                           [3/7, 4/7]]\n",
    "    # sum of row elements = 1\n",
    "    \n",
    "    B =(C/C.sum(axis=0))\n",
    "    #recall = true positive/ actual positive\n",
    "    #divid each element of the confusion matrix with the sum of elements in that row\n",
    "    # C = [[1, 2],\n",
    "    #     [3, 4]]\n",
    "    # C.sum(axis = 0)  axis=0 corresonds to columns and axis=1 corresponds to rows in two diamensional array\n",
    "    # C.sum(axix =0) = [[4, 6]]\n",
    "    # (C/C.sum(axis=0)) = [[1/4, 2/6],\n",
    "    #                      [3/4, 4/6]] \n",
    "    plt.figure(figsize=(20,4))\n",
    "    \n",
    "    labels = [1,2]\n",
    "    # representing A in heatmap format\n",
    "    cmap=sns.light_palette(\"blue\")\n",
    "    plt.subplot(1, 3, 1)\n",
    "    sns.heatmap(C, annot=True, cmap=cmap, fmt=\".3f\", xticklabels=labels, yticklabels=labels)\n",
    "    plt.xlabel('Predicted Class')\n",
    "    plt.ylabel('Original Class')\n",
    "    plt.title(\"Confusion matrix\")\n",
    "    \n",
    "    plt.subplot(1, 3, 2)\n",
    "    sns.heatmap(B, annot=True, cmap=cmap, fmt=\".3f\", xticklabels=labels, yticklabels=labels)\n",
    "    plt.xlabel('Predicted Class')\n",
    "    plt.ylabel('Original Class')\n",
    "    plt.title(\"Precision matrix\")\n",
    "    \n",
    "    plt.subplot(1, 3, 3)\n",
    "    # representing B in heatmap format\n",
    "    sns.heatmap(A, annot=True, cmap=cmap, fmt=\".3f\", xticklabels=labels, yticklabels=labels)\n",
    "    plt.xlabel('Predicted Class')\n",
    "    plt.ylabel('Original Class')\n",
    "    plt.title(\"Recall matrix\")\n",
    "    \n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "UStQJ5F_tASk"
   },
   "source": [
    "# TFIDF W2V"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Random Model- A Worst Case"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "qwMDqcU7tASl",
    "outputId": "c1e90d53-25ec-445b-e33a-299538520e32"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Log loss on Test Data using Random Model 0.8757882627312501\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x288 with 6 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# we need to generate 9 numbers and the sum of numbers should be 1\n",
    "# one solution is to genarate 9 numbers and divide each of the numbers by their sum\n",
    "# ref: https://stackoverflow.com/a/18662466/4084039\n",
    "# we create a output array that has exactly same size as the CV data\n",
    "predicted_y = np.zeros((test_len,2))\n",
    "for i in range(test_len):\n",
    "    rand_probs = np.random.rand(1,2)\n",
    "    predicted_y[i] = ((rand_probs/sum(sum(rand_probs)))[0])\n",
    "print(\"Log loss on Test Data using Random Model\",log_loss(y_test, predicted_y, eps=1e-15))\n",
    "\n",
    "predicted_y =np.argmax(predicted_y, axis=1)\n",
    "plot_confusion_matrix(y_test, predicted_y)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "YgY29g_qtASq"
   },
   "source": [
    "## Logistics Regression with Hyperparameter Tuning"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "Wb2tOE3GtASr",
    "outputId": "d7e4fc88-7d4e-4313-cda7-462a2409292e"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "For values of alpha =  1e-05 The log loss is: 0.6582243102020666\n",
      "For values of alpha =  0.0001 The log loss is: 0.5239731117023871\n",
      "For values of alpha =  0.001 The log loss is: 0.531030933770585\n",
      "For values of alpha =  0.01 The log loss is: 0.5238064340340219\n",
      "For values of alpha =  0.1 The log loss is: 0.5206027673404094\n",
      "For values of alpha =  1 The log loss is: 0.5145708244250301\n",
      "For values of alpha =  10 The log loss is: 0.5204401069177507\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "For values of best alpha =  1 The train log loss is: 0.5101453297915229\n",
      "For values of best alpha =  1 The test log loss is: 0.5145708244250301\n"
     ]
    }
   ],
   "source": [
    "alpha = [10 ** x for x in range(-5, 2)] # hyperparam for SGD classifier.\n",
    "\n",
    "# read more about SGDClassifier() at http://scikit-learn.org/stable/modules/generated/sklearn.linear_model.SGDClassifier.html\n",
    "# ------------------------------\n",
    "# default parameters\n",
    "# SGDClassifier(loss=’hinge’, penalty=’l2’, alpha=0.0001, l1_ratio=0.15, fit_intercept=True, max_iter=None, tol=None, \n",
    "# shuffle=True, verbose=0, epsilon=0.1, n_jobs=1, random_state=None, learning_rate=’optimal’, eta0=0.0, power_t=0.5, \n",
    "# class_weight=None, warm_start=False, average=False, n_iter=None)\n",
    "\n",
    "# some of methods\n",
    "# fit(X, y[, coef_init, intercept_init, …])\tFit linear model with Stochastic Gradient Descent.\n",
    "# predict(X)\tPredict class labels for samples in X.\n",
    "\n",
    "#-------------------------------\n",
    "# video link: \n",
    "#------------------------------\n",
    "\n",
    "\n",
    "log_error_array=[]\n",
    "for i in alpha:\n",
    "    clf = SGDClassifier(alpha=i, penalty='l2', loss='log', random_state=42)\n",
    "    clf.fit(X_train, y_train)\n",
    "    sig_clf = CalibratedClassifierCV(clf, method=\"sigmoid\")\n",
    "    sig_clf.fit(X_train, y_train)\n",
    "    predict_y = sig_clf.predict_proba(X_test)\n",
    "    log_error_array.append(log_loss(y_test, predict_y, labels=clf.classes_, eps=1e-15))\n",
    "    print('For values of alpha = ', i, \"The log loss is:\",log_loss(y_test, predict_y, labels=clf.classes_, eps=1e-15))\n",
    "\n",
    "fig, ax = plt.subplots()\n",
    "ax.plot(alpha, log_error_array,c='g')\n",
    "for i, txt in enumerate(np.round(log_error_array,3)):\n",
    "    ax.annotate((alpha[i],np.round(txt,3)), (alpha[i],log_error_array[i]))\n",
    "plt.grid()\n",
    "plt.title(\"Cross Validation Error for each alpha\")\n",
    "plt.xlabel(\"Alpha i's\")\n",
    "plt.ylabel(\"Error measure\")\n",
    "plt.show()\n",
    "\n",
    "\n",
    "best_alpha = np.argmin(log_error_array)\n",
    "clf = SGDClassifier(alpha=alpha[best_alpha], penalty='l2', loss='log', random_state=42)\n",
    "clf.fit(X_train, y_train)\n",
    "sig_clf = CalibratedClassifierCV(clf, method=\"sigmoid\")\n",
    "sig_clf.fit(X_train, y_train)\n",
    "\n",
    "predict_y = sig_clf.predict_proba(X_train)\n",
    "print('For values of best alpha = ', alpha[best_alpha], \"The train log loss is:\",log_loss(y_train, predict_y, labels=clf.classes_, eps=1e-15))\n",
    "predict_y = sig_clf.predict_proba(X_test)\n",
    "print('For values of best alpha = ', alpha[best_alpha], \"The test log loss is:\",log_loss(y_test, predict_y, labels=clf.classes_, eps=1e-15))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Total number of data points : 30000\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x288 with 6 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "predicted_y =np.argmax(predict_y,axis=1)\n",
    "print(\"Total number of data points :\", len(predicted_y))\n",
    "plot_confusion_matrix(y_test, predicted_y)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "ouQSEnr3tASy"
   },
   "source": [
    "## Linear SVM with Hyperparameter Tuning"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "AOFfZ5PLtAS0",
    "outputId": "d31eb598-e275-48cb-c49b-98e9eb76d8ba"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "For values of alpha =  1e-05 The log loss is: 0.6582243102020666\n",
      "For values of alpha =  0.0001 The log loss is: 0.4885874377203666\n",
      "For values of alpha =  0.001 The log loss is: 0.5206700965267733\n",
      "For values of alpha =  0.01 The log loss is: 0.562698888770099\n",
      "For values of alpha =  0.1 The log loss is: 0.5849170654028049\n",
      "For values of alpha =  1 The log loss is: 0.633347767594272\n",
      "For values of alpha =  10 The log loss is: 0.6472063567834271\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "For values of best alpha =  0.0001 The train log loss is: 0.4815323702975848\n",
      "For values of best alpha =  0.0001 The test log loss is: 0.4885874377203666\n",
      "Total number of data points : 30000\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x288 with 6 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "alpha = [10 ** x for x in range(-5, 2)] # hyperparam for SGD classifier.\n",
    "\n",
    "# read more about SGDClassifier() at http://scikit-learn.org/stable/modules/generated/sklearn.linear_model.SGDClassifier.html\n",
    "# ------------------------------\n",
    "# default parameters\n",
    "# SGDClassifier(loss=’hinge’, penalty=’l2’, alpha=0.0001, l1_ratio=0.15, fit_intercept=True, max_iter=None, tol=None, \n",
    "# shuffle=True, verbose=0, epsilon=0.1, n_jobs=1, random_state=None, learning_rate=’optimal’, eta0=0.0, power_t=0.5, \n",
    "# class_weight=None, warm_start=False, average=False, n_iter=None)\n",
    "\n",
    "# some of methods\n",
    "# fit(X, y[, coef_init, intercept_init, …])\tFit linear model with Stochastic Gradient Descent.\n",
    "# predict(X)\tPredict class labels for samples in X.\n",
    "\n",
    "#-------------------------------\n",
    "# video link: \n",
    "#------------------------------\n",
    "\n",
    "\n",
    "log_error_array=[]\n",
    "for i in alpha:\n",
    "    clf = SGDClassifier(alpha=i, penalty='l1', loss='hinge', random_state=42)\n",
    "    clf.fit(X_train, y_train)\n",
    "    sig_clf = CalibratedClassifierCV(clf, method=\"sigmoid\")\n",
    "    sig_clf.fit(X_train, y_train)\n",
    "    predict_y = sig_clf.predict_proba(X_test)\n",
    "    log_error_array.append(log_loss(y_test, predict_y, labels=clf.classes_, eps=1e-15))\n",
    "    print('For values of alpha = ', i, \"The log loss is:\",log_loss(y_test, predict_y, labels=clf.classes_, eps=1e-15))\n",
    "\n",
    "fig, ax = plt.subplots()\n",
    "ax.plot(alpha, log_error_array,c='g')\n",
    "for i, txt in enumerate(np.round(log_error_array,3)):\n",
    "    ax.annotate((alpha[i],np.round(txt,3)), (alpha[i],log_error_array[i]))\n",
    "plt.grid()\n",
    "plt.title(\"Cross Validation Error for each alpha\")\n",
    "plt.xlabel(\"Alpha i's\")\n",
    "plt.ylabel(\"Error measure\")\n",
    "plt.show()\n",
    "\n",
    "\n",
    "best_alpha = np.argmin(log_error_array)\n",
    "clf = SGDClassifier(alpha=alpha[best_alpha], penalty='l1', loss='hinge', random_state=42)\n",
    "clf.fit(X_train, y_train)\n",
    "sig_clf = CalibratedClassifierCV(clf, method=\"sigmoid\")\n",
    "sig_clf.fit(X_train, y_train)\n",
    "\n",
    "predict_y = sig_clf.predict_proba(X_train)\n",
    "print('For values of best alpha = ', alpha[best_alpha], \"The train log loss is:\",log_loss(y_train, predict_y, labels=clf.classes_, eps=1e-15))\n",
    "predict_y = sig_clf.predict_proba(X_test)\n",
    "print('For values of best alpha = ', alpha[best_alpha], \"The test log loss is:\",log_loss(y_test, predict_y, labels=clf.classes_, eps=1e-15))\n",
    "predicted_y =np.argmax(predict_y,axis=1)\n",
    "print(\"Total number of data points :\", len(predicted_y))\n",
    "plot_confusion_matrix(y_test, predicted_y)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "ZhTJgclztAS6"
   },
   "source": [
    "## XG Boost"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "9U367-xetAS7",
    "outputId": "167e8588-2ac4-4c6d-ac22-f56a2fce5657"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0]\ttrain-logloss:0.684702\tvalid-logloss:0.684774\n",
      "Multiple eval metrics have been passed: 'valid-logloss' will be used for early stopping.\n",
      "\n",
      "Will train until valid-logloss hasn't improved in 20 rounds.\n",
      "[10]\ttrain-logloss:0.61461\tvalid-logloss:0.615057\n",
      "[20]\ttrain-logloss:0.56415\tvalid-logloss:0.564908\n",
      "[30]\ttrain-logloss:0.526479\tvalid-logloss:0.527611\n",
      "[40]\ttrain-logloss:0.497495\tvalid-logloss:0.498991\n",
      "[50]\ttrain-logloss:0.474632\tvalid-logloss:0.476433\n",
      "[60]\ttrain-logloss:0.456517\tvalid-logloss:0.458502\n",
      "[70]\ttrain-logloss:0.44184\tvalid-logloss:0.443982\n",
      "[80]\ttrain-logloss:0.430138\tvalid-logloss:0.432491\n",
      "[90]\ttrain-logloss:0.420364\tvalid-logloss:0.422893\n",
      "[100]\ttrain-logloss:0.412255\tvalid-logloss:0.414887\n",
      "[110]\ttrain-logloss:0.405624\tvalid-logloss:0.408426\n",
      "[120]\ttrain-logloss:0.399741\tvalid-logloss:0.402732\n",
      "[130]\ttrain-logloss:0.394943\tvalid-logloss:0.398146\n",
      "[140]\ttrain-logloss:0.390933\tvalid-logloss:0.39442\n",
      "[150]\ttrain-logloss:0.387193\tvalid-logloss:0.390891\n",
      "[160]\ttrain-logloss:0.383957\tvalid-logloss:0.387947\n",
      "[170]\ttrain-logloss:0.381141\tvalid-logloss:0.385409\n",
      "[180]\ttrain-logloss:0.378606\tvalid-logloss:0.383082\n",
      "[190]\ttrain-logloss:0.376186\tvalid-logloss:0.381026\n",
      "[200]\ttrain-logloss:0.373989\tvalid-logloss:0.379127\n",
      "[210]\ttrain-logloss:0.371899\tvalid-logloss:0.377363\n",
      "[220]\ttrain-logloss:0.369783\tvalid-logloss:0.375594\n",
      "[230]\ttrain-logloss:0.367874\tvalid-logloss:0.37398\n",
      "[240]\ttrain-logloss:0.366206\tvalid-logloss:0.372621\n",
      "[250]\ttrain-logloss:0.36445\tvalid-logloss:0.37122\n",
      "[260]\ttrain-logloss:0.362802\tvalid-logloss:0.369839\n",
      "[270]\ttrain-logloss:0.361309\tvalid-logloss:0.368695\n",
      "[280]\ttrain-logloss:0.359849\tvalid-logloss:0.36756\n",
      "[290]\ttrain-logloss:0.358444\tvalid-logloss:0.366541\n",
      "[300]\ttrain-logloss:0.357029\tvalid-logloss:0.365522\n",
      "[310]\ttrain-logloss:0.355775\tvalid-logloss:0.364589\n",
      "[320]\ttrain-logloss:0.35456\tvalid-logloss:0.363718\n",
      "[330]\ttrain-logloss:0.353343\tvalid-logloss:0.362912\n",
      "[340]\ttrain-logloss:0.352059\tvalid-logloss:0.361985\n",
      "[350]\ttrain-logloss:0.351003\tvalid-logloss:0.361202\n",
      "[360]\ttrain-logloss:0.350049\tvalid-logloss:0.36056\n",
      "[370]\ttrain-logloss:0.34899\tvalid-logloss:0.359856\n",
      "[380]\ttrain-logloss:0.347852\tvalid-logloss:0.359049\n",
      "[390]\ttrain-logloss:0.346855\tvalid-logloss:0.358359\n",
      "[399]\ttrain-logloss:0.345942\tvalid-logloss:0.357785\n",
      "The test log loss is: 0.3577844689081753\n"
     ]
    }
   ],
   "source": [
    "import xgboost as xgb\n",
    "params = {}\n",
    "params['objective'] = 'binary:logistic'\n",
    "params['eval_metric'] = 'logloss'\n",
    "params['eta'] = 0.02\n",
    "params['max_depth'] = 4\n",
    "\n",
    "d_train = xgb.DMatrix(X_train, label=y_train)\n",
    "d_test = xgb.DMatrix(X_test, label=y_test)\n",
    "\n",
    "watchlist = [(d_train, 'train'), (d_test, 'valid')]\n",
    "\n",
    "bst = xgb.train(params, d_train, 400, watchlist, early_stopping_rounds=20, verbose_eval=10)\n",
    "\n",
    "xgdmat = xgb.DMatrix(X_train,y_train)\n",
    "predict_y = bst.predict(d_test)\n",
    "print(\"The test log loss is:\",log_loss(y_test, predict_y, labels=clf.classes_, eps=1e-15))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "6U5b17AatAS_",
    "outputId": "ca83b680-023b-4bc5-f499-8d8d85c2ff5e"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Total number of data points : 30000\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x288 with 6 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "predicted_y =np.array(predict_y>0.5,dtype=int)\n",
    "print(\"Total number of data points :\", len(predicted_y))\n",
    "plot_confusion_matrix(y_test, predicted_y)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## XG Boost with Hyperparameter Tuning"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 3 folds for each of 10 candidates, totalling 30 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Using backend LokyBackend with 16 concurrent workers.\n",
      "[Parallel(n_jobs=-1)]: Done   3 out of  30 | elapsed:  1.6min remaining: 14.8min\n",
      "[Parallel(n_jobs=-1)]: Done   7 out of  30 | elapsed:  3.4min remaining: 11.0min\n",
      "[Parallel(n_jobs=-1)]: Done  11 out of  30 | elapsed:  4.9min remaining:  8.5min\n",
      "[Parallel(n_jobs=-1)]: Done  15 out of  30 | elapsed:  6.4min remaining:  6.4min\n",
      "[Parallel(n_jobs=-1)]: Done  19 out of  30 | elapsed:  8.1min remaining:  4.7min\n",
      "[Parallel(n_jobs=-1)]: Done  23 out of  30 | elapsed:  8.9min remaining:  2.7min\n",
      "[Parallel(n_jobs=-1)]: Done  27 out of  30 | elapsed: 11.0min remaining:  1.2min\n"
     ]
    }
   ],
   "source": [
    "import xgboost as xgb\n",
    "from sklearn.model_selection import RandomizedSearchCV\n",
    "\n",
    "model = xgb.XGBClassifier(objective='binary:logistic', eval_metric='logloss',n_jobs=-1)\n",
    "params = {\n",
    "        'max_depth': [i for i in range(1,8,2)],\n",
    "        'n_estimators' : [i*10 for i in range(1, 10)],\n",
    "         'gamma': [0, 0.25, 0.5, 1.0],'learning_rate': [0.001, 0.01, 0.1, 0.2, 0,3]         \n",
    "        }\n",
    "\n",
    "xgb_search = RandomizedSearchCV(model, param_distributions=params, scoring = 'neg_log_loss', n_jobs = -1, cv=3, verbose=10)\n",
    "\n",
    "xgb_search.fit(X_train, y_train)\n",
    "\n",
    "print(\"Best score : \",xgb_search.best_score_)\n",
    "print(\"Best Parameters\",xgb_search.best_params_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best score(Logloss) :  0.3550591030395856\n",
      "Best Parameters {'n_estimators': 40, 'max_depth': 7, 'learning_rate': 0.1, 'gamma': 1.0}\n"
     ]
    }
   ],
   "source": [
    "print(\"Best score(Logloss) : \",-(xgb_search.best_score_))\n",
    "print(\"Best Parameters\",xgb_search.best_params_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Logloss of train dataset = 0.3147937625931934\n",
      "Logloss of test dataset = 0.3572127318255224\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x288 with 6 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "model1  = xgb.XGBClassifier(max_depth=7, learning_rate=0.1, n_estimators=40, gamma=1.0, objective='binary:logistic',n_jobs=-1,random_state=21)\n",
    "model1.fit(X_train, y_train)\n",
    "model1_cal = CalibratedClassifierCV(model1, method='sigmoid')\n",
    "model1_cal.fit(X_train,y_train)\n",
    "y_pred_proba = model1_cal.predict_proba(X_test)\n",
    "y_train_proba=  model1_cal.predict_proba(X_train)\n",
    "print(f\"Logloss of train dataset = {log_loss(y_train, y_train_proba)}\")\n",
    "print(f\"Logloss of test dataset = {log_loss(y_test, y_pred_proba)}\")\n",
    "y_pred = np.argmax(y_pred_proba,axis=1)\n",
    "plot_confusion_matrix(y_test,y_pred)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# TFIDF"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [],
   "source": [
    "if os.path.isfile('nlp_features_train.csv'):\n",
    "    dfnlp = pd.read_csv(\"nlp_features_train.csv\",encoding='latin-1')\n",
    "else:\n",
    "    print(\"download nlp_features_train.csv from drive or run previous notebook\")\n",
    "\n",
    "if os.path.isfile('df_fe_without_preprocessing_train.csv'):\n",
    "    dfppro = pd.read_csv(\"df_fe_without_preprocessing_train.csv\",encoding='latin-1')\n",
    "else:\n",
    "    print(\"download df_fe_without_preprocessing_train.csv from drive or run previous notebook\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.read_csv(\"train.csv\")\n",
    "df['question1'] = df['question1'].apply(lambda x: str(x))\n",
    "df['question2'] = df['question2'].apply(lambda x: str(x))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.feature_extraction.text import TfidfVectorizer\n",
    "from sklearn.feature_extraction.text import CountVectorizer\n",
    "import scipy\n",
    "\n",
    "#Vectorise words using TFIDF\n",
    "tfidf = TfidfVectorizer(lowercase=False, max_features=20000)\n",
    "tfidf_q1 = tfidf.fit_transform(df[\"question1\"])\n",
    "tfidf_q2 = tfidf.fit_transform(df[\"question2\"])\n",
    "#Merge the vertorised columns of question 1 and question 2 into a single column.\n",
    "tfidf_vec = hstack((tfidf_q1, tfidf_q2))\n",
    "\n",
    "#Drop repeating and unwanted features from the dataframe.\n",
    "df1 = dfnlp.drop(['qid1','qid2','question1','question2'],axis=1)\n",
    "df2 = dfppro.drop(['qid1','qid2','question1','question2','is_duplicate'],axis=1)\n",
    "#Merge the two dataframe df1 and df2 as a single df.\n",
    "df1 = df1.merge(df2, on='id',how='left')\n",
    "\n",
    "df_x = df1.drop(columns='is_duplicate', axis=1)\n",
    "#Convert the df_x dataframe to sparse matrix\n",
    "df1_sparse = scipy.sparse.csr_matrix(df_x)\n",
    "x = hstack((df1_sparse, tfidf_vec))\n",
    "y = df1[\"is_duplicate\"]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Train test split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 144,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train,X_test, y_train, y_test = train_test_split(x, y, test_size=0.3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 145,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Shape of X Train: (283003, 40027)\n",
      "Shape of X Test: (121287, 40027)\n",
      "Shape of Y Train: (283003,)\n",
      "Shape of Y Test: (121287,)\n"
     ]
    }
   ],
   "source": [
    "print(f\"Shape of X Train: {X_train.shape}\")\n",
    "print(f\"Shape of X Test: {X_test.shape}\")\n",
    "print(f\"Shape of Y Train: {y_train.shape}\")\n",
    "print(f\"Shape of Y Test: {y_test.shape}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Random model - A worst Case"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 146,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "---------- Distribution of output variable in train data ----------\n",
      "Class 0:  0.6295763649148596 Class 1:  0.3704236350851404\n",
      "---------- Distribution of output variable in test data ----------\n",
      "Class 0:  0.3663376948889823 Class 1:  0.3663376948889823\n"
     ]
    }
   ],
   "source": [
    "print(\"-\"*10, \"Distribution of output variable in train data\", \"-\"*10)\n",
    "train_distr = Counter(y_train)  #will return a dictionaly of howmany 1 and zero. eg {0: 34, 1:78}\n",
    "train_len = len(y_train)\n",
    "print(\"Class 0: \",int(train_distr[0])/train_len,\"Class 1: \", int(train_distr[1])/train_len)\n",
    "print(\"-\"*10, \"Distribution of output variable in test data\", \"-\"*10)\n",
    "test_distr = Counter(y_test)\n",
    "test_len = len(y_test)\n",
    "print(\"Class 0: \",int(test_distr[1])/test_len, \"Class 1: \",int(test_distr[1])/test_len)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 147,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Log loss on Test Data using Random Model 0.886910813515842\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x288 with 6 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# we need to generate 9 numbers and the sum of numbers should be 1\n",
    "# one solution is to genarate 9 numbers and divide each of the numbers by their sum\n",
    "# ref: https://stackoverflow.com/a/18662466/4084039\n",
    "# we create a output array that has exactly same size as the CV data\n",
    "predicted_y = np.zeros((test_len,2))\n",
    "for i in range(test_len):\n",
    "    rand_probs = np.random.rand(1,2)\n",
    "    predicted_y[i] = ((rand_probs/sum(sum(rand_probs)))[0])\n",
    "print(\"Log loss on Test Data using Random Model\",log_loss(y_test, predicted_y, eps=1e-15))\n",
    "\n",
    "predicted_y =np.argmax(predicted_y, axis=1)\n",
    "plot_confusion_matrix(y_test, predicted_y)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Logistic Regression with Hyperparameter Tuning"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 148,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "For values of alpha =  1e-05 The log loss is: 0.6570136275973378\n",
      "For values of alpha =  0.0001 The log loss is: 0.6570136275973378\n",
      "For values of alpha =  0.001 The log loss is: 0.6570136275973378\n",
      "For values of alpha =  0.01 The log loss is: 0.6570136275973378\n",
      "For values of alpha =  0.1 The log loss is: 0.6570136275973378\n",
      "For values of alpha =  1 The log loss is: 0.6570136275973378\n",
      "For values of alpha =  10 The log loss is: 0.6570136275973378\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "For values of best alpha =  1e-05 The train log loss is: 0.6591807615185924\n",
      "For values of best alpha =  1e-05 The test log loss is: 0.6570136275973378\n",
      "Total number of data points : 121287\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x288 with 6 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "alpha = [10 ** x for x in range(-5, 2)] # hyperparam for SGD classifier.\n",
    "\n",
    "# read more about SGDClassifier() at http://scikit-learn.org/stable/modules/generated/sklearn.linear_model.SGDClassifier.html\n",
    "# ------------------------------\n",
    "# default parameters\n",
    "# SGDClassifier(loss=’hinge’, penalty=’l2’, alpha=0.0001, l1_ratio=0.15, fit_intercept=True, max_iter=None, tol=None, \n",
    "# shuffle=True, verbose=0, epsilon=0.1, n_jobs=1, random_state=None, learning_rate=’optimal’, eta0=0.0, power_t=0.5, \n",
    "# class_weight=None, warm_start=False, average=False, n_iter=None)\n",
    "\n",
    "# some of methods\n",
    "# fit(X, y[, coef_init, intercept_init, …])\tFit linear model with Stochastic Gradient Descent.\n",
    "# predict(X)\tPredict class labels for samples in X.\n",
    "\n",
    "#-------------------------------\n",
    "# video link: \n",
    "#------------------------------\n",
    "\n",
    "\n",
    "log_error_array=[]\n",
    "for i in alpha:\n",
    "    clf = SGDClassifier(alpha=i, penalty='l2', loss='log', random_state=42)\n",
    "    clf.fit(X_train, y_train)\n",
    "    sig_clf = CalibratedClassifierCV(clf, method=\"sigmoid\")\n",
    "    sig_clf.fit(X_train, y_train)\n",
    "    predict_y = sig_clf.predict_proba(X_test)\n",
    "    log_error_array.append(log_loss(y_test, predict_y, labels=clf.classes_, eps=1e-15))\n",
    "    print('For values of alpha = ', i, \"The log loss is:\",log_loss(y_test, predict_y, labels=clf.classes_, eps=1e-15))\n",
    "\n",
    "fig, ax = plt.subplots()\n",
    "ax.plot(alpha, log_error_array,c='g')\n",
    "for i, txt in enumerate(np.round(log_error_array,3)):\n",
    "    ax.annotate((alpha[i],np.round(txt,3)), (alpha[i],log_error_array[i]))\n",
    "plt.grid()\n",
    "plt.title(\"Cross Validation Error for each alpha\")\n",
    "plt.xlabel(\"Alpha i's\")\n",
    "plt.ylabel(\"Error measure\")\n",
    "plt.show()\n",
    "\n",
    "\n",
    "best_alpha = np.argmin(log_error_array)\n",
    "clf = SGDClassifier(alpha=alpha[best_alpha], penalty='l2', loss='log', random_state=42)\n",
    "clf.fit(X_train, y_train)\n",
    "sig_clf = CalibratedClassifierCV(clf, method=\"sigmoid\")\n",
    "sig_clf.fit(X_train, y_train)\n",
    "\n",
    "predict_y = sig_clf.predict_proba(X_train)\n",
    "print('For values of best alpha = ', alpha[best_alpha], \"The train log loss is:\",log_loss(y_train, predict_y, labels=clf.classes_, eps=1e-15))\n",
    "predict_y = sig_clf.predict_proba(X_test)\n",
    "print('For values of best alpha = ', alpha[best_alpha], \"The test log loss is:\",log_loss(y_test, predict_y, labels=clf.classes_, eps=1e-15))\n",
    "predicted_y =np.argmax(predict_y,axis=1)\n",
    "print(\"Total number of data points :\", len(predicted_y))\n",
    "plot_confusion_matrix(y_test, predicted_y)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Linear SVM with Hyperparameter Tuning"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 149,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "For values of alpha =  1e-05 The log loss is: 0.6570136275973378\n",
      "For values of alpha =  0.0001 The log loss is: 0.6570136275973378\n",
      "For values of alpha =  0.001 The log loss is: 0.6570136275973378\n",
      "For values of alpha =  0.01 The log loss is: 0.6570136275973378\n",
      "For values of alpha =  0.1 The log loss is: 0.6570136275973378\n",
      "For values of alpha =  1 The log loss is: 0.6393926484757468\n",
      "For values of alpha =  10 The log loss is: 0.6509614055726312\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "For values of best alpha =  1 The train log loss is: 0.6416219868892176\n",
      "For values of best alpha =  1 The test log loss is: 0.6393926484757468\n",
      "Total number of data points : 121287\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x288 with 6 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "alpha = [10 ** x for x in range(-5, 2)] # hyperparam for SGD classifier.\n",
    "\n",
    "# read more about SGDClassifier() at http://scikit-learn.org/stable/modules/generated/sklearn.linear_model.SGDClassifier.html\n",
    "# ------------------------------\n",
    "# default parameters\n",
    "# SGDClassifier(loss=’hinge’, penalty=’l2’, alpha=0.0001, l1_ratio=0.15, fit_intercept=True, max_iter=None, tol=None, \n",
    "# shuffle=True, verbose=0, epsilon=0.1, n_jobs=1, random_state=None, learning_rate=’optimal’, eta0=0.0, power_t=0.5, \n",
    "# class_weight=None, warm_start=False, average=False, n_iter=None)\n",
    "\n",
    "# some of methods\n",
    "# fit(X, y[, coef_init, intercept_init, …])\tFit linear model with Stochastic Gradient Descent.\n",
    "# predict(X)\tPredict class labels for samples in X.\n",
    "\n",
    "#-------------------------------\n",
    "# video link: \n",
    "#------------------------------\n",
    "\n",
    "\n",
    "log_error_array=[]\n",
    "for i in alpha:\n",
    "    clf = SGDClassifier(alpha=i, penalty='l1', loss='hinge', random_state=42)\n",
    "    clf.fit(X_train, y_train)\n",
    "    sig_clf = CalibratedClassifierCV(clf, method=\"sigmoid\")\n",
    "    sig_clf.fit(X_train, y_train)\n",
    "    predict_y = sig_clf.predict_proba(X_test)\n",
    "    log_error_array.append(log_loss(y_test, predict_y, labels=clf.classes_, eps=1e-15))\n",
    "    print('For values of alpha = ', i, \"The log loss is:\",log_loss(y_test, predict_y, labels=clf.classes_, eps=1e-15))\n",
    "\n",
    "fig, ax = plt.subplots()\n",
    "ax.plot(alpha, log_error_array,c='g')\n",
    "for i, txt in enumerate(np.round(log_error_array,3)):\n",
    "    ax.annotate((alpha[i],np.round(txt,3)), (alpha[i],log_error_array[i]))\n",
    "plt.grid()\n",
    "plt.title(\"Cross Validation Error for each alpha\")\n",
    "plt.xlabel(\"Alpha i's\")\n",
    "plt.ylabel(\"Error measure\")\n",
    "plt.show()\n",
    "\n",
    "\n",
    "best_alpha = np.argmin(log_error_array)\n",
    "clf = SGDClassifier(alpha=alpha[best_alpha], penalty='l1', loss='hinge', random_state=42)\n",
    "clf.fit(X_train, y_train)\n",
    "sig_clf = CalibratedClassifierCV(clf, method=\"sigmoid\")\n",
    "sig_clf.fit(X_train, y_train)\n",
    "\n",
    "predict_y = sig_clf.predict_proba(X_train)\n",
    "print('For values of best alpha = ', alpha[best_alpha], \"The train log loss is:\",log_loss(y_train, predict_y, labels=clf.classes_, eps=1e-15))\n",
    "predict_y = sig_clf.predict_proba(X_test)\n",
    "print('For values of best alpha = ', alpha[best_alpha], \"The test log loss is:\",log_loss(y_test, predict_y, labels=clf.classes_, eps=1e-15))\n",
    "predicted_y =np.argmax(predict_y,axis=1)\n",
    "print(\"Total number of data points :\", len(predicted_y))\n",
    "plot_confusion_matrix(y_test, predicted_y)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## XGBoost"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 150,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0]\ttrain-logloss:0.684909\tvalid-logloss:0.684832\n",
      "Multiple eval metrics have been passed: 'valid-logloss' will be used for early stopping.\n",
      "\n",
      "Will train until valid-logloss hasn't improved in 20 rounds.\n",
      "[10]\ttrain-logloss:0.615575\tvalid-logloss:0.615056\n",
      "[20]\ttrain-logloss:0.564875\tvalid-logloss:0.564039\n",
      "[30]\ttrain-logloss:0.526904\tvalid-logloss:0.525881\n",
      "[40]\ttrain-logloss:0.497752\tvalid-logloss:0.496588\n",
      "[50]\ttrain-logloss:0.474809\tvalid-logloss:0.473433\n",
      "[60]\ttrain-logloss:0.456484\tvalid-logloss:0.454979\n",
      "[70]\ttrain-logloss:0.441771\tvalid-logloss:0.440188\n",
      "[80]\ttrain-logloss:0.429998\tvalid-logloss:0.428366\n",
      "[90]\ttrain-logloss:0.420332\tvalid-logloss:0.418676\n",
      "[100]\ttrain-logloss:0.412104\tvalid-logloss:0.410416\n",
      "[110]\ttrain-logloss:0.405405\tvalid-logloss:0.40372\n",
      "[120]\ttrain-logloss:0.399691\tvalid-logloss:0.398011\n",
      "[130]\ttrain-logloss:0.394907\tvalid-logloss:0.393238\n",
      "[140]\ttrain-logloss:0.390972\tvalid-logloss:0.389278\n",
      "[150]\ttrain-logloss:0.387399\tvalid-logloss:0.385707\n",
      "[160]\ttrain-logloss:0.384474\tvalid-logloss:0.382787\n",
      "[170]\ttrain-logloss:0.381926\tvalid-logloss:0.38023\n",
      "[180]\ttrain-logloss:0.379543\tvalid-logloss:0.377848\n",
      "[190]\ttrain-logloss:0.377678\tvalid-logloss:0.375991\n",
      "[200]\ttrain-logloss:0.375921\tvalid-logloss:0.37427\n",
      "[210]\ttrain-logloss:0.3744\tvalid-logloss:0.372789\n",
      "[220]\ttrain-logloss:0.372787\tvalid-logloss:0.37121\n",
      "[230]\ttrain-logloss:0.37135\tvalid-logloss:0.369804\n",
      "[240]\ttrain-logloss:0.36998\tvalid-logloss:0.368477\n",
      "[250]\ttrain-logloss:0.368282\tvalid-logloss:0.366831\n",
      "[260]\ttrain-logloss:0.366692\tvalid-logloss:0.365284\n",
      "[270]\ttrain-logloss:0.365428\tvalid-logloss:0.364051\n",
      "[280]\ttrain-logloss:0.364414\tvalid-logloss:0.363076\n",
      "[290]\ttrain-logloss:0.363545\tvalid-logloss:0.362232\n",
      "[300]\ttrain-logloss:0.362608\tvalid-logloss:0.361334\n",
      "[310]\ttrain-logloss:0.361735\tvalid-logloss:0.360502\n",
      "[320]\ttrain-logloss:0.36099\tvalid-logloss:0.359785\n",
      "[330]\ttrain-logloss:0.360084\tvalid-logloss:0.358922\n",
      "[340]\ttrain-logloss:0.359343\tvalid-logloss:0.358225\n",
      "[350]\ttrain-logloss:0.358601\tvalid-logloss:0.357525\n",
      "[360]\ttrain-logloss:0.357955\tvalid-logloss:0.356921\n",
      "[370]\ttrain-logloss:0.357281\tvalid-logloss:0.356313\n",
      "[380]\ttrain-logloss:0.35656\tvalid-logloss:0.355638\n",
      "[390]\ttrain-logloss:0.355987\tvalid-logloss:0.355116\n",
      "[399]\ttrain-logloss:0.355472\tvalid-logloss:0.354647\n",
      "The test log loss is: 0.35464699561102575\n"
     ]
    }
   ],
   "source": [
    "import xgboost as xgb\n",
    "params = {}\n",
    "params['objective'] = 'binary:logistic'\n",
    "params['eval_metric'] = 'logloss'\n",
    "params['eta'] = 0.02\n",
    "params['max_depth'] = 4\n",
    "\n",
    "d_train = xgb.DMatrix(X_train, label=y_train)\n",
    "d_test = xgb.DMatrix(X_test, label=y_test)\n",
    "\n",
    "watchlist = [(d_train, 'train'), (d_test, 'valid')]\n",
    "\n",
    "bst = xgb.train(params, d_train, 400, watchlist, early_stopping_rounds=20, verbose_eval=10)\n",
    "\n",
    "xgdmat = xgb.DMatrix(X_train,y_train)\n",
    "predict_y = bst.predict(d_test)\n",
    "print(\"The test log loss is:\",log_loss(y_test, predict_y, labels=clf.classes_, eps=1e-15))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## XGBOOST using Hyperparameter Tuning"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 152,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 3 folds for each of 30 candidates, totalling 90 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Using backend LokyBackend with 16 concurrent workers.\n",
      "[Parallel(n_jobs=-1)]: Done   1 tasks      | elapsed:   42.6s\n",
      "[Parallel(n_jobs=-1)]: Done   2 tasks      | elapsed:   45.4s\n",
      "[Parallel(n_jobs=-1)]: Done   3 tasks      | elapsed:   46.5s\n",
      "[Parallel(n_jobs=-1)]: Done   4 tasks      | elapsed:   46.8s\n",
      "[Parallel(n_jobs=-1)]: Done   5 tasks      | elapsed:   47.4s\n",
      "[Parallel(n_jobs=-1)]: Done   6 tasks      | elapsed:   51.3s\n",
      "[Parallel(n_jobs=-1)]: Done   7 tasks      | elapsed:   51.4s\n",
      "[Parallel(n_jobs=-1)]: Done   8 tasks      | elapsed:   51.5s\n",
      "[Parallel(n_jobs=-1)]: Done   9 tasks      | elapsed:   52.4s\n",
      "[Parallel(n_jobs=-1)]: Done  10 tasks      | elapsed:  1.6min\n",
      "[Parallel(n_jobs=-1)]: Done  11 tasks      | elapsed:  1.6min\n",
      "[Parallel(n_jobs=-1)]: Done  12 tasks      | elapsed:  1.7min\n",
      "[Parallel(n_jobs=-1)]: Done  13 tasks      | elapsed:  1.8min\n",
      "[Parallel(n_jobs=-1)]: Done  14 tasks      | elapsed:  2.0min\n",
      "[Parallel(n_jobs=-1)]: Done  15 tasks      | elapsed:  2.0min\n",
      "[Parallel(n_jobs=-1)]: Done  16 tasks      | elapsed:  2.2min\n",
      "[Parallel(n_jobs=-1)]: Done  17 tasks      | elapsed:  2.3min\n",
      "[Parallel(n_jobs=-1)]: Done  18 tasks      | elapsed:  2.3min\n",
      "[Parallel(n_jobs=-1)]: Done  19 tasks      | elapsed:  2.3min\n",
      "[Parallel(n_jobs=-1)]: Done  20 tasks      | elapsed:  2.4min\n",
      "[Parallel(n_jobs=-1)]: Done  21 tasks      | elapsed:  2.4min\n",
      "[Parallel(n_jobs=-1)]: Done  22 tasks      | elapsed:  3.3min\n",
      "[Parallel(n_jobs=-1)]: Done  23 tasks      | elapsed:  3.5min\n",
      "[Parallel(n_jobs=-1)]: Done  24 tasks      | elapsed:  3.7min\n",
      "[Parallel(n_jobs=-1)]: Done  25 tasks      | elapsed:  3.7min\n",
      "[Parallel(n_jobs=-1)]: Done  26 tasks      | elapsed:  3.8min\n",
      "[Parallel(n_jobs=-1)]: Done  27 tasks      | elapsed:  3.8min\n",
      "[Parallel(n_jobs=-1)]: Done  28 tasks      | elapsed:  3.9min\n",
      "[Parallel(n_jobs=-1)]: Done  29 tasks      | elapsed:  4.2min\n",
      "[Parallel(n_jobs=-1)]: Done  30 tasks      | elapsed:  4.3min\n",
      "[Parallel(n_jobs=-1)]: Done  31 tasks      | elapsed:  4.3min\n",
      "[Parallel(n_jobs=-1)]: Done  32 tasks      | elapsed:  4.3min\n",
      "[Parallel(n_jobs=-1)]: Done  33 tasks      | elapsed:  4.4min\n",
      "[Parallel(n_jobs=-1)]: Done  34 tasks      | elapsed:  4.4min\n",
      "[Parallel(n_jobs=-1)]: Done  35 tasks      | elapsed:  4.4min\n",
      "[Parallel(n_jobs=-1)]: Done  36 tasks      | elapsed:  4.5min\n",
      "[Parallel(n_jobs=-1)]: Done  37 tasks      | elapsed:  4.5min\n",
      "[Parallel(n_jobs=-1)]: Done  38 tasks      | elapsed:  5.0min\n",
      "[Parallel(n_jobs=-1)]: Done  39 tasks      | elapsed:  5.0min\n",
      "[Parallel(n_jobs=-1)]: Done  40 tasks      | elapsed:  5.1min\n",
      "[Parallel(n_jobs=-1)]: Done  41 tasks      | elapsed:  5.2min\n",
      "[Parallel(n_jobs=-1)]: Done  42 tasks      | elapsed:  5.5min\n",
      "[Parallel(n_jobs=-1)]: Done  43 tasks      | elapsed:  5.6min\n",
      "[Parallel(n_jobs=-1)]: Done  44 tasks      | elapsed:  5.7min\n",
      "[Parallel(n_jobs=-1)]: Done  45 tasks      | elapsed:  5.7min\n",
      "[Parallel(n_jobs=-1)]: Done  46 tasks      | elapsed:  5.7min\n",
      "[Parallel(n_jobs=-1)]: Done  47 tasks      | elapsed:  5.7min\n",
      "[Parallel(n_jobs=-1)]: Done  48 tasks      | elapsed:  5.8min\n",
      "[Parallel(n_jobs=-1)]: Done  49 tasks      | elapsed:  5.9min\n",
      "[Parallel(n_jobs=-1)]: Done  50 tasks      | elapsed:  6.0min\n",
      "[Parallel(n_jobs=-1)]: Done  51 tasks      | elapsed:  6.0min\n",
      "[Parallel(n_jobs=-1)]: Done  52 tasks      | elapsed:  6.4min\n",
      "[Parallel(n_jobs=-1)]: Done  53 tasks      | elapsed:  6.5min\n",
      "[Parallel(n_jobs=-1)]: Done  54 tasks      | elapsed:  6.5min\n",
      "[Parallel(n_jobs=-1)]: Done  55 tasks      | elapsed:  6.6min\n",
      "[Parallel(n_jobs=-1)]: Done  56 tasks      | elapsed:  6.7min\n",
      "[Parallel(n_jobs=-1)]: Done  57 tasks      | elapsed:  6.7min\n",
      "[Parallel(n_jobs=-1)]: Done  58 tasks      | elapsed:  6.7min\n",
      "[Parallel(n_jobs=-1)]: Done  59 tasks      | elapsed:  6.8min\n",
      "[Parallel(n_jobs=-1)]: Done  64 out of  90 | elapsed:  7.5min remaining:  3.0min\n",
      "[Parallel(n_jobs=-1)]: Done  69 out of  90 | elapsed:  7.9min remaining:  2.4min\n",
      "[Parallel(n_jobs=-1)]: Done  74 out of  90 | elapsed:  8.1min remaining:  1.7min\n",
      "[Parallel(n_jobs=-1)]: Done  79 out of  90 | elapsed:  8.9min remaining:  1.2min\n",
      "[Parallel(n_jobs=-1)]: Done  84 out of  90 | elapsed:  9.2min remaining:   39.5s\n",
      "[Parallel(n_jobs=-1)]: Done  90 out of  90 | elapsed: 10.9min finished\n",
      "C:\\ProgramData\\Anaconda3\\lib\\site-packages\\sklearn\\model_selection\\_search.py:841: DeprecationWarning: The default of the `iid` parameter will change from True to False in version 0.22 and will be removed in 0.24. This will change numeric results when test-set sizes are unequal.\n",
      "  DeprecationWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best score :  -0.3328970966606883\n",
      "Best Parameters {'n_estimators': 90, 'max_depth': 5, 'learning_rate': 0.2, 'gamma': 1.0}\n"
     ]
    }
   ],
   "source": [
    "import xgboost as xgb\n",
    "from sklearn.model_selection import RandomizedSearchCV\n",
    "\n",
    "params = {\n",
    "        'max_depth': [i for i in range(1,8,2)],\n",
    "        'n_estimators' : [i*10 for i in range(1, 10)],\n",
    "         'gamma': [0, 0.25, 0.5, 1.0],'learning_rate': [0.001, 0.01, 0.1, 0.2, 0,3]         \n",
    "        }\n",
    "xgboost_model = xgb.XGBClassifier(objective='binary:logistic', eval_metric='logloss',n_jobs=-1)\n",
    "\n",
    "xgb_search = RandomizedSearchCV(xgboost_model, param_distributions = params,n_iter=30,\n",
    "                         scoring = 'neg_log_loss', n_jobs = -1,cv=3, verbose=20)\n",
    "\n",
    "xgb_search.fit(X_train, y_train)\n",
    "\n",
    "print(\"Best score : \",xgb_search.best_score_)\n",
    "print(\"Best Parameters\",xgb_search.best_params_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 153,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Log loss for tfidf train data :  0.32958004235646765\n",
      "Log loss for tfidf test data :  0.33745821903961576\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x288 with 6 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "clf = xgb.XGBClassifier(max_depth=5,learning_rate=0.2,objective='binary:logistic',gamma=1.0,n_estimators=90,n_jobs=-1)\n",
    "clf.fit(X_train, y_train)   \n",
    "\n",
    "#log loss uses a exact prob score \n",
    "cal_clf = CalibratedClassifierCV(clf, method=\"sigmoid\")\n",
    "cal_clf.fit(X_train, y_train)\n",
    "\n",
    "predict_y = cal_clf.predict_proba(X_train)\n",
    "\n",
    "print(\"Log loss for tfidf train data : \",log_loss(y_train, predict_y,  eps=1e-15))\n",
    "predict_y = cal_clf.predict_proba(X_test)\n",
    "\n",
    "print(\"Log loss for tfidf test data : \",log_loss(y_test, predict_y,  eps=1e-15))\n",
    "predicted_y =np.argmax(predict_y,axis=1)\n",
    "plot_confusion_matrix(y_test, predicted_y)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Conclusion"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 156,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "+------------+----------------------+---------------------+---------------------+\n",
      "| Vectorizer |      Estimator       |    Train log loss   |    Test log loss    |\n",
      "+------------+----------------------+---------------------+---------------------+\n",
      "| TFIDF W2V  | Logistics Regression |  0.5101453297915229 |  0.5145708244250301 |\n",
      "| TFIDF W2V  |      Linear SVM      |  0.4815323702975848 |  0.4885874377203666 |\n",
      "| TFIDF W2V  |       XG Boost       |  0.3147937625931934 |  0.3572127318255224 |\n",
      "|   TFIDF    | Logistics Regression |  0.6591807615185924 |  0.6570136275973378 |\n",
      "|   TFIDF    |      Linear SVM      |  0.6416219868892176 |  0.6393926484757468 |\n",
      "|   TFIDF    |       XG Boost       | 0.32958004235646765 | 0.33745821903961576 |\n",
      "+------------+----------------------+---------------------+---------------------+\n"
     ]
    }
   ],
   "source": [
    "from prettytable import PrettyTable\n",
    "\n",
    "x = PrettyTable()\n",
    "x.field_names = [\"Vectorizer\", \"Estimator\", \"Train log loss\", \"Test log loss\"]\n",
    "x.add_row([\"TFIDF W2V\", \"Logistics Regression\", 0.5101453297915229, 0.5145708244250301])\n",
    "x.add_row([\"TFIDF W2V\", \"Linear SVM\", 0.4815323702975848, 0.4885874377203666])\n",
    "x.add_row([\"TFIDF W2V\", \"XG Boost\", 0.3147937625931934, 0.3572127318255224])\n",
    "x.add_row([\"TFIDF\", \"Logistics Regression\", 0.6591807615185924, 0.6570136275973378])\n",
    "x.add_row([\"TFIDF\", \"Linear SVM\", 0.6416219868892176, 0.6393926484757468])\n",
    "x.add_row([\"TFIDF\", \"XG Boost\", 0.32958004235646765, 0.33745821903961576])\n",
    "\n",
    "print(x)"
   ]
  }
 ],
 "metadata": {
  "colab": {
   "name": "4.ML_models.ipynb",
   "provenance": [],
   "version": "0.3.2"
  },
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.3"
  },
  "toc": {
   "base_numbering": 1,
   "nav_menu": {},
   "number_sections": true,
   "sideBar": true,
   "skip_h1_title": false,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": false,
   "toc_position": {},
   "toc_section_display": true,
   "toc_window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}