05 Amazon Fine Food Reviews Analysis_Logistic Regression.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "AM7jsfwALPlp"
   },
   "source": [
    "# Amazon Fine Food Reviews Analysis\n",
    "\n",
    "\n",
    "Data Source: https://www.kaggle.com/snap/amazon-fine-food-reviews <br>\n",
    "\n",
    "EDA: https://nycdatascience.com/blog/student-works/amazon-fine-foods-visualization/\n",
    "\n",
    "\n",
    "The Amazon Fine Food Reviews dataset consists of reviews of fine foods from Amazon.<br>\n",
    "\n",
    "Number of reviews: 568,454<br>\n",
    "Number of users: 256,059<br>\n",
    "Number of products: 74,258<br>\n",
    "Timespan: Oct 1999 - Oct 2012<br>\n",
    "Number of Attributes/Columns in data: 10 \n",
    "\n",
    "Attribute Information:\n",
    "\n",
    "1. Id\n",
    "2. ProductId - unique identifier for the product\n",
    "3. UserId - unqiue identifier for the user\n",
    "4. ProfileName\n",
    "5. HelpfulnessNumerator - number of users who found the review helpful\n",
    "6. HelpfulnessDenominator - number of users who indicated whether they found the review helpful or not\n",
    "7. Score - rating between 1 and 5\n",
    "8. Time - timestamp for the review\n",
    "9. Summary - brief summary of the review\n",
    "10. Text - text of the review\n",
    "\n",
    "\n",
    "#### Objective:\n",
    "Given a review, determine whether the review is positive (rating of 4 or 5) or negative (rating of 1 or 2).\n",
    "\n",
    "<br>\n",
    "[Q] How to determine if a review is positive or negative?<br>\n",
    "<br> \n",
    "[Ans] We could use Score/Rating. A rating of 4 or 5 can be cosnidered as a positive review. A rating of 1 or 2 can be considered as negative one. A review of rating 3 is considered nuetral and such reviews are ignored from our analysis. This is an approximate and proxy way of determining the polarity (positivity/negativity) of a review.\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "CaQ7XJ4OLPls"
   },
   "source": [
    "# [1]. Reading Data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "5VnEBjKRLPlu"
   },
   "source": [
    "## [1.1] Loading the data\n",
    "\n",
    "The dataset is available in two forms\n",
    "1. .csv file\n",
    "2. SQLite Database\n",
    "\n",
    "In order to load the data, We have used the SQLITE dataset as it is easier to query the data and visualise the data efficiently.\n",
    "<br> \n",
    "\n",
    "Here as we only want to get the global sentiment of the recommendations (positive or negative), we will purposefully ignore all Scores equal to 3. If the score is above 3, then the recommendation wil be set to \"positive\". Otherwise, it will be set to \"negative\"."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "collapsed": true,
    "id": "x6YclOzXLPlv"
   },
   "outputs": [],
   "source": [
    "%matplotlib inline\n",
    "import warnings\n",
    "warnings.filterwarnings(\"ignore\")\n",
    "\n",
    "\n",
    "import sqlite3\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import nltk\n",
    "import string\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "from sklearn.feature_extraction.text import TfidfTransformer\n",
    "from sklearn.feature_extraction.text import TfidfVectorizer\n",
    "\n",
    "from sklearn.feature_extraction.text import CountVectorizer\n",
    "from sklearn.metrics import confusion_matrix\n",
    "from sklearn import metrics\n",
    "from sklearn.metrics import roc_curve, auc, roc_auc_score\n",
    "from nltk.stem.porter import PorterStemmer\n",
    "\n",
    "import re\n",
    "# Tutorial about Python regular expressions: https://pymotw.com/2/re/\n",
    "import string\n",
    "from nltk.corpus import stopwords\n",
    "from nltk.stem import PorterStemmer\n",
    "from nltk.stem.wordnet import WordNetLemmatizer\n",
    "\n",
    "from gensim.models import Word2Vec\n",
    "from gensim.models import KeyedVectors\n",
    "import pickle\n",
    "\n",
    "from tqdm import tqdm\n",
    "import os"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "Kxx72u1hLPo1"
   },
   "source": [
    "# [5] Assignment 5: Apply Logistic Regression"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "KDR9We0oLPo2"
   },
   "source": [
    "<ol>\n",
    "    <li><strong>Apply Logistic Regression on these feature sets</strong>\n",
    "        <ul>\n",
    "            <li><font color='red'>SET 1:</font>Review text, preprocessed one converted into vectors using (BOW)</li>\n",
    "            <li><font color='red'>SET 2:</font>Review text, preprocessed one converted into vectors using (TFIDF)</li>\n",
    "            <li><font color='red'>SET 3:</font>Review text, preprocessed one converted into vectors using (AVG W2v)</li>\n",
    "            <li><font color='red'>SET 4:</font>Review text, preprocessed one converted into vectors using (TFIDF W2v)</li>\n",
    "        </ul>\n",
    "    </li>\n",
    "    <br>\n",
    "    <li><strong>Hyper paramter tuning (find best hyper parameters corresponding the algorithm that you choose)</strong>\n",
    "        <ul>\n",
    "    <li>Find the best hyper parameter which will give the maximum <a href='https://www.appliedaicourse.com/course/applied-ai-course-online/lessons/receiver-operating-characteristic-curve-roc-curve-and-auc-1/'>AUC</a> value</li>\n",
    "    <li>Find the best hyper paramter using k-fold cross validation or simple cross validation data</li>\n",
    "    <li>Use gridsearch cv or randomsearch cv or you can also write your own for loops to do this task of hyperparameter tuning</li>          \n",
    "        </ul>\n",
    "    </li>\n",
    "    <br>\n",
    "    <li><strong>Pertubation Test</strong>\n",
    "        <ul>\n",
    "    <li>Get the weights W after fit your model with the data X i.e Train data.</li>\n",
    "    <li>Add a noise to the X (X' = X + e) and get the new data set X' (if X is a sparse\n",
    "matrix, X.data+=e)</li>\n",
    "    <li>Fit the model again on data X' and get the weights W'</li>\n",
    "    <li>Add a small eps value(to eliminate the divisible by zero error) to W and W’ i.e\n",
    "W=W+10^-6 and W’ = W’+10^-6</li>\n",
    "    <li>Now find the % change between W and W' (| (W-W') / (W) |)*100)</li>\n",
    "    <li>Calculate the 0th, 10th, 20th, 30th, ...100th percentiles, and observe any sudden rise in the values of percentage_change_vector</li>\n",
    "    <li> Ex: consider your 99th percentile is 1.3 and your 100th percentiles are 34.6, there is sudden rise from 1.3 to 34.6, now calculate the 99.1, 99.2, 99.3,..., 100th percentile values and get the proper value after which there is sudden rise the values, assume it is 2.5</li>\n",
    "            <li> Print the feature names whose % change is more than a threshold x(in our example it's 2.5)</li>\n",
    "        </ul>\n",
    "    </li>\n",
    "    <br>\n",
    "    <li><strong>Sparsity</strong>\n",
    "        <ul>\n",
    "    <li>Calculate sparsity on weight vector obtained after using L1 regularization</li>\n",
    "        </ul>\n",
    "    </li>\n",
    "    <br><font color='red'>NOTE: Do sparsity and multicollinearity for any one of the vectorizers. Bow or tf-idf is recommended.</font>\n",
    "    <br>\n",
    "    <br>\n",
    "    <li><strong>Feature importance</strong>\n",
    "        <ul>\n",
    "    <li>Get top 10 important features for both positive and negative classes separately.</li>\n",
    "        </ul>\n",
    "    </li>\n",
    "    <br>\n",
    "    <li><strong>Feature engineering</strong>\n",
    "        <ul>\n",
    "    <li>To increase the performance of your model, you can also experiment with with feature engineering like :</li>\n",
    "            <ul>\n",
    "            <li>Taking length of reviews as another feature.</li>\n",
    "            <li>Considering some features from review summary as well.</li>\n",
    "        </ul>\n",
    "        </ul>\n",
    "    </li>\n",
    "    <br>\n",
    "    <li><strong>Representation of results</strong>\n",
    "        <ul>\n",
    "    <li>You need to plot the performance of model both on train data and cross validation data for each hyper parameter, like shown in the figure.\n",
    "    <img src='train_cv_auc.JPG' width=300px></li>\n",
    "    <li>Once after you found the best hyper parameter, you need to train your model with it, and find the AUC on test data and plot the ROC curve on both train and test.\n",
    "    <img src='train_test_auc.JPG' width=300px></li>\n",
    "    <li>Along with plotting ROC curve, you need to print the <a href='https://www.appliedaicourse.com/course/applied-ai-course-online/lessons/confusion-matrix-tpr-fpr-fnr-tnr-1/'>confusion matrix</a> with predicted and original labels of test data points. Please visualize your confusion matrices using <a href='https://seaborn.pydata.org/generated/seaborn.heatmap.html'>seaborn heatmaps.\n",
    "    <img src='confusion_matrix.png' width=300px></li>\n",
    "        </ul>\n",
    "    </li>\n",
    "    <br>\n",
    "    <li><strong>Conclusion</strong>\n",
    "        <ul>\n",
    "    <li>You need to summarize the results at the end of the notebook, summarize it in the table format. To print out a table please refer to this prettytable library<a href='http://zetcode.com/python/prettytable/'>  link</a> \n",
    "        <img src='summary.JPG' width=400px>\n",
    "    </li>\n",
    "        </ul>\n",
    "</ol>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "gilb-DOVLPo2"
   },
   "source": [
    "<h4><font color='red'>Note: Data Leakage</font></h4>\n",
    "\n",
    "1. There will be an issue of data-leakage if you vectorize the entire data and then split it into train/cv/test.\n",
    "2. To avoid the issue of data-leakag, make sure to split your data first and then vectorize it. \n",
    "3. While vectorizing your data, apply the method fit_transform() on you train data, and apply the method transform() on cv/test data.\n",
    "4. For more details please go through this <a href='https://soundcloud.com/applied-ai-course/leakage-bow-and-tfidf'>link.</a>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Reading the csv file (Which is already saved on my system (saved while doing naive bayes assignment))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "review_dataset_filtered = pd.read_csv('filtered_reviews_data.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(364173, 12)"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "review_dataset_filtered.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#removing all the null value represnted rows (removing rows even if one column has na in entire 12 columns in given dataset)\n",
    "review_dataset_filtered.dropna(inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(364158, 12)"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "review_dataset_filtered.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "### As it is temporal data, I will be performing time-based splitting"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#sorting the dataset according to time as we need time-based splitting\n",
    "review_dataset_final_sorted = review_dataset_filtered.sort_values(by ='Time',axis=0,ascending=True,inplace=False,kind='quicksort')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "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>Unnamed: 0</th>\n",
       "      <th>Id</th>\n",
       "      <th>ProductId</th>\n",
       "      <th>UserId</th>\n",
       "      <th>ProfileName</th>\n",
       "      <th>HelpfulnessNumerator</th>\n",
       "      <th>HelpfulnessDenominator</th>\n",
       "      <th>Score</th>\n",
       "      <th>Time</th>\n",
       "      <th>Summary</th>\n",
       "      <th>Text</th>\n",
       "      <th>Cleaned_text</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>117926</th>\n",
       "      <td>150523</td>\n",
       "      <td>150524</td>\n",
       "      <td>0006641040</td>\n",
       "      <td>ACITT7DI6IDDL</td>\n",
       "      <td>shari zychinski</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>positive</td>\n",
       "      <td>939340800</td>\n",
       "      <td>EVERY book is educational</td>\n",
       "      <td>this witty little book makes my son laugh at l...</td>\n",
       "      <td>witti littl book make son laugh loud recit car...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>117903</th>\n",
       "      <td>150500</td>\n",
       "      <td>150501</td>\n",
       "      <td>0006641040</td>\n",
       "      <td>AJ46FKXOVC7NR</td>\n",
       "      <td>Nicholas A Mesiano</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>positive</td>\n",
       "      <td>940809600</td>\n",
       "      <td>This whole series is great way to spend time w...</td>\n",
       "      <td>I can remember seeing the show when it aired o...</td>\n",
       "      <td>rememb see show air televis year ago child sis...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>298794</th>\n",
       "      <td>451855</td>\n",
       "      <td>451856</td>\n",
       "      <td>B00004CXX9</td>\n",
       "      <td>AIUWLEQ1ADEG5</td>\n",
       "      <td>Elizabeth Medina</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>positive</td>\n",
       "      <td>944092800</td>\n",
       "      <td>Entertainingl Funny!</td>\n",
       "      <td>Beetlejuice is a well written movie ..... ever...</td>\n",
       "      <td>beetlejuic well written movi everyth excel act...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>169283</th>\n",
       "      <td>230284</td>\n",
       "      <td>230285</td>\n",
       "      <td>B00004RYGX</td>\n",
       "      <td>A344SMIA5JECGM</td>\n",
       "      <td>Vincent P. Ross</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>positive</td>\n",
       "      <td>944438400</td>\n",
       "      <td>A modern day fairy tale</td>\n",
       "      <td>A twist of rumplestiskin captured on film, sta...</td>\n",
       "      <td>twist rumplestiskin captur film star michael k...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>298793</th>\n",
       "      <td>451854</td>\n",
       "      <td>451855</td>\n",
       "      <td>B00004CXX9</td>\n",
       "      <td>AJH6LUC1UT1ON</td>\n",
       "      <td>The Phantom of the Opera</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>positive</td>\n",
       "      <td>946857600</td>\n",
       "      <td>FANTASTIC!</td>\n",
       "      <td>Beetlejuice is an excellent and funny movie. K...</td>\n",
       "      <td>beetlejuic excel funni movi keaton hilari wack...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        Unnamed: 0      Id   ProductId          UserId  \\\n",
       "117926      150523  150524  0006641040   ACITT7DI6IDDL   \n",
       "117903      150500  150501  0006641040   AJ46FKXOVC7NR   \n",
       "298794      451855  451856  B00004CXX9   AIUWLEQ1ADEG5   \n",
       "169283      230284  230285  B00004RYGX  A344SMIA5JECGM   \n",
       "298793      451854  451855  B00004CXX9   AJH6LUC1UT1ON   \n",
       "\n",
       "                     ProfileName  HelpfulnessNumerator  \\\n",
       "117926           shari zychinski                     0   \n",
       "117903        Nicholas A Mesiano                     2   \n",
       "298794          Elizabeth Medina                     0   \n",
       "169283           Vincent P. Ross                     1   \n",
       "298793  The Phantom of the Opera                     0   \n",
       "\n",
       "        HelpfulnessDenominator     Score       Time  \\\n",
       "117926                       0  positive  939340800   \n",
       "117903                       2  positive  940809600   \n",
       "298794                       0  positive  944092800   \n",
       "169283                       2  positive  944438400   \n",
       "298793                       0  positive  946857600   \n",
       "\n",
       "                                                  Summary  \\\n",
       "117926                          EVERY book is educational   \n",
       "117903  This whole series is great way to spend time w...   \n",
       "298794                               Entertainingl Funny!   \n",
       "169283                            A modern day fairy tale   \n",
       "298793                                         FANTASTIC!   \n",
       "\n",
       "                                                     Text  \\\n",
       "117926  this witty little book makes my son laugh at l...   \n",
       "117903  I can remember seeing the show when it aired o...   \n",
       "298794  Beetlejuice is a well written movie ..... ever...   \n",
       "169283  A twist of rumplestiskin captured on film, sta...   \n",
       "298793  Beetlejuice is an excellent and funny movie. K...   \n",
       "\n",
       "                                             Cleaned_text  \n",
       "117926  witti littl book make son laugh loud recit car...  \n",
       "117903  rememb see show air televis year ago child sis...  \n",
       "298794  beetlejuic well written movi everyth excel act...  \n",
       "169283  twist rumplestiskin captur film star michael k...  \n",
       "298793  beetlejuic excel funni movi keaton hilari wack...  "
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "review_dataset_final_sorted.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "labels = review_dataset_final_sorted['Score']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "LABELS = labels.value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAgIAAAFlCAYAAACZXICzAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<Figure size 576x396 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.barplot(LABELS.index,LABELS.values,alpha=0.8)\n",
    "plt.title(\"distribution of positive and negative classes in original dataset\")\n",
    "plt.xlabel(\"Positive  Negative\")\n",
    "plt.ylabel(\"Count\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "POSITIVE class labels in original data is 84.32 %\n",
      "NEGATIVE class labels in original data is 15.68 %\n"
     ]
    }
   ],
   "source": [
    "print(f\"POSITIVE class labels in original data is {round(((labels.value_counts()[0]/review_dataset_final_sorted.shape[0])*100),2)} %\")\n",
    "print(f\"NEGATIVE class labels in original data is {round(((labels.value_counts()[1]/review_dataset_final_sorted.shape[0])*100),2)} %\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#Now, let's take only 100k datapoints for computation because 300K+ would not be ideal to choose among\n",
    "review_dataset_final_sorted_sample = review_dataset_final_sorted.sample(n=100000)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#Now, let's make a check distribution of data or percentage of data present\n",
    "sample_labels = review_dataset_final_sorted_sample['Score'] "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "positive    84215\n",
       "negative    15785\n",
       "Name: Score, dtype: int64"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sample_labels.value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAgcAAAFlCAYAAAB/dUv3AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi4yLCBo\ndHRwOi8vbWF0cGxvdGxpYi5vcmcvhp/UCwAAIABJREFUeJzt3Xl8TPf+x/H3ZBMkSmzlqhZVW+xc\ntVdRa8S+tpSqvYpSSqpBxL4rv2pVXdRSIqqUKkUttda+loqlKpaQhJJk5vz+cM31JSFaSYjX8/Hw\neJgz3znn8zlnlveccybHZlmWJQAAgP9ySekCAADAk4VwAAAADIQDAABgIBwAAAAD4QAAABgIBwAA\nwEA4gCRp1apVeuuttyRJkyZNUmho6APHT506VT/++GO89939+AIFCujKlSuPVMu+ffs0ePBgSdL+\n/fvVs2fPR3r832G329W1a1fVqlVLc+fOTbLlrF27VkFBQZKk9evXa9KkSfdNfxLMnDlTAwYMSNEa\nknv9DBgwQDNnzkzSZTyqpH7++/v7KzIy8oFjBg0apC1btvztZbz++uvav3//Q8eVLFlSZ8+efeCY\nu98bHrdvvvlG8+bNS5J5P43cUroAPHnef//9h47Ztm2bXn755b/9+Af57bffdOHCBUlS0aJFNXny\n5H80v8S4cOGCNm3apD179sjV1TXJllO9enVVr15d0u03/mvXrt03HbexfpL++b9s2bKHjhk+fHiS\nLf9R3f3e8Ljt2rVL+fPnT5J5P40IB8+wSZMmafny5cqYMaNefPFF5/QBAwYof/78eueddzR58mSt\nWbNG7u7uypQpk0aMGKE1a9bowIEDGj16tFxdXbV27VpdvXpVZ86c0WuvvabLly87Hy9JEydO1P79\n++VwONSrVy9Vq1ZNISEhWr16tT777DNJct4ODAzU5MmTFRUVpY8++kgNGzbUsGHD9N133ykqKkpD\nhgzRkSNHZLPZVLlyZfXp00dubm4qWrSoOnXqpM2bNys8PFwdO3ZU69at7+t5586dGj16tP766y+5\nu7urV69eKlWqlDp27Ki4uDg1btxYU6ZMUe7cuY31kSZNGh05ckSXL19WxYoVFRAQIHd393jnV6VK\nFV28eFH9+/dXRESEJKlq1arq1auXs89u3bppwYIFstvt8vb21osvvqjVq1drwIABatmypX7++Wd5\neHjIbrfrtdde01dffaVs2bJp+PDhOnbsmGJjY1W+fHl9+OGHcnMzX8a///67hg4dquvXr+vixYsq\nWLCgJk6cqDRp0iS4nmJjYxUUFKQtW7Yoc+bMypw5s7y9ve9bfyEhIVqzZo1cXFwUFhYmT09PjRo1\nSvny5VNUVFSC9W3YsEFjx46Vi4uLChUqpC1btujrr7+Wj4+PAgMDFRYWpqtXryp9+vQaO3asoqKi\nkmz9XL9+XUFBQdq9e7dcXV1Vo0YN9e7d2xizePFiLVy4ULGxsbp27ZreffddtW7dOsHtmtB06fY3\n0vnz58vhcChjxoz6+OOPlS9fPu3cuVMjR46Uw+GQJHXu3Fm1atUy6ti2bZvz+T9gwAB5eXnp6NGj\n+vPPP1WgQAGNGjVK6dOnNx7zoNeJr6+vqlevriNHjmjs2LFq2rSptm7dqueee06jR4/WunXr5O3t\nrWLFiunEiROaM2eO3nrrLbVp00a+vr56++23VbVqVe3du1eRkZHq16+fatasqUuXLmnw4MG6fPmy\nLl68qH/961+aOHGiMmfOfN9z6O7X4rBhw2Sz2VS0aFHnenA4HAoODtbevXt1/fp1WZaloKAg5cyZ\n03hvGD58eLzjSpcuneC6jYmJ0dixY7Vjxw7Z7XYVLlxYAQEB2rp1q9atW6fNmzfL09NTbdq0SbDu\nZwWHFZ5RP/74o3744QeFhoZqwYIFio6Ovm/M+fPnNXv2bC1ZskQhISGqWLGi9u3b53yj+PDDD1Wz\nZk1J0s2bN7VixQr169fvvvnkypVLS5cu1ZgxYzRgwIAHHmbIkSOHevbsqTJlymjEiBHGfUFBQcqY\nMaOWL1+uJUuW6OjRo/ryyy8lSTExMcqUKZMWLFigyZMna8SIEbp165bx+IiICPXs2VODBg3S8uXL\nNWrUKPXr108RERGaMWOGPD09tWzZMiMY3LFv3z59+eWXWrlypU6cOKGFCxcmOL8zZ85o0aJFzr7n\nzZunsLAwRUVFOedXvHhxtWzZUnXr1jU+mPLkyaP8+fNr3bp1kqRNmzYpV65cypcvn4KDg1WkSBGF\nhIQoNDRUERERmjVr1n21Llq0SA0bNtSiRYv0ww8/6OzZs1q/fv0D19PXX3+tU6dOacWKFfryyy91\n/vz5BLfRjh079PHHH+u7775T8eLFNWPGDElKsL6IiAh9+OGHGjNmjJYtW6Zy5co5v/1t3LhRGTJk\n0MKFC7V69Wr5+vpq3rx5Sbp+Jk+erFu3bmnlypUKDQ3V7t27tX37duf9169f1zfffKMZM2YoNDRU\nEyZM0JgxY5zrNr7tmtD07du3KzQ0VPPmzVNoaKg6duyoHj16SJKmTJmi9u3bKyQkRMHBwfrll18S\nXOd3HDhwQDNnztTKlSt17tw5rVq16r4xD3qdxMbGqlq1alq9erWKFi3qfMw333yjgwcP6rvvvtOC\nBQt05syZeJd/5swZVapUSYsXL9YHH3yg4OBgSdKKFStUokQJLVy4UGvXrnW+lhISExOj999/XwMG\nDFBoaKjKlSunmzdvSpL27t2r8PBwLVy4UCtXrlSjRo30+eef3/fekNC4B63bGTNmyNXVVSEhIfr2\n22+VLVs2jR07VjVr1tTrr7+ut99+m2DwX+w5eEZt3bpVNWvWlJeXlySpSZMmmjNnjjEme/bsKliw\noBo1aqQqVaqoSpUqKl++fLzzK126dILLatWqlSTplVdeUb58+fTrr7/+rZo3btyo+fPny2azycPD\nQy1bttTs2bPVqVMnSXLudi5SpIhiYmJ048YNpUmTxvn4ffv2KXfu3CpevLgkKX/+/CpVqpS2b9+u\ncuXKPXDZjRo1cn5D8/f319q1a/XCCy8kOL/KlSurU6dOOn/+vCpUqKAPPvgg3m/i8WnatKmWLl2q\n2rVrKyQkRM2bN5d0+xj8/v37tXjxYklyvpneq1+/ftq8ebM+//xznTp1SuHh4bpx44bz/vjW09at\nW1W/fn15eHjIw8NDfn5+Onr0aLzzL1KkiJ5//nlJUuHChbVmzZoH1rdz507ly5dPBQsWdK7LO+cP\n1K5dWy+88ILmzJmjsLAwbd++XSVLlkzS9bNlyxZ99NFHcnV1laurq/Mck6VLl0qS0qdPr//7v//T\nhg0bdOrUKR05csS5/hLarglNX79+vcLCwtSyZUvn8iMjI3X16lXVqVNHQ4cO1bp161ShQgX16dPn\ngX3fWb6Hh4ek26+nO4dd7vaw10mZMmXue8yGDRvk7+/vfL20aNHivvcDSXJ3d1fVqlUl3d72V69e\nlSS1a9dOO3fu1KxZs3Tq1CkdP37c+bqIz7Fjx+Tm5uZ8P6lfv77zXIKSJUvqueeec4aUbdu23bd3\n5GHjElq369evV1RUlPMcitjY2Afu3XiWEQ6eYXdfViO+4+wuLi6aO3eu9u/fr61btyo4OFiVK1fW\nhx9+eN/YdOnSJbgcF5f/7aByOBxyc3OTzWYzlh8bG/vQeh0Oh2w2m3E7Li7OefvOG9udMfdeNsRu\ntxuPvzPm7nkk5O71Y1mWXFxcHji/YsWKae3atdq6dat++eUXNWvWzPmt5mHq1KmjkSNH6sSJE9qx\nY4dGjhzp7HfSpEnKly+fpNsfMvcuX5L69Okju92uOnXq6LXXXtP58+eNdfGw9XRvv/fy9PR0/v/u\n7ZhQfTt27LhvGXeeE19//bUWLVqkNm3ayM/PTxkzZnzoSWn/dP3cef7dcf78eaOnP//8Uy1atFDz\n5s1VunRp1a5dWz/99JMkJbhdE5rucDjk7+/v3KPmcDgUHh6u5557Ti1btlS1atW0efNm/fzzz5o6\ndapWrVplBNrErvu7Pex1Et9r9d5DL3e/Zu/m7u7uvO/uZYwZM0b79u1TkyZNVK5cOcXFxcVb293u\nvf9ODevXr9fw4cPVvn17Va9eXXnz5tW333573+MfNC6hdetwODRw4EBnwLl+/fp9exhxG4cVnlFV\nqlTRqlWrFBkZKYfDEe8uwCNHjqh+/frKly+fOnfurLffftt51rGrq2uiPlSl/30jO3jwoE6fPq3i\nxYvLx8dHx48f161btxQbG6vVq1c7xyc070qVKmnu3LmyLEsxMTFatGiRKlSokOieS5QooZMnT2rf\nvn2SpOPHj2vHjh3697///dDHfv/994qJidGtW7e0dOlSVatW7YHzGzt2rKZNm6YaNWpo0KBBevnl\nl3X8+HFjngn1mSZNGtWrV08DBgzQG2+8obRp0zr7/+qrr5z9d+3aNd5fVmzatEndu3dX3bp1Jd3e\nTWu32x/YX+XKlRUaGqpbt245d7k/qoTqK1WqlPMbuCStXr3a+cG9adMmNWrUSM2aNVOePHm0bt06\nZ61JtX7Kly+vpUuXyuFwKCYmRj179tSOHTuc9x84cEA+Pj7q1q2bKlWq5AwGdrs9we2a0PRKlSpp\nxYoVCg8PlyTNnz9f7dq1k3T7A+zw4cNq3Lixhg0bpsjISF28ePGR13t82+FRXydVq1bVt99+q5iY\nGMXFxTlfs4m1adMmtWvXTg0bNlTmzJm1ZcuWBz7nChQoIMuytGHDBkm3f41yZy/I5s2bVa1aNbVu\n3Vq+vr768ccf431OPGhcQuu2UqVKmjdvnmJiYuRwOPTxxx9r/Pjx980b7Dl4ZlWtWlVHjx5VkyZN\nlCFDBhUsWNB5MtUdBQsWVJ06ddSkSROlS5dOnp6eCggIkHT750njx49P1Df+M2fOqGHDhrLZbBo/\nfrwyZsyoihUrqmzZsqpTp46yZs2qcuXKOXdjlyhRQp9++ql69Ojh/HmlJAUEBCgoKEh+fn6KjY1V\n5cqV1aVLl0T37OPjo0mTJmnYsGG6efOmbDabRowYoTx58jz026qnp6dat26tyMhI1apVS02aNJGL\ni0uC82vXrp0GDBjg3FVfoEAB1atXT999951znq+++qr69u2rYcOGqUiRIsbymjVrprlz5yowMNA5\nbdCgQRo+fLiz/woVKqhjx4731dq7d291795d6dKlk5eXl8qWLavTp08/sL+WLVvq9OnTql+//n0n\nqCZWQvW5u7tr/Pjx6t+/v1xcXOTr6ys3NzelTZtWHTp00ODBg52HAkqUKKFjx44l6frp0aOHhg8f\nLn9/f9ntdtWtW1dvvPGG8zyGihUravHixapdu7ZsNpv+/e9/y8fHR2FhYQlu12vXrsU73cPDQ+++\n+646dOggm80mLy8vTZ06VTabTX379lVwcLAmTpwom82mHj16KFeuXI+83u/1d14njRs31u+//66G\nDRsqXbp0ypUrlzN0JUb37t01evRoTZo0Se7u7ipVqtQDn3Pu7u769NNPFRgYqPHjx6tQoULO3fst\nW7bUBx98ID8/P8XFxalixYr64Ycf5HA4jPeG3r17JzguoXXbrVs3jRo1So0aNZLdblehQoWcP9mt\nUqWKcy9U586dE917amXjks3Ag9396w08uujoaE2bNk3vvfee0qZNq4MHD6pz5876+eef493tj+S3\nadMmXb58Wf7+/pJun9SYJk2aeE8wxrOBPQcAkpSXl5fc3d3VtGlTubm5yc3NzfmNDk+G/Pnza+bM\nmfriiy/kcDhUsGBBY68Mnj3sOQAAAAZOSAQAAAbCAQAAMBAOAACAgRMS/ysuzq6IiBsPH/iUy5Qp\nHX2mIvSZutBn6vI09Jk1a/x/uZU9B//l5pZ0V+J7ktBn6kKfqQt9pi5Pc5+EAwAAYCAcAAAAA+EA\nAAAYCAcAAMBAOAAAAAbCAQAAMBAOAACAgXAAAAAMhAMAAGAgHAAAAAPhAAAAGAgHAADAQDgAAAAG\nLtn8X4Mmfa+4OHtKl5Hk3Nxc6TMVoc/UhT5Tl8fVZ9+2VR9DNY+GPQcAAMBAOAAAAAbCAQAAMBAO\nAACAgXAAAAAMhAMAAGAgHAAAAAPhAAAAGAgHAADAQDgAAAAGwgEAADAQDgAAgIFwAAAADIQDAABg\nIBwAAAAD4QAAABgIBwAAwEA4AAAABsIBAAAwEA4AAICBcAAAAAyEAwAAYCAcAAAAA+EAAAAYCAcA\nAMBAOAAAAAbCAQAAMBAOAACAgXAAAAAMhAMAAGAgHAAAAAPhAAAAGAgHAADAQDgAAAAGwgEAADAQ\nDgAAgIFwAAAADIQDAABgIBwAAAAD4QAAABgIBwAAwEA4AAAABsIBAAAwEA4AAICBcAAAAAyEAwAA\nYCAcAAAAA+EAAAAYCAcAAMBAOAAAAAbCAQAAMBAOAACAgXAAAAAMhAMAAGAgHAAAAMNTEQ4uXryo\nwMBASdKOHTt05MgRSVKPHj1SsCoAAFKnpyIcZM2a1RkOlixZovDwcEnS1KlTU7AqAABSJ7fkWlBI\nSIjWrl2r6OhoRUREqHv37vLy8tLEiROVJk0aZcyYUcHBwYqLi1OvXr1kWZZiY2M1ZMgQpU+fXn36\n9NHgwYP1888/6+DBg3r55ZfVrFkzLV++XG3atNHKlStls9k0ZMgQVahQQblz51ZQUJAkOeft7e2d\nXO0CAPCP2ONiZDnsioqKlLu7uzw90ybbspMtHEjSjRs3NGvWLF25ckXNmjWTzWbT/PnzlT17ds2e\nPVvTp09XuXLl5O3trXHjxum3335TdHS00qdPL0ny9fVV5cqVVbduXeXMmVOS5OPjowIFCmjnzp0q\nXry4tm/frkGDBql169YKDg7Wyy+/rG+++UZffPGFevfunZztAgDwt4QdXKeLYXslSe/++H+y2VxU\ns2ZtdejQKVmWn6zhoGzZsnJxcVGWLFmULl06xcXFKXv27M77xo8fr379+unUqVPq1q2b3Nzc1LVr\n14fOt3nz5lq6dKkuXryo119/XW5ubjpx4oSGDBkiSYqNjVWePHmStDcAAB6XO8HgDsty6IcfVqbO\ncHDw4EFJ0qVLl/TXX39JksLDw5UtWzZt375dL730krZt26Zs2bLpyy+/1K+//qrx48drxIgRznnY\nbDZZlmXMt3z58hozZowuXLigwYMHS5Ly5MmjUaNGKWfOnNq1a5cuXryYTF0CAPDPZH2xuC6G7ZN0\n+/Puzp6D5JKs4eDSpUtq166doqKiFBgYKDc3N7333nuy2Wx67rnnNGLECNlsNvXu3VuzZ8+Wi4uL\nunfvbsyjePHiGjt2rHLlyuWcZrPZVKtWLW3ZskUvvviiJCkwMFD9+/eX3W6XJA0fPjz5GgUA4B94\nscjrylWgkiyHXd1bVEj2cw5s1r1fw5NISEiITp48qb59+ybH4h7ZoEnfKy7OntJlJDk3N1f6TEXo\nM3Whz9TlcfXZt23Vx1BN/LJmjf9E/afip4wAACD5JNthhcaNGyfXogAAwD/AngMAAGAgHAAAAAPh\nAAAAGAgHAADAQDgAAAAGwgEAADAQDgAAgIFwAAAADIQDAABgIBwAAAAD4QAAABgIBwAAwEA4AAAA\nBsIBAAAwEA4AAICBcAAAAAyEAwAAYCAcAAAAA+EAAAAYCAcAAMBAOAAAAAbCAQAAMBAOAACAgXAA\nAAAMhAMAAGAgHAAAAAPhAAAAGAgHAADAQDgAAAAGwgEAADAQDgAAgIFwAAAADIQDAABgIBwAAAAD\n4QAAABgIBwAAwEA4AAAABsIBAAAwEA4AAICBcAAAAAyEAwAAYCAcAAAAA+EAAAAYCAcAAMBAOAAA\nAAbCAQAAMBAOAACAgXAAAAAMhAMAAGAgHAAAAAPhAAAAGAgHAADAQDgAAAAGm2VZVkoX8aS4eDEq\npUtIclmzetNnKkKfqQt9pi5PQ59Zs3rHO509BwAAwJCocHD8+PH7pu3Zs+exFwMAAFKe24Pu3LVr\nlxwOhwICAjR8+HDdOQIRFxenwMBArV69OlmKBAAAyeeB4WDLli3avn27wsPDNWnSpP89yM1NLVq0\nSPLiAABA8ntgOHjvvfckSaGhoWrYsGGyFAQAAFLWA8PBHWXLltWoUaN07do13f3jhhEjRiRZYQAA\nIGUkKhz06tVLZcqUUZkyZWSz2ZK6JgAAkIISFQ7i4uLUv3//pK4FAAA8ARL1U8bSpUtr3bp1iomJ\nSep6AABACkvUnoNVq1Zp7ty5xjSbzabDhw8nSVEAACDlJCocbNq0KanrAAAAT4hEhYOpU6fGO71H\njx6PtRgAAJDyHvnaCrGxsVq3bp0uX76cFPUAAIAUlqg9B/fuIejevbs6dOiQJAUBAICU9beuynj9\n+nX98ccfj7sWAADwBEjUnoPXX3/d+cePLMvStWvX1LFjxyQtDAAApIxEhYM5c+Y4/2+z2ZQhQwZ5\neXklWVEAACDlJCoc5MyZU/Pnz9cvv/yiuLg4vfrqq3rzzTfl4vK3jkoAAIAnWKLCwejRoxUWFqYm\nTZrIsiyFhITo9OnTCggISOr6AABAMktUONi8ebNCQ0Odewpee+01+fn5JWlhAAAgZSTquIDdbldc\nXJxx29XVNcmKAgAAKSdRew78/PzUtm1b1atXT5K0YsUK1a9fP0kLAwAAKeOh4eDatWtq3ry5Chcu\nrK1bt2rbtm1q27atGjZsmBz1AQCAZPbAwwqHDh1SvXr1dODAAVWpUkX9+/dXpUqVNG7cOB05ciS5\nagQAAMnogeFg1KhRGjdunKpUqeKc1qdPHwUHB2vkyJFJXhwAAEh+DwwHkZGRKleu3H3TK1eurIiI\niCQrCgAApJwHhoO4uDg5HI77pjscDsXGxiZZUQAAIOU8MByULVtWU6dOvW/6tGnT5Ovrm2RFAQCA\nlPPAXyv06dNHnTp1UmhoqAoWLKg0adLo0KFD8vHx0fTp05OrRgAAkIxslmVZDxpgWZZ++eUXHT58\nWC4uLvL19VWZMmWSq75kE/jtWMXF2VO6jCTn5uZKn6kIfaYu9Jm6/N0+36/cLQmqiV/WrN7xTn/o\n3zmw2WwqX768ypcv/9iLAgAATx4uqwgAAAyEAwAAYCAcAAAAA+EAAAAYCAcAAMBAOAAAAAbCAQAA\nMBAOAACAgXAAAAAMhAMAAGAgHAAAAAPhAAAAGAgHAADAQDgAAAAGwgEAADAQDgAAgIFwAAAADIQD\nAABgIBwAAAAD4QAAABgIBwAAwEA4AAAABsIBAAAwEA4AAICBcAAAAAyEAwAAYCAcAAAAA+EAAAAY\nCAcAAMBAOAAAAAbCAQAAMBAOAACAgXAAAAAMhAMAAGAgHAAAAAPhAAAAGAgHAADAQDgAAAAGwgEA\nADAQDgAAgIFwAAAADIQDAABgIBwAAAAD4QAAABgIBwAAwEA4AAAABsIBAAAwEA4AAICBcAAAAAyE\nAwAAYCAcAAAAA+EAAAAYCAcAAMBAOAAAAAbCAQAAMDx14WDNmjW6cOGCLl68qMDAwJQuBwCAVOep\nCwf/+c9/FB0draxZsxIOAABPPXtMnGJvxjj/RUVFGv9u3vwr2WtyS6oZh4SEaMOGDbp586ZOnz6t\nd999V0WKFFFQUJAkKWPGjAoODpaXl5eGDBmiAwcOKEuWLDp37pymT5+uGzduaOTIkXI4HIqMjFRA\nQIAiIyN1+PBh9e/fX2PGjFH//v01dOhQBQcH6z//+Y8kqXPnznr//fcVHR2tCRMmyNXVVS+88IKG\nDh0qd3f3pGoXAIBHdnLjYV3Yf8aYtnPmeuO2zeaimjVrq0OHTslWV5KFA0mKjo7WzJkzderUKXXp\n0kUZMmRQcHCwXn75ZX3zzTf64osvVLRoUV29elWLFy/WlStX9MYbb0iSfvvtN/Xv318FChTQ8uXL\nFRISoqCgIBUqVEiBgYHOD/qCBQvq1q1bOnfunNzd3RUREaFChQqpdu3a+vrrr5U5c2ZNnDhRS5cu\nVfPmzZOyXQAAHsm9wSA+luXQDz+sTD3hoGDBgpKkHDlyKCYmRidOnNCQIUMkSbGxscqTJ49Onjyp\nEiVKSJJ8fHyUN29eSVK2bNk0bdo0eXp66vr16/Ly8kpwOU2bNlVoaKg8PDzUuHFjXblyReHh4erV\nq5ck6ebNm6pYsWJStgoAwCPLXvQFXThwRrISHnNnz0FyStJwYLPZjNt58uTRqFGjlDNnTu3atUsX\nL15UmjRptGzZMknStWvXdOrUKUnS8OHDNXbsWOXLl0+TJ0/WuXPnnPO0LHMt1q1bV2+//bZsNpu+\n/PJLpUuXTs8//7ymTZsmb29vrV27VunSpUvKVgEAeGR5qxTSi6/ml8PhcE7r/GoHY4y7u7s8PdMm\na11JGg7uFRgYqP79+8tut0u6HQBeeuklbdy4US1btlSWLFnk6ekpd3d3NWjQQN26dVPmzJn1/PPP\nKyIiQpJUsmRJffjhhxo2bJhzvunTp1fBggUVFxfn3MMwaNAgderUSZZlKX369Bo9enRytgoAQKK4\nerjJ9a7b3t4ZUqyWO2zWvV/Dk9mJEyd05MgR1atXTxEREapfv75++ukneXh4JGsdgd+OVVycPVmX\nmRLc3FzpMxWhz9SFPlOXv9vn+5W7JUE18cua1Tve6cm65yA+OXLk0NixYzV79mzZ7Xb17ds32YMB\nAAD4nxQPB+nSpdP06dNTugwAAPBfT90fQQIAAEmLcAAAAAyEAwAAYCAcAAAAA+EAAAAYCAcAAMBA\nOAAAAAbCAQAAMBAOAACAgXAAAAAMhAMAAGAgHAAAAAPhAAAAGAgHAADAQDgAAAAGwgEAADAQDgAA\ngIFwAAAADIQDAABgIBwAAAAD4QAAABgIBwAAwEA4AAAABsIBAAAwEA4AAICBcAAAAAyEAwAAYCAc\nAAAAA+EAAAAYCAcAAMBAOAAAAAbCAQAAMBAOAACAgXAAAAAMhAMAAGAgHAAAAAPhAAAAGAgHAADA\nQDgAAAAGwgEAADAQDgAAgIFwAAAADIQDAABgIBwAAAAD4QAAABgIBwAAwEA4AAAABsIBAAAwEA4A\nAICBcAAAAAyEAwAAYCAcAAAAA+EAAAAYbJZlWSldxJPi4sWolC4hyWXN6k2fqQh9pi70mbo8DX1m\nzeod73T2HAAAAAPhAAAAGAghG+wLAAAPI0lEQVQHAADAQDgAAAAGwgEAADAQDgAAgIFwAAAADIQD\nAABgIBwAAAAD4QAAABgIBwAAwEA4AAAABsIBAAAwEA4AAICBSzYDAAADew4AAICBcAAAAAyEAwAA\nYCAcAAAAA+EAAAAYCAcAAMDgltIFpDSHw6HAwEAdPXpUHh4eCgoK0osvvpjSZSVKbGysBg4cqHPn\nzikmJkZdu3bV888/ry5duuill16SJLVq1Up169bV1KlTtX79erm5uWngwIEqVqyYwsLCNGDAANls\nNuXPn1+ffPKJXFxc4h2b0ho2bChvb29JUq5cudSiRQsNHz5crq6uqlSpknr06JHgttyzZ0+ix6ak\nkJAQLV26VJJ069YtHT58WOPGjdPo0aOVI0cOSdJ7772nMmXKPLV97t27V2PHjtWcOXMe6fn3OMam\nVJ+HDx/WsGHD5OrqKg8PD40aNUpZsmRRUFCQdu/erfTp00uSpk2bptjYWPXt21c3b95UtmzZNGLE\nCKVNm1aLFi3SggUL5Obmpq5du6patWq6cuVKvGNTqs+DBw8myXvPk7Y9e/furUuXLkmSzp07p+LF\ni2vChAnq0qWLrl69Knd3d6VJk0ZffPHFU9enwXrGrV692urfv79lWZb166+/Wl26dEnhihJv8eLF\nVlBQkGVZlnXlyhWratWq1qJFi6yZM2ca4w4cOGC99dZblsPhsM6dO2c1btzYsizL6ty5s/XLL79Y\nlmVZH3/8sfXDDz8kODYl3bx50/L39zemNWjQwAoLC7McDofVsWNH68CBAwluy0cZ+6QIDAy0FixY\nYI0fP95atWqVcd/T2ueMGTOs+vXrW82aNbMs69Gef/90bEr22aZNG+vQoUOWZVnW/PnzreDgYMuy\nLKtly5bW5cuXjccOGzbMWrJkiWVZlvXZZ59Zs2bNssLDw6369etbt27dsiIjI53/j29scrq3z6R6\n73nStucdV69etRo0aGBduHDBsizLqlOnjuVwOIwxT1Of93rmDyvs2rVLlStXliSVKFFCBw4cSOGK\nEq927dp6//33nbddXV114MABrV+/Xm3atNHAgQMVHR2tXbt2qVKlSrLZbMqZM6fsdruuXLmigwcP\n6t///rckqUqVKtqyZUuCY1PSkSNH9Ndff6lDhw5q27atduzYoZiYGOXOnVs2m02VKlXS1q1b492W\n0dHRiR77pNi/f79+++03tWjRQgcPHtSSJUvUunVrjRw5UnFxcU9tn7lz59aUKVOctx/l+fdPx6Zk\nn+PHj1ehQoUkSXa7XWnSpJHD4VBYWJgGDx6sli1bavHixZLM96M7te/bt08lS5aUh4eHvL29lTt3\nbh05ciTesSnZZ1K99zxp2/OOKVOm6M0331S2bNl06dIlRUZGqkuXLmrVqpV++uknSf/8OZ6SnvnD\nCtHR0fLy8nLednV1VVxcnNzcnvxVc2d3ZHR0tHr27KlevXopJiZGzZo1k6+vr6ZPn65PP/1U3t7e\nypgxo/G4qKgoWZYlm81mTIuOjo53rI+PT/I2dxdPT0+98847atasmU6dOqV3331XGTJkMGo8c+ZM\nvNvy3mkPGvukbPfPPvtM3bt3lyRVrFhRNWrUUK5cufTJJ59owYIFT22ftWrV0tmzZ523H+X590/H\nJqd7+8yWLZskaffu3Zo7d67mzZunGzdu6M0331T79u1lt9vVtm1b+fr6Kjo62nn47O4+70y7Mz06\nOjrescnp3j6LFSuWJO89T9r2lKTLly9r69at+uijjyTdPsR758vLtWvX1KpVKxUrVuyp6vNez/ye\nAy8vL12/ft152+FwPBEfEIl1/vx5tW3bVv7+/vLz81PNmjXl6+srSapZs6YOHTp0X4/Xr1+Xt7e3\ncTzr+vXrypAhQ4JjU1KePHnUoEED2Ww25cmTR97e3rp69arz/oRqdzgc8faT0NgnYbtHRkbq5MmT\nevXVVyVJTZo00QsvvCCbzabq1avHuz2fxj4lPdLz75+OTWkrV67UJ598ohkzZsjHx0dp06ZV27Zt\nlTZtWnl5eenVV1/VkSNHjJ4e1md8Y1NSUr33PInbc9WqVapfv75cXV0lSVmyZFHLli3l5uamzJkz\nq1ChQvr999+f6j6f+XBQqlQpbdy4UZK0Z88evfLKKylcUeJdunRJHTp0UL9+/dS0aVNJ0jvvvKN9\n+/ZJkrZu3aoiRYqoVKlS2rRpkxwOh/744w85HA75+PiocOHC2rZtmyRp48aNKlOmTIJjU9LixYs1\ncuRISdKFCxf0119/KV26dDp9+rQsy9KmTZuctd+7Lb28vOTu7p6osU+CHTt2qEKFCpJuf7Nu0KCB\n/vzzT0nm9nza+5T0SM+/fzo2JS1btkxz587VnDlz9MILL0iSTp06pdatW8tutys2Nla7d+92btsN\nGzZIul176dKlVaxYMe3atUu3bt1SVFSUTpw4oVdeeSXesSkpqd57nrTtKd3ur0qVKs7bW7ZsUa9e\nvSTd/mA/fvy48ubN+1T3+cxfeOnO2dzHjh2TZVkKDg5Wvnz5UrqsRAkKCtL333+vvHnzOqf16tVL\nY8aMkbu7u7JkyaJhw4bJy8tLU6ZM0caNG+VwOPTRRx+pTJky+v333/Xxxx8rNjZWefPmVVBQkFxd\nXeMdm5JiYmL00Ucf6Y8//pDNZlPfvn3l4uKi4OBg2e12VapUSb17905wW+7ZsyfRY1PaF198ITc3\nN7399tuSpE2bNmnixIny9PRUvnz5FBAQIFdX16e2z7Nnz6pPnz5atGjRIz3/HsfYlOhz/vz5Kl++\nvHLkyOH8Jli2bFn17NlTn3/+uVatWiV3d3f5+/urVatWunTpkvr376/r168rU6ZMGjdunNKlS6dF\nixZp4cKFsixLnTt3Vq1atRIcmxJ9Llq0SAcPHtSwYcMe+3vPk7Q9Fy1aJEmqV6+e5s+fb3y7Hz58\nuPbu3SsXFxd17NhRNWrUeOr6vNszHw4AAIDpmT+sAAAATIQDAABgIBwAAAAD4QAAABgIBwAAwEA4\nAFKZs2fPytfXV/7+/mrYsKHq1aun9u3bO/9ewqNYu3atJk2aJEmaPHmydu7cKUkaNGiQ9u/f/9hq\nnjJlimrVqqWbN286p23btk1vvfXWY1uGJEVFRTn/+uSFCxf07rvvPtb5A6kF4QBIhbJly6Zly5Yp\nNDRUK1asUIECBTR69OhHnk/16tWd1+/YsWOH7Ha7pNu/6S5atOhjrfncuXMaP378Y53nva5du6bD\nhw9LkrJnz67PP/88SZcHPK0IB8AzoFy5cjp+/Lik238psVmzZmrQoIHatWunsLAwSdKsWbPUoEED\nNWzYUIMHD5Z0+zLSAwYMUGhoqA4cOKCAgAAdPXpUb731lrZt26YePXpo9erVzuU0btxYhw4dUlhY\nmNq3b69GjRqpVatWOnTo0ENrbNGihVauXOncO3G3S5cuqVu3bmrcuLGaNGnivChNVFSUunbtqnr1\n6qlLly5q2LChzp4967zeSIsWLVStWjUNHDhQlmUpKChI4eHh6t69u86ePavXX39dERERqlixomJj\nYyVJx44dU4MGDSRJoaGhatSokfz9/TVw4EDdunXrH2wF4OlBOABSudjYWK1evVolSpRQTEyM+vTp\no48//ljffvutWrZsqT59+shut+uzzz7TkiVLFBISotjYWF24cME5j4YNG8rX11dBQUEqUKCAc7q/\nv79WrFgh6fafBL5165YKFy6s/v37q1+/flq6dKmGDRum3r17P7TOjBkzKjAwUIMGDTIOL0i391Q0\nadJEISEhmj59ugYPHqzo6Gh9+umnypMnj1asWKHu3bvr2LFjkqT169erUKFCWrhwoVavXq0dO3bo\n4MGDCggIULZs2fTpp586550pUyYVK1ZMmzZtkiStWLFCDRo00PHjx7Vo0SItWLBAy5YtU+bMmTVz\n5sy/vyGAp8iTcQUWAI9VeHi4/P39Jd3+89PFihXTBx98oFOnTilDhgwqVqyYJKlOnToaPHiwbty4\noZIlS6pp06aqXr262rdvr+zZsz90OVWrVtXQoUMVHR2t7777Tg0aNND169d14MAB5xXrJOnGjRuK\niIhQpkyZHji/GjVq6Pvvv9f48eNVvXp15/QtW7bo5MmTmjx5siQpLi5OZ86c0ebNmzV27FhJUtGi\nRZ3Xjqhfv7727dunr776SidPntTVq1d148YN42p4d2vQoIFWrFihatWq6fvvv9ecOXO0Zs0ahYWF\nqXnz5pJuh6zChQs/dJ0AqQHhAEiF7pxzcK/4Tkq0LEt2u13Tpk3Tnj17tHHjRnXs2NH5ofsgHh4e\nqlatmtatW6dVq1bps88+k8PhkIeHh7H8P//8M8EP5nsFBATIz8/PGO9wODR79mzntPDwcGXOnFmu\nrq6K7y/Az5kzR6tXr1bz5s1VoUIF57UlElK9enWNHDlSO3bsUI4cOZQ9e3bZ7XbVqVNHAQEBkm5f\nUOfOORdAasdhBeAZkjdvXl29etV59byVK1cqZ86ccjgcqlu3rl555RW9//77qlixoo4ePWo81tXV\nNd4PR39/f82aNUsZM2bUv/71L3l7e+ull15yhoPNmzerTZs2ia4xU6ZMCgwM1LRp05zTXn31VX39\n9deSpN9++01+fn7666+/VL58eS1fvlySdPToUR0/flw2m02bN29WixYt1KBBA926dUtHjhxxXq46\nLi7uvmV6eHiocuXKCg4Odp5vUK5cOa1Zs0aXL1+WZVkKDAzU7NmzE90H8DQjHADPEA8PD02YMEHD\nhg1T/fr1NW/ePE2YMEE+Pj5q0aKFmjZtqsaNGysmJkZNmjQxHlu5cmV98skn2r17tzG9dOnSioqK\ncn6oStKYMWO0ePFi+fn5ady4cZowYYJsNlui66xRo4Zq1arlvB0QEKC9e/fKz89PvXv31ujRo+Xl\n5aXu3bvr9OnT8vPz0+TJk5UlSxZ5enqqXbt2mjp1qvz8/BQcHKySJUvq7Nmzypw5s3LmzBnvTyT9\n/f114sQJ53ILFiyoHj16qF27dqpXr54cDoc6deqU6B6ApxlXZQTw1Fq2bJly5cql0qVL648//tCb\nb76pH3/8US4ufO8B/gnOOQDw1MqbN68++eQTORwOubi4aOjQoQQD4DFgzwEAADAQsQEAgIFwAAAA\nDIQDAABgIBwAAAAD4QAAABgIBwAAwPD/+ZgrC05/aksAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 576x396 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.barplot(sample_labels.index,sample_labels.values,alpha=0.8)\n",
    "plt.title(\"distribution of positive and negative classes in original dataset\")\n",
    "plt.xlabel(\"Positive  Negative\")\n",
    "plt.ylabel(\"Count\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "POSITIVE class labels in sample data is 84.21 %\n",
      "NEGATIVE class labels in sample data is 15.78 %\n"
     ]
    }
   ],
   "source": [
    "print(f\"POSITIVE class labels in sample data is {round(((sample_labels.value_counts()[0]/review_dataset_final_sorted_sample.shape[0])*100),2)} %\")\n",
    "print(f\"NEGATIVE class labels in sample data is {round(((sample_labels.value_counts()[1]/review_dataset_final_sorted_sample.shape[0])*100),2)} %\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "So, it is almost same as our original dataset. Now, we can proceed on with further operations"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Actually it is reviews dataset. So,I think it is better if we perform time-based splitting rather than random split\n",
    "So, we will sort it first and then divide the data in 80:20 ratio and from train we can divide cross validation data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#sorting the dataset according to time as we need time-based splitting\n",
    "final_data = review_dataset_final_sorted_sample.sort_values(by ='Time',axis=0,ascending=True,inplace=False,kind='quicksort')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(100000, 12)"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "final_data.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "So, 80000 will be train data and 20000 will be test data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "train_data = final_data['Cleaned_text'][:80000]\n",
    "test_data = final_data['Cleaned_text'][80000:]\n",
    "train_data_labels = final_data['Score'][:80000]\n",
    "test_data_labels = final_data['Score'][80000:]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "positive labels distribution for TRAIN data 84.5\n",
      "negative labels distribution for TRAIN data 15.5\n"
     ]
    }
   ],
   "source": [
    "#TRAIN DATA\n",
    "print(f\"positive labels distribution for TRAIN data {round((train_data_labels.value_counts()[0]/train_data_labels.shape[0])*100,2)}\")\n",
    "print(f\"negative labels distribution for TRAIN data {round((train_data_labels.value_counts()[1]/train_data_labels.shape[0])*100,2)}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "positive labels distribution for TEST data 83.06\n",
      "negative labels distribution for TEST data 16.94\n"
     ]
    }
   ],
   "source": [
    "#TEST DATA\n",
    "print(f\"positive labels distribution for TEST data {round((test_data_labels.value_counts()[0]/test_data_labels.shape[0])*100,2)}\")\n",
    "print(f\"negative labels distribution for TEST data {round((test_data_labels.value_counts()[1]/test_data_labels.shape[0])*100,2)}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "So, as we can see that there is slight variation in both distributions but it is fine to some extent"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "jiYocvSHLPo4"
   },
   "source": [
    "# Applying Logistic Regression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#importing all the necessary packages needed for Logistic regression\n",
    "from sklearn.model_selection import GridSearchCV\n",
    "from sklearn.linear_model import LogisticRegression"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "LE_XPVZJLPo4"
   },
   "source": [
    "## [5.1] Logistic Regression on BOW,<font color='red'> SET 1</font>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "_iciJvj2LPo4"
   },
   "source": [
    "### [5.1.1] Applying Logistic Regression with L1 regularization on BOW,<font color='red'> SET 1</font>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 1. Initially applying BOW on train data and then will apply BOW on test data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "MyOVi-xYLPo5"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "some feature names  ['aa', 'aaa', 'aaaaa', 'aaaaaaaaaaaaaaaaaaaargh', 'aaaaah', 'aaaah', 'aaaand', 'aaaarrrrghh', 'aaah', 'aachen']\n",
      "==================================================\n",
      "the type of count vectorizer  <class 'scipy.sparse.csr.csr_matrix'>\n",
      "the shape of out text BOW vectorizer  (80000, 34040)\n",
      "the number of unique words  34040\n"
     ]
    }
   ],
   "source": [
    "#BoW on train data\n",
    "count_vect = CountVectorizer() #in scikit-learn\n",
    "bow_train_data = count_vect.fit_transform(train_data)\n",
    "print(\"some feature names \", count_vect.get_feature_names()[:10])\n",
    "print('='*50)\n",
    "print(\"the type of count vectorizer \",type(bow_train_data))\n",
    "print(\"the shape of out text BOW vectorizer \",bow_train_data.get_shape())\n",
    "print(\"the number of unique words \", bow_train_data.get_shape()[1])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "GridSearchCV(cv=3, error_score='raise-deprecating',\n",
       "       estimator=LogisticRegression(C=1.0, class_weight=None, dual=False, fit_intercept=True,\n",
       "          intercept_scaling=1, max_iter=100, multi_class='warn',\n",
       "          n_jobs=None, penalty='l1', random_state=None, solver='warn',\n",
       "          tol=0.0001, verbose=0, warm_start=False),\n",
       "       fit_params=None, iid='warn', n_jobs=None,\n",
       "       param_grid=[{'C': [0.0001, 0.01, 1, 100, 10000]}],\n",
       "       pre_dispatch='2*n_jobs', refit=True, return_train_score='warn',\n",
       "       scoring='roc_auc', verbose=0)"
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tuned_parameters = [{'C': [10**-4, 10**-2, 10**0, 10**2, 10**4]}]\n",
    "\n",
    "#using GridSearchCV\n",
    "model = GridSearchCV(LogisticRegression(penalty='l1'),tuned_parameters,scoring='roc_auc',cv=3)\n",
    "model.fit(bow_train_data,train_data_labels.values)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "LogisticRegression(C=1, class_weight=None, dual=False, fit_intercept=True,\n",
      "          intercept_scaling=1, max_iter=100, multi_class='warn',\n",
      "          n_jobs=None, penalty='l1', random_state=None, solver='warn',\n",
      "          tol=0.0001, verbose=0, warm_start=False)\n"
     ]
    }
   ],
   "source": [
    "print(model.best_estimator_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#https://stackoverflow.com/questions/38692520/what-is-the-difference-between-fit-transform-and-transform-in-sklearn-countvecto\n",
    "bow_test_data = count_vect.transform(test_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train accuracy is 97.18%\n"
     ]
    }
   ],
   "source": [
    "print(f\"Train accuracy is {round(model.score(bow_train_data, train_data_labels.values)*100,2)}%\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test accuracy is 93.18%\n"
     ]
    }
   ],
   "source": [
    "print(f\"Test accuracy is {round(model.score(bow_test_data, test_data_labels.values)*100,2)}%\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "5h4NhywcLPo7"
   },
   "source": [
    "#### [5.1.1.1] Calculating sparsity on weight vector obtained using L1 regularization on BOW,<font color='red'> SET 1</font>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "-mmdLzUALPo-"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "4578\n"
     ]
    }
   ],
   "source": [
    "# Please write all the code with proper documentation\n",
    "l1_model = LogisticRegression(C=1.0,penalty='l1')\n",
    "l1_model.fit(bow_train_data,train_data_labels.values)\n",
    "weight_vec = l1_model.coef_\n",
    "print(np.count_nonzero(weight_vec))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Confusion matrix and AUC Scores"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "predictions = l1_model.predict(bow_test_data)\n",
    "prediction_probability_test = l1_model.predict_proba(bow_test_data)\n",
    "prediction_probability_train = l1_model.predict_proba(bow_train_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAdkAAAFlCAYAAACnT5IMAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\ngHYAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 576x396 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Code for drawing seaborn heatmaps\n",
    "class_names = ['positive','negative']\n",
    "heatmap_dataframe = pd.DataFrame(confusion_matrix(test_data_labels.values, predictions), index=class_names, columns=class_names )\n",
    "heatmap = sns.heatmap(heatmap_dataframe, annot=True, fmt=\"d\")\n",
    "\n",
    "# Setting tick labels for heatmap\n",
    "heatmap.xaxis.set_ticklabels(heatmap.xaxis.get_ticklabels())\n",
    "heatmap.yaxis.set_ticklabels(heatmap.yaxis.get_ticklabels())\n",
    "plt.xlabel('True label')\n",
    "plt.ylabel('Predicted label')\n",
    "plt.title(\"Confusion Matrix\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The AUC value for test data is 0.9318569699988875\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAlgAAAFlCAYAAAA3apYyAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAXpea5Ku9uJHAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 576x396 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#https://machinelearningmastery.com/roc-curves-and-precision-recall-curves-for-classification-in-python/\n",
    "#https://medium.com/greyatom/lets-learn-about-auc-roc-curve-4a94b4d88152\n",
    "#AUC and ROC Values\n",
    "\n",
    "fpr_train, tpr_train, threshold_train = roc_curve(train_data_labels, prediction_probability_train [:,1],pos_label='positive')\n",
    "fpr_test, tpr_test, threshold_test = roc_curve(test_data_labels, prediction_probability_test[:,1],pos_label='positive')\n",
    "print(f\"The AUC value for test data is {roc_auc_score(test_data_labels, prediction_probability_test[:,1])}\")\n",
    "\n",
    "plt.plot(fpr_train, tpr_train,'g', label = 'Train Data')\n",
    "plt.plot(fpr_test, tpr_test,'y', label = 'Test Data')\n",
    "plt.ylim(0,1)\n",
    "plt.legend(bbox_to_anchor=(1.05, 1), loc='upper left', borderaxespad=0.)\n",
    "plt.grid(True)\n",
    "plt.title(\"ROC Curve for Train and Test Data\")\n",
    "plt.xlabel(\"FPR\")\n",
    "plt.ylabel(\"TPR\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "aRHMa52qLPpA"
   },
   "source": [
    "### [5.1.2] Applying Logistic Regression with L2 regularization on BOW,<font color='red'> SET 1</font>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 101,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "GBbb9lxlLPpB"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "GridSearchCV(cv=3, error_score='raise-deprecating',\n",
       "       estimator=LogisticRegression(C=1.0, class_weight=None, dual=False, fit_intercept=True,\n",
       "          intercept_scaling=1, max_iter=100, multi_class='warn',\n",
       "          n_jobs=None, penalty='l2', random_state=None, solver='warn',\n",
       "          tol=0.0001, verbose=0, warm_start=False),\n",
       "       fit_params=None, iid='warn', n_jobs=None,\n",
       "       param_grid=[{'C': [0.0001, 0.01, 1, 100, 10000]}],\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": [
    "# Please write all the code with proper documentation\n",
    "tuned_parameters = [{'C': [10**-4, 10**-2, 10**0, 10**2, 10**4]}]\n",
    "\n",
    "#using GridSearchCV\n",
    "model = GridSearchCV(LogisticRegression(penalty='l2'),tuned_parameters,scoring='roc_auc',cv=3)\n",
    "model.fit(bow_train_data,train_data_labels.values)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 102,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "LogisticRegression(C=0.01, class_weight=None, dual=False, fit_intercept=True,\n",
      "          intercept_scaling=1, max_iter=100, multi_class='warn',\n",
      "          n_jobs=None, penalty='l2', random_state=None, solver='warn',\n",
      "          tol=0.0001, verbose=0, warm_start=False)\n"
     ]
    }
   ],
   "source": [
    "print(model.best_estimator_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 103,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "34040\n"
     ]
    }
   ],
   "source": [
    "l2_model = LogisticRegression(C=0.01,penalty='l2')\n",
    "l2_model.fit(bow_train_data,train_data_labels.values)\n",
    "weight_vec = l2_model.coef_\n",
    "print(np.count_nonzero(weight_vec))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train accuracy is 91.11%\n"
     ]
    }
   ],
   "source": [
    "print(f\"Train accuracy is {round(l2_model.score(bow_train_data,train_data_labels.values)*100,2)}%\") #Test accuracy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 105,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test accuracy is 89.73%\n"
     ]
    }
   ],
   "source": [
    "print(f\"Test accuracy is {round(l2_model.score(bow_test_data, test_data_labels.values)*100,2)}%\") #Test accuracy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 106,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "predictions = l2_model.predict(bow_test_data)\n",
    "prediction_probability_test = l2_model.predict_proba(bow_test_data)\n",
    "prediction_probability_train = l2_model.predict_proba(bow_train_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 107,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAdkAAAFlCAYAAACnT5IMAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nTkSuQmCC\n",
      "text/plain": [
       "<Figure size 576x396 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Code for drawing seaborn heatmaps\n",
    "class_names = ['positive','negative']\n",
    "heatmap_dataframe = pd.DataFrame(confusion_matrix(test_data_labels.values, predictions), index=class_names, columns=class_names )\n",
    "heatmap = sns.heatmap(heatmap_dataframe, annot=True, fmt=\"d\")\n",
    "\n",
    "# Setting tick labels for heatmap\n",
    "heatmap.xaxis.set_ticklabels(heatmap.xaxis.get_ticklabels())\n",
    "heatmap.yaxis.set_ticklabels(heatmap.yaxis.get_ticklabels())\n",
    "plt.xlabel('True label')\n",
    "plt.ylabel('Predicted label')\n",
    "plt.title(\"Confusion Matrix\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 108,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The AUC value for test data is 0.9329126470372093\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAlgAAAFlCAYAAAA3apYyAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nTkSuQmCC\n",
      "text/plain": [
       "<Figure size 576x396 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fpr_train, tpr_train, threshold_train = roc_curve(train_data_labels, prediction_probability_train [:,1],pos_label='positive')\n",
    "fpr_test, tpr_test, threshold_test = roc_curve(test_data_labels, prediction_probability_test[:,1],pos_label='positive')\n",
    "print(f\"The AUC value for test data is {roc_auc_score(test_data_labels, prediction_probability_test[:,1])}\")\n",
    "\n",
    "plt.plot(fpr_train, tpr_train,'g', label = 'Train Data')\n",
    "plt.plot(fpr_test, tpr_test,'y', label = 'Test Data')\n",
    "plt.ylim(0,1)\n",
    "plt.legend(bbox_to_anchor=(1.05, 1), loc='upper left', borderaxespad=0.)\n",
    "plt.grid(True)\n",
    "plt.title(\"ROC Curve for Train and Test Data\")\n",
    "plt.xlabel(\"FPR\")\n",
    "plt.ylabel(\"TPR\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 109,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1, 34040)"
      ]
     },
     "execution_count": 109,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "weight_vec.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "UaJkugTKLPpE"
   },
   "source": [
    "#### [5.1.2.1] Performing pertubation test (multicollinearity check) on BOW,<font color='red'> SET 1</font>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 110,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import numpy as np"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 111,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "collapsed": true,
    "id": "E1c5B1nELPpF"
   },
   "outputs": [],
   "source": [
    "#To understand multicollinearity , we perform pertubation test.Actual Definition: a deviation of a system, moving object, or process from its regular or normal state or path, caused by an outside influence which even seems same what we thought of\n",
    "#bow_train_data is my original data , variable new_data is pertubated data , adding error value =,1\n",
    "\n",
    "#let's consider our error value from uniform distribution b/w +0.0001 and -0.0001\n",
    "error = np.random.uniform(-0.0001,0.0001,1 ) #error value which is to be added as a part of pertubation\n",
    "new_data = bow_train_data.astype(float) #new_data is pertubated data (as we are adding float value.So, it is recommended to use float value over there)\n",
    "new_data.data = new_data.data+error\n",
    "\n",
    "\n",
    "pertubated_model = LogisticRegression(C=0.01,penalty='l2') #using the values obtained at l2 regularization\n",
    "pertubated_model.fit(new_data,train_data_labels)\n",
    "pertubated_weight_vec = pertubated_model.coef_\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 112,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<80000x34040 sparse matrix of type '<class 'numpy.int64'>'\n",
       "\twith 2522048 stored elements in Compressed Sparse Row format>"
      ]
     },
     "execution_count": 112,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "bow_train_data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 113,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1, 34040)"
      ]
     },
     "execution_count": 113,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pertubated_weight_vec.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 114,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "l2_weight_vec = weight_vec+10**-6 #adding 10**-6 to avoid divison by zero error\n",
    "p_weight_vec = pertubated_weight_vec+10**-6 #adding 10**-6 to avoid divison by zero error"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 115,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#percentage change\n",
    "percentage_change = (abs((l2_weight_vec - p_weight_vec)/l2_weight_vec)*100)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 116,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[ 0.00186035,  0.00175517,  0.0003536 , ...,  0.00048644,\n",
       "         0.00260827,  0.00329643]])"
      ]
     },
     "execution_count": 116,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "percentage_change"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 117,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "10th percentile value is 0.0011304274649346481\n",
      "20th percentile value is 0.002304258882984027\n",
      "30th percentile value is 0.0034516122733440685\n",
      "40th percentile value is 0.0045656470829227355\n",
      "50th percentile value is 0.005661931709816043\n",
      "60th percentile value is 0.0068176285735520896\n",
      "70th percentile value is 0.00807445293513242\n",
      "80th percentile value is 0.009449424201676423\n",
      "90th percentile value is 0.012544384508900365\n",
      "100th percentile value is 22.973575983298296\n"
     ]
    }
   ],
   "source": [
    "percentiles = [10,20,30,40,50,60,70,80,90,100]\n",
    "for each_percentile in percentiles:\n",
    "    print(f\"{each_percentile}th percentile value is {np.percentile(percentage_change,each_percentile)}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "mostly all of the values seem abnormal but b/w 90th and 100th percentile there is huge change. So, we will try to find b/w them "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 118,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "99.1th percentile value is 0.07910010315705121\n",
      "99.2th percentile value is 0.09093058087751701\n",
      "99.3th percentile value is 0.10745793501490912\n",
      "99.4th percentile value is 0.12276533415846307\n",
      "99.5th percentile value is 0.1438691173605509\n",
      "99.6th percentile value is 0.17063213460136128\n",
      "99.7th percentile value is 0.22968404506856513\n",
      "99.8th percentile value is 0.4164062723484683\n",
      "99.9th percentile value is 0.952100064403833\n",
      "100th percentile value is 22.973575983298296\n"
     ]
    }
   ],
   "source": [
    "percentiles = [99.1,99.2,99.3,99.4,99.5,99.6,99.7,99.8,99.9,100]\n",
    "for each_percentile in percentiles:\n",
    "    print(f\"{each_percentile}th percentile value is {np.percentile(percentage_change,each_percentile)}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 119,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "99.0th percentile value is 0.06744852272295626\n",
      "99.1th percentile value is 0.07910010315705121\n",
      "99.2th percentile value is 0.09093058087751701\n",
      "99.3th percentile value is 0.10745793501490912\n",
      "99.4th percentile value is 0.12276533415846307\n",
      "99.5th percentile value is 0.1438691173605509\n",
      "99.6th percentile value is 0.17063213460136128\n",
      "99.7th percentile value is 0.22968404506856513\n",
      "99.8th percentile value is 0.4164062723484683\n",
      "99.9th percentile value is 0.952100064403833\n",
      "100.0th percentile value is 22.973575983298296\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x292b4362978>]"
      ]
     },
     "execution_count": 119,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAd4AAAFJCAYAAAA1/Uw4AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nBC8AAA4ieAEAcND/B/AsdfButnyjAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 576x396 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "percentiles = np.linspace(99,100,11)\n",
    "for each_percentile in percentiles:\n",
    "    print(f\"{each_percentile}th percentile value is {np.percentile(percentage_change,each_percentile)}\")\n",
    "\n",
    "plt.plot(percentiles,np.percentile(percentage_change,percentiles) )"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "So, let's consider our threshold as >0.75, will try to find the features which have similar value"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 120,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "% of features which didnot change by more than 0.69 is 99.87074030552292\n"
     ]
    }
   ],
   "source": [
    "no_of_multicollinear_features = percentage_change[np.where(percentage_change>0.69)].size\n",
    "print(f\"% of features which didnot change by more than 0.69 is {(l2_weight_vec.size-no_of_multicollinear_features)/l2_weight_vec.size*100}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### So, only very small percentage of features have changed. So, we can conclude that there is no multi-collinearity"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 121,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "% of features which didnot change by more than 0.69 is 99.87074030552292\n"
     ]
    }
   ],
   "source": [
    "print(f\"% of features which didnot change by more than 0.69 is {(l2_weight_vec.size-no_of_multicollinear_features)/l2_weight_vec.size*100}\")"
   ]
  },
  {
   "cell_type": "raw",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "# https://gist.github.com/larsmans/3745866\n",
    "# https://stackoverflow.com/questions/15361885/linking-text-feature-names-to-their-tfidf-value\n",
    "\n",
    "# https://stackoverflow.com/questions/51462979/how-to-get-feature-importance-in-logistic-regression-using-weights"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "eB1rF0BQLPpG"
   },
   "source": [
    "### [5.1.3] Feature Importance on BOW,<font color='red'> SET 1</font>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "JZTP19dqLPpI"
   },
   "source": [
    "#### [5.1.3.1] Top 10 important features of positive class from<font color='red'> SET 1</font>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 122,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "oGRpCVJgLPpI"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Top 10 important features of positive reviews are : ['delici' 'great' 'best' 'perfect' 'excel' 'love' 'nice' 'amaz' 'thank'\n",
      " 'favorit']\n"
     ]
    }
   ],
   "source": [
    "imp_feat_df = pd.DataFrame(weight_vec.T,columns=['positive'])\n",
    "imp_feat_df['feature'] = count_vect.get_feature_names()\n",
    "top_features = imp_feat_df.sort_values(by='positive',kind='quicksort',ascending=False) # ascending=False because we want to arrange in descending order\n",
    "print(f\"Top 10 important features of positive reviews are : {np.array(top_features['feature'][:10])}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "UJLYBE1gLPpJ"
   },
   "source": [
    "#### [5.1.3.2] Top 10 important features of negative class from<font color='red'> SET 1</font>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 123,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "c9r7Zjx1LPpK"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Top 10 important features of negative reviews are : ['bland' 'threw' 'return' 'unfortun' 'money' 'horribl' 'terribl' 'aw'\n",
      " 'worst' 'disappoint']\n"
     ]
    }
   ],
   "source": [
    "print(f\"Top 10 important features of negative reviews are : {np.array(top_features['feature'].tail(10))}\") #tail becuase we know last values will be negative when top of them are positive"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "nKn3VXcZLPpN"
   },
   "source": [
    "## [5.2] Logistic Regression on TFIDF,<font color='red'> SET 2</font>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "naTBSzm6LPpN"
   },
   "source": [
    "### [5.2.1] Applying Logistic Regression with L1 regularization on TFIDF,<font color='red'> SET 2</font>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 124,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "PEhaagXXLPpS"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "some sample features(unique words in the corpus) ['aback', 'abandon', 'abdomin', 'abil', 'abl', 'abl add', 'abl buy', 'abl chew', 'abl drink', 'abl eat']\n",
      "==================================================\n",
      "the type of count vectorizer  <class 'scipy.sparse.csr.csr_matrix'>\n",
      "the shape of out text TFIDF vectorizer  (80000, 48758)\n",
      "the number of unique words including both unigrams and bigrams  48758\n"
     ]
    }
   ],
   "source": [
    "# Please write all the code with proper documentation\n",
    "tf_idf_vectorizer = TfidfVectorizer(ngram_range=(1,2),min_df=10)\n",
    "tf_idf_train_data = tf_idf_vectorizer.fit_transform(train_data)\n",
    "print(\"some sample features(unique words in the corpus)\",tf_idf_vectorizer.get_feature_names()[0:10])\n",
    "print('='*50)\n",
    "print(\"the type of count vectorizer \",type(tf_idf_train_data))\n",
    "print(\"the shape of out text TFIDF vectorizer \",tf_idf_train_data.get_shape())\n",
    "print(\"the number of unique words including both unigrams and bigrams \", tf_idf_train_data.get_shape()[1])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "NdFShHQJLPpV"
   },
   "source": [
    "### [5.2.2] Applying Logistic Regression with L2 regularization on TFIDF,<font color='red'> SET 2</font>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 125,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#https://stackoverflow.com/questions/1557571/how-do-i-get-time-of-a-python-programs-execution\n",
    "import timeit"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 126,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "Jyhq3JDzLPpW"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Program Executed in 54.1976971\n"
     ]
    }
   ],
   "source": [
    "# Please write all the code with proper documentation\n",
    "tuned_parameters = [{'C': [10**-4, 10**-2, 10**0, 10**2, 10**4]}]\n",
    "\n",
    "start = timeit.default_timer() #program starting time\n",
    "\n",
    "#using GridSearchCV\n",
    "model = GridSearchCV(LogisticRegression(penalty='l2'),tuned_parameters,scoring='roc_auc',cv=3)\n",
    "model.fit(tf_idf_train_data,train_data_labels.values)\n",
    "\n",
    "stop = timeit.default_timer()  #program end time\n",
    "execution_time = stop - start  #total execution time\n",
    "\n",
    "print(f\"Program Executed in {execution_time}\") #It returns time in sec"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 127,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "LogisticRegression(C=1, class_weight=None, dual=False, fit_intercept=True,\n",
      "          intercept_scaling=1, max_iter=100, multi_class='warn',\n",
      "          n_jobs=None, penalty='l2', random_state=None, solver='warn',\n",
      "          tol=0.0001, verbose=0, warm_start=False)\n"
     ]
    }
   ],
   "source": [
    "print(model.best_estimator_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 128,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "tf_idf_test_data = tf_idf_vectorizer.transform(test_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 129,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "48758\n"
     ]
    }
   ],
   "source": [
    "l2_model = LogisticRegression(C=1,penalty='l2')\n",
    "l2_model.fit(tf_idf_train_data,train_data_labels.values)\n",
    "weight_vec = l2_model.coef_\n",
    "print(np.count_nonzero(weight_vec))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 130,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train accuracy is 93.88%\n"
     ]
    }
   ],
   "source": [
    "print(f\"Train accuracy is {round(l2_model.score(tf_idf_train_data,train_data_labels.values)*100,2)}%\") #Test accuracy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 131,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test accuracy is 91.54%\n"
     ]
    }
   ],
   "source": [
    "print(f\"Test accuracy is {round(l2_model.score(tf_idf_test_data,test_data_labels.values)*100,2)}%\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 132,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "predictions = l2_model.predict(tf_idf_test_data)\n",
    "prediction_probability_test = l2_model.predict_proba(tf_idf_test_data)\n",
    "prediction_probability_train = l2_model.predict_proba(tf_idf_train_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 133,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAdkAAAFlCAYAAACnT5IMAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\niJKsiIiIQZRkRUREDKIkKyIiYpD/A5ASNmKHcri4AAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 576x396 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Code for drawing seaborn heatmaps\n",
    "class_names = ['positive','negative']\n",
    "heatmap_dataframe = pd.DataFrame(confusion_matrix(test_data_labels.values, predictions), index=class_names, columns=class_names )\n",
    "heatmap = sns.heatmap(heatmap_dataframe, annot=True, fmt=\"d\")\n",
    "\n",
    "# Setting tick labels for heatmap\n",
    "heatmap.xaxis.set_ticklabels(heatmap.xaxis.get_ticklabels())\n",
    "heatmap.yaxis.set_ticklabels(heatmap.yaxis.get_ticklabels())\n",
    "plt.xlabel('True label')\n",
    "plt.ylabel('Predicted label')\n",
    "plt.title(\"Confusion Matrix\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 134,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The AUC value for test data is 0.9561577036437138\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAlgAAAFlCAYAAAA3apYyAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<Figure size 576x396 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fpr_train, tpr_train, threshold_train = roc_curve(train_data_labels, prediction_probability_train [:,1],pos_label='positive')\n",
    "fpr_test, tpr_test, threshold_test = roc_curve(test_data_labels, prediction_probability_test[:,1],pos_label='positive')\n",
    "print(f\"The AUC value for test data is {roc_auc_score(test_data_labels, prediction_probability_test[:,1])}\")\n",
    "\n",
    "plt.plot(fpr_train, tpr_train,'g', label = 'Train Data')\n",
    "plt.plot(fpr_test, tpr_test,'y', label = 'Test Data')\n",
    "plt.ylim(0,1)\n",
    "plt.legend(bbox_to_anchor=(1.05, 1), loc='upper left', borderaxespad=0.)\n",
    "plt.grid(True)\n",
    "plt.title(\"ROC Curve for Train and Test Data\")\n",
    "plt.xlabel(\"FPR\")\n",
    "plt.ylabel(\"TPR\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "xeeJ8eVvLPpX"
   },
   "source": [
    "### [5.2.3] Feature Importance on TFIDF,<font color='red'> SET 2</font>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "Cj9XEmqQLPpY"
   },
   "source": [
    "#### [5.2.3.1] Top 10 important features of positive class from<font color='red'> SET 2</font>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 135,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "RG10txygLPpb"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Top 10 important features of positive reviews are : ['great' 'best' 'love' 'delici' 'perfect' 'excel' 'good' 'high recommend'\n",
      " 'nice' 'favorit']\n"
     ]
    }
   ],
   "source": [
    "imp_feat_df = pd.DataFrame(weight_vec.T,columns=['positive'])\n",
    "imp_feat_df['feature'] = tf_idf_vectorizer.get_feature_names()\n",
    "top_features = imp_feat_df.sort_values(by='positive',kind='quicksort',ascending=False) # ascending=False because we want to arrange in descending order\n",
    "print(f\"Top 10 important features of positive reviews are : {np.array(top_features['feature'][:10])}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "PowEfns4LPpg"
   },
   "source": [
    "#### [5.2.3.2] Top 10 important features of negative class from<font color='red'> SET 2</font>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 136,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "G7jrma95LPpg"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Top 10 important features of negative reviews are : ['money' 'bland' 'threw' 'return' 'unfortun' 'horribl' 'terribl' 'aw'\n",
      " 'worst' 'disappoint']\n"
     ]
    }
   ],
   "source": [
    "print(f\"Top 10 important features of negative reviews are : {np.array(top_features['feature'].tail(10))}\") #tail becuase we know last values will be negative when top of them are positive"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 94,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#saving features for future use\n",
    "np.savez('bow_train_and_test_data.npz',bow_train_data,bow_test_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "np.savez('tf_idf_train_and_test_data.npz',tf_idf_train_data,tf_idf_test_data)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "sqYadGOgLPpi"
   },
   "source": [
    "## [5.3] Logistic Regression on AVG W2V,<font color='red'> SET 3</font>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "kByBCEA1LPpj"
   },
   "source": [
    "### [5.3.1] Applying Logistic Regression with L1 regularization on AVG W2V<font color='red'> SET 3</font>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 197,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "collapsed": true,
    "id": "yCQ_4iJkLPpj"
   },
   "outputs": [],
   "source": [
    "# Train your own Word2Vec model using your own text corpus with train_data\n",
    "list_of_sentences_train=[]\n",
    "for sentance in train_data:\n",
    "    list_of_sentences_train.append(sentance.split())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 198,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#training W2V using test data\n",
    "list_of_sentences_test=[]\n",
    "for sentance in test_data:\n",
    "    list_of_sentences_test.append(sentance.split())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 199,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "number of words that occured minimum 5 times  11536\n"
     ]
    }
   ],
   "source": [
    "# Train your own Word2Vec model using your own text corpus with train_data with min_count = 5\n",
    "w2v_model=Word2Vec(list_of_sentences_train,min_count=5,size=50, workers=4)\n",
    "\n",
    "w2v_words = list(w2v_model.wv.vocab)\n",
    "print(\"number of words that occured minimum 5 times \",len(w2v_words))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 200,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|███████████████████████████████████████████████████████████████████████████| 80000/80000 [04:19<00:00, 307.76it/s]\n"
     ]
    }
   ],
   "source": [
    "#computing avgW2V for train data (for each review)\n",
    "train_vectors = []; # the avg-w2v for each sentence/review is stored in this list\n",
    "for sent in tqdm(list_of_sentences_train): # for each review/sentence\n",
    "    sent_vec = np.zeros(50) # as word vectors are of zero length 50, you might need to change this to 300 if you use google's w2v\n",
    "    cnt_words =0; # num of words with a valid vector in the sentence/review\n",
    "    for word in sent: # for each word in a review/sentence\n",
    "        if word in w2v_words:\n",
    "            vec = w2v_model.wv[word]\n",
    "            sent_vec += vec\n",
    "            cnt_words += 1\n",
    "    if cnt_words != 0:\n",
    "        sent_vec /= cnt_words\n",
    "    train_vectors.append(sent_vec)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 201,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|███████████████████████████████████████████████████████████████████████████| 20000/20000 [00:49<00:00, 400.73it/s]\n"
     ]
    }
   ],
   "source": [
    "#computing avgW2V for train data (for each review)\n",
    "test_vectors = []; # the avg-w2v for each sentence/review is stored in this list\n",
    "for sent in tqdm(list_of_sentences_test): # for each review/sentence\n",
    "    sent_vec = np.zeros(50) # as word vectors are of zero length 50, you might need to change this to 300 if you use google's w2v\n",
    "    cnt_words =0; # num of words with a valid vector in the sentence/review\n",
    "    for word in sent: # for each word in a review/sentence\n",
    "        if word in w2v_words:\n",
    "            vec = w2v_model.wv[word]\n",
    "            sent_vec += vec\n",
    "            cnt_words += 1\n",
    "    if cnt_words != 0:\n",
    "        sent_vec /= cnt_words\n",
    "    test_vectors.append(sent_vec)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 103,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#np.savez('avgW2V_train_and_test.npz',train_vectors,test_vectors)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 158,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#avgW2V = np.load('avgW2V_train_and_test.npz')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 159,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['arr_0', 'arr_1']"
      ]
     },
     "execution_count": 159,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#avgW2V.files"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 160,
   "metadata": {},
   "outputs": [],
   "source": [
    "#train_vectors = avgW2V['arr_0']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 161,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#test_vectors = avgW2V['arr_1']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 202,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Data-preprocessing: Standardizing the data , We can even proceed without standardizing but if we standardize our data then whole values will be in same range.So, we are doing it\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "standardization = StandardScaler()\n",
    "train_vector_standardized =standardization.fit_transform(train_vectors)\n",
    "test_vector_standardized = standardization.transform(test_vectors)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 203,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "GridSearchCV(cv=3, error_score='raise-deprecating',\n",
       "       estimator=LogisticRegression(C=1.0, class_weight=None, dual=False, fit_intercept=True,\n",
       "          intercept_scaling=1, max_iter=100, multi_class='warn',\n",
       "          n_jobs=None, penalty='l1', random_state=None, solver='warn',\n",
       "          tol=0.0001, verbose=0, warm_start=False),\n",
       "       fit_params=None, iid='warn', n_jobs=None,\n",
       "       param_grid=[{'C': [0.0001, 0.01, 1, 100, 10000]}],\n",
       "       pre_dispatch='2*n_jobs', refit=True, return_train_score='warn',\n",
       "       scoring='roc_auc', verbose=0)"
      ]
     },
     "execution_count": 203,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Please write all the code with proper documentation\n",
    "tuned_parameters = [{'C': [10**-4, 10**-2, 10**0, 10**2, 10**4]}]\n",
    "\n",
    "#using GridSearchCV\n",
    "model = GridSearchCV(LogisticRegression(penalty='l1'),tuned_parameters,scoring='roc_auc',cv=3)\n",
    "model.fit(train_vector_standardized,train_data_labels.values)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 204,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "LogisticRegression(C=1, class_weight=None, dual=False, fit_intercept=True,\n",
      "          intercept_scaling=1, max_iter=100, multi_class='warn',\n",
      "          n_jobs=None, penalty='l1', random_state=None, solver='warn',\n",
      "          tol=0.0001, verbose=0, warm_start=False)\n"
     ]
    }
   ],
   "source": [
    "print(model.best_estimator_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 205,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LogisticRegression(C=1, class_weight=None, dual=False, fit_intercept=True,\n",
       "          intercept_scaling=1, max_iter=100, multi_class='warn',\n",
       "          n_jobs=None, penalty='l1', random_state=None, solver='warn',\n",
       "          tol=0.0001, verbose=0, warm_start=False)"
      ]
     },
     "execution_count": 205,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "l1_model = LogisticRegression(C=1,penalty='l1')\n",
    "l1_model.fit(train_vector_standardized,train_data_labels.values)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 206,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train accuracy is 89.21%\n",
      "Test accuracy is 88.34%\n"
     ]
    }
   ],
   "source": [
    "print(f\"Train accuracy is {round(l1_model.score(train_vector_standardized,train_data_labels.values)*100,2)}%\")\n",
    "print(f\"Test accuracy is {round(l1_model.score(test_vector_standardized,test_data_labels.values)*100,2)}%\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 207,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "predictions = l1_model.predict(test_vector_standardized)\n",
    "prediction_probability_test = l1_model.predict_proba(test_vector_standardized)\n",
    "prediction_probability_train = l1_model.predict_proba(train_vector_standardized)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 208,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAdkAAAFlCAYAAACnT5IMAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<Figure size 576x396 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Code for drawing seaborn heatmaps\n",
    "class_names = ['positive','negative']\n",
    "heatmap_dataframe = pd.DataFrame(confusion_matrix(test_data_labels.values, predictions), index=class_names, columns=class_names )\n",
    "heatmap = sns.heatmap(heatmap_dataframe, annot=True, fmt=\"d\")\n",
    "\n",
    "# Setting tick labels for heatmap\n",
    "heatmap.xaxis.set_ticklabels(heatmap.xaxis.get_ticklabels())\n",
    "heatmap.yaxis.set_ticklabels(heatmap.yaxis.get_ticklabels())\n",
    "plt.xlabel('True label')\n",
    "plt.ylabel('Predicted label')\n",
    "plt.title(\"Confusion Matrix\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 209,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The AUC value for test data is 0.9077587954627114\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAlgAAAFlCAYAAAA3apYyAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<Figure size 576x396 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fpr_train, tpr_train, threshold_train = roc_curve(train_data_labels, prediction_probability_train [:,1],pos_label='positive')\n",
    "fpr_test, tpr_test, threshold_test = roc_curve(test_data_labels, prediction_probability_test[:,1],pos_label='positive')\n",
    "print(f\"The AUC value for test data is {roc_auc_score(test_data_labels, prediction_probability_test[:,1])}\")\n",
    "\n",
    "plt.plot(fpr_train, tpr_train,'g', label = 'Train Data')\n",
    "plt.plot(fpr_test, tpr_test,'y', label = 'Test Data')\n",
    "plt.ylim(0,1)\n",
    "plt.legend(bbox_to_anchor=(1.05, 1), loc='upper left', borderaxespad=0.)\n",
    "plt.grid(True)\n",
    "plt.title(\"ROC Curve for Train and Test Data\")\n",
    "plt.xlabel(\"FPR\")\n",
    "plt.ylabel(\"TPR\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "O_8Gu2SfLPpn"
   },
   "source": [
    "### [5.3.2] Applying Logistic Regression with L2 regularization on AVG W2V,<font color='red'> SET 3</font>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 210,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "Zy3s4jk1LPpn"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "GridSearchCV(cv=3, error_score='raise-deprecating',\n",
       "       estimator=LogisticRegression(C=1.0, class_weight=None, dual=False, fit_intercept=True,\n",
       "          intercept_scaling=1, max_iter=100, multi_class='warn',\n",
       "          n_jobs=None, penalty='l2', random_state=None, solver='warn',\n",
       "          tol=0.0001, verbose=0, warm_start=False),\n",
       "       fit_params=None, iid='warn', n_jobs=None,\n",
       "       param_grid=[{'C': [0.0001, 0.01, 1, 100, 10000]}],\n",
       "       pre_dispatch='2*n_jobs', refit=True, return_train_score='warn',\n",
       "       scoring='roc_auc', verbose=0)"
      ]
     },
     "execution_count": 210,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Please write all the code with proper documentation\n",
    "tuned_parameters = [{'C': [10**-4, 10**-2, 10**0, 10**2, 10**4]}]\n",
    "\n",
    "#using GridSearchCV\n",
    "model = GridSearchCV(LogisticRegression(penalty='l2'),tuned_parameters,scoring='roc_auc',cv=3)\n",
    "model.fit(train_vector_standardized,train_data_labels.values)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 211,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "LogisticRegression(C=0.01, class_weight=None, dual=False, fit_intercept=True,\n",
      "          intercept_scaling=1, max_iter=100, multi_class='warn',\n",
      "          n_jobs=None, penalty='l2', random_state=None, solver='warn',\n",
      "          tol=0.0001, verbose=0, warm_start=False)\n"
     ]
    }
   ],
   "source": [
    "print(model.best_estimator_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 212,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LogisticRegression(C=0.01, class_weight=None, dual=False, fit_intercept=True,\n",
       "          intercept_scaling=1, max_iter=100, multi_class='warn',\n",
       "          n_jobs=None, penalty='l2', random_state=None, solver='warn',\n",
       "          tol=0.0001, verbose=0, warm_start=False)"
      ]
     },
     "execution_count": 212,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "l2_model = LogisticRegression(C=0.01,penalty='l2')\n",
    "l2_model.fit(train_vector_standardized,train_data_labels.values)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 213,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train accuracy is 89.17%\n",
      "Test accuracy is 88.33%\n"
     ]
    }
   ],
   "source": [
    "print(f\"Train accuracy is {round(l2_model.score(train_vector_standardized,train_data_labels.values)*100,2)}%\")\n",
    "print(f\"Test accuracy is {round(l2_model.score(test_vector_standardized,test_data_labels.values)*100,2)}%\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 214,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "predictions = l2_model.predict(test_vector_standardized)\n",
    "prediction_probability_test = l2_model.predict_proba(test_vector_standardized)\n",
    "prediction_probability_train = l2_model.predict_proba(train_vector_standardized)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 215,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAdkAAAFlCAYAAACnT5IMAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nhP6v9a/f/FMAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 576x396 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Code for drawing seaborn heatmaps\n",
    "class_names = ['positive','negative']\n",
    "heatmap_dataframe = pd.DataFrame(confusion_matrix(test_data_labels.values, predictions), index=class_names, columns=class_names )\n",
    "heatmap = sns.heatmap(heatmap_dataframe, annot=True, fmt=\"d\")\n",
    "\n",
    "# Setting tick labels for heatmap\n",
    "heatmap.xaxis.set_ticklabels(heatmap.xaxis.get_ticklabels())\n",
    "heatmap.yaxis.set_ticklabels(heatmap.yaxis.get_ticklabels())\n",
    "plt.xlabel('True label')\n",
    "plt.ylabel('Predicted label')\n",
    "plt.title(\"Confusion Matrix\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 216,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The AUC value for test data is 0.9076082839019297\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAlgAAAFlCAYAAAA3apYyAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 576x396 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fpr_train, tpr_train, threshold_train = roc_curve(train_data_labels, prediction_probability_train [:,1],pos_label='positive')\n",
    "fpr_test, tpr_test, threshold_test = roc_curve(test_data_labels, prediction_probability_test[:,1],pos_label='positive')\n",
    "print(f\"The AUC value for test data is {roc_auc_score(test_data_labels, prediction_probability_test[:,1])}\")\n",
    "\n",
    "plt.plot(fpr_train, tpr_train,'g', label = 'Train Data')\n",
    "plt.plot(fpr_test, tpr_test,'y', label = 'Test Data')\n",
    "plt.ylim(0,1)\n",
    "plt.legend(bbox_to_anchor=(1.05, 1), loc='upper left', borderaxespad=0.)\n",
    "plt.grid(True)\n",
    "plt.title(\"ROC Curve for Train and Test Data\")\n",
    "plt.xlabel(\"FPR\")\n",
    "plt.ylabel(\"TPR\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "pdXTwVIPLPpp"
   },
   "source": [
    "## [5.4] Logistic Regression on TFIDF W2V,<font color='red'> SET 4</font>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "WP9NjwNSLPpp"
   },
   "source": [
    "### [5.4.1] Applying Logistic Regression with L1 regularization on TFIDF W2V,<font color='red'> SET 4</font>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 217,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "207gYymwLPpq"
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████████████████████████████████████████████████████████████████████| 80000/80000 [2:32:40<00:00,  8.73it/s]\n"
     ]
    }
   ],
   "source": [
    "# Please write all the code with proper documentation\n",
    "#I am using tfidf_train and test which is already done in previous steps and also using Only Word2Vec from previous one\n",
    "dictionary = dict(zip(tf_idf_vectorizer.get_feature_names(), list(tf_idf_vectorizer.idf_)))\n",
    "tfidf_features = tf_idf_vectorizer.get_feature_names()\n",
    "tfidf_sent_train = []; # the tfidf-w2v for each sentence/review is stored in this list\n",
    "row=0;\n",
    "for sent in tqdm(list_of_sentences_train): # for each review/sentence \n",
    "    sent_vec = np.zeros(50) # as word vectors are of zero length\n",
    "    weight_sum =0; # num of words with a valid vector in the sentence/review\n",
    "    for word in sent: # for each word in a review/sentence\n",
    "        if word in w2v_words and word in tfidf_features:\n",
    "            vec = w2v_model.wv[word]\n",
    "            #  tf_idf = tf_idf_matrix[row, tfidf_feat.index(word)]\n",
    "            # to reduce the computation we are \n",
    "            # dictionary[word] = idf value of word in whole courpus\n",
    "            # sent.count(word) = tf valeus of word in this review\n",
    "            tf_idf = dictionary[word]*(sent.count(word)/len(sent))\n",
    "            sent_vec += (vec * tf_idf)\n",
    "            weight_sum += tf_idf\n",
    "    if weight_sum != 0:\n",
    "        sent_vec /= weight_sum\n",
    "    tfidf_sent_train.append(sent_vec)\n",
    "    row += 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 218,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|████████████████████████████████████████████████████████████████████████████| 20000/20000 [32:12<00:00, 10.35it/s]\n"
     ]
    }
   ],
   "source": [
    "#test-data (for test data)\n",
    "dictionary = dict(zip(tf_idf_vectorizer.get_feature_names(), list(tf_idf_vectorizer.idf_)))\n",
    "tfidf_features = tf_idf_vectorizer.get_feature_names()\n",
    "tfidf_sent_test = []; # the tfidf-w2v for each sentence/review is stored in this list\n",
    "row=0;\n",
    "for sent in tqdm(list_of_sentences_test): # for each review/sentence \n",
    "    sent_vec = np.zeros(50) # as word vectors are of zero length\n",
    "    weight_sum =0; # num of words with a valid vector in the sentence/review\n",
    "    for word in sent: # for each word in a review/sentence\n",
    "        if word in w2v_words and word in tfidf_features:\n",
    "            vec = w2v_model.wv[word]\n",
    "            #  tf_idf = tf_idf_matrix[row, tfidf_feat.index(word)]\n",
    "            # to reduce the computation we are \n",
    "            # dictionary[word] = idf value of word in whole courpus\n",
    "            # sent.count(word) = tf valeus of word in this review\n",
    "            tf_idf = dictionary[word]*(sent.count(word)/len(sent))\n",
    "            sent_vec += (vec * tf_idf)\n",
    "            weight_sum += tf_idf\n",
    "    if weight_sum != 0:\n",
    "        sent_vec /= weight_sum\n",
    "    tfidf_sent_test.append(sent_vec)\n",
    "    row += 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 118,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#np.savez('tf_idf_W2V_train_and_test.npz',tfidf_sent_train,tfidf_sent_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 177,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#tfidfW2V = np.load('tf_idf_W2V_train_and_test.npz')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 178,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#tfidf_sent_train = tfidfW2V['arr_0']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 179,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#tfidf_sent_test = tfidfW2V['arr_1']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 219,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Data-preprocessing: Standardizing the data , We can even proceed without standardizing but if we standardize our data then whole values will be in same range.So, we are doing it\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "standardization = StandardScaler()\n",
    "tfidf_sent_train_standardized =standardization.fit_transform(tfidf_sent_train)\n",
    "tfidf_sent_test_standardized = standardization.transform(tfidf_sent_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 220,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "GridSearchCV(cv=3, error_score='raise-deprecating',\n",
       "       estimator=LogisticRegression(C=1.0, class_weight=None, dual=False, fit_intercept=True,\n",
       "          intercept_scaling=1, max_iter=100, multi_class='warn',\n",
       "          n_jobs=None, penalty='l1', random_state=None, solver='warn',\n",
       "          tol=0.0001, verbose=0, warm_start=False),\n",
       "       fit_params=None, iid='warn', n_jobs=None,\n",
       "       param_grid=[{'C': [0.0001, 0.01, 1, 100, 10000]}],\n",
       "       pre_dispatch='2*n_jobs', refit=True, return_train_score='warn',\n",
       "       scoring='roc_auc', verbose=0)"
      ]
     },
     "execution_count": 220,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Please write all the code with proper documentation\n",
    "tuned_parameters = [{'C': [10**-4, 10**-2, 10**0, 10**2, 10**4]}]\n",
    "\n",
    "#using GridSearchCV\n",
    "model = GridSearchCV(LogisticRegression(penalty='l1'),tuned_parameters,scoring='roc_auc',cv=3)\n",
    "model.fit(tfidf_sent_train_standardized,train_data_labels.values)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 221,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "LogisticRegression(C=1, class_weight=None, dual=False, fit_intercept=True,\n",
      "          intercept_scaling=1, max_iter=100, multi_class='warn',\n",
      "          n_jobs=None, penalty='l1', random_state=None, solver='warn',\n",
      "          tol=0.0001, verbose=0, warm_start=False)\n"
     ]
    }
   ],
   "source": [
    "print(model.best_estimator_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 222,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LogisticRegression(C=1, class_weight=None, dual=False, fit_intercept=True,\n",
       "          intercept_scaling=1, max_iter=100, multi_class='warn',\n",
       "          n_jobs=None, penalty='l1', random_state=None, solver='warn',\n",
       "          tol=0.0001, verbose=0, warm_start=False)"
      ]
     },
     "execution_count": 222,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "l1_model = LogisticRegression(C=1,penalty='l1')\n",
    "l1_model.fit(tfidf_sent_train_standardized,train_data_labels.values)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 223,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train accuracy is 87.86%\n",
      "Test accuracy is 87.2%\n"
     ]
    }
   ],
   "source": [
    "print(f\"Train accuracy is {round(l1_model.score(tfidf_sent_train_standardized,train_data_labels.values)*100,2)}%\")\n",
    "print(f\"Test accuracy is {round(l1_model.score(tfidf_sent_test_standardized,test_data_labels.values)*100,2)}%\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 224,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "predictions = l1_model.predict(tfidf_sent_test_standardized)\n",
    "prediction_probability_test = l1_model.predict_proba(tfidf_sent_test_standardized)\n",
    "prediction_probability_train = l1_model.predict_proba(tfidf_sent_train_standardized)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 225,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAdkAAAFlCAYAAACnT5IMAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nf6Gi99DnucpJAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 576x396 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Code for drawing seaborn heatmaps\n",
    "class_names = ['positive','negative']\n",
    "heatmap_dataframe = pd.DataFrame(confusion_matrix(test_data_labels.values, predictions), index=class_names, columns=class_names )\n",
    "heatmap = sns.heatmap(heatmap_dataframe, annot=True, fmt=\"d\")\n",
    "\n",
    "# Setting tick labels for heatmap\n",
    "heatmap.xaxis.set_ticklabels(heatmap.xaxis.get_ticklabels())\n",
    "heatmap.yaxis.set_ticklabels(heatmap.yaxis.get_ticklabels())\n",
    "plt.xlabel('True label')\n",
    "plt.ylabel('Predicted label')\n",
    "plt.title(\"Confusion Matrix\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 226,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The AUC value for test data is 0.879929646636763\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAlgAAAFlCAYAAAA3apYyAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<Figure size 576x396 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fpr_train, tpr_train, threshold_train = roc_curve(train_data_labels, prediction_probability_train [:,1],pos_label='positive')\n",
    "fpr_test, tpr_test, threshold_test = roc_curve(test_data_labels, prediction_probability_test[:,1],pos_label='positive')\n",
    "print(f\"The AUC value for test data is {roc_auc_score(test_data_labels, prediction_probability_test[:,1])}\")\n",
    "\n",
    "plt.plot(fpr_train, tpr_train,'g', label = 'Train Data')\n",
    "plt.plot(fpr_test, tpr_test,'y', label = 'Test Data')\n",
    "plt.ylim(0,1)\n",
    "plt.legend(bbox_to_anchor=(1.05, 1), loc='upper left', borderaxespad=0.)\n",
    "plt.grid(True)\n",
    "plt.title(\"ROC Curve for Train and Test Data\")\n",
    "plt.xlabel(\"FPR\")\n",
    "plt.ylabel(\"TPR\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "ynUK2tYDLPps"
   },
   "source": [
    "### [5.4.2] Applying Logistic Regression with L2 regularization on TFIDF W2V,<font color='red'> SET 4</font>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 227,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "hVYuiXl2LPpt"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "GridSearchCV(cv=3, error_score='raise-deprecating',\n",
       "       estimator=LogisticRegression(C=1.0, class_weight=None, dual=False, fit_intercept=True,\n",
       "          intercept_scaling=1, max_iter=100, multi_class='warn',\n",
       "          n_jobs=None, penalty='l2', random_state=None, solver='warn',\n",
       "          tol=0.0001, verbose=0, warm_start=False),\n",
       "       fit_params=None, iid='warn', n_jobs=None,\n",
       "       param_grid=[{'C': [0.0001, 0.01, 1, 100, 10000]}],\n",
       "       pre_dispatch='2*n_jobs', refit=True, return_train_score='warn',\n",
       "       scoring='roc_auc', verbose=0)"
      ]
     },
     "execution_count": 227,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Please write all the code with proper documentation\n",
    "tuned_parameters = [{'C': [10**-4, 10**-2, 10**0, 10**2, 10**4]}]\n",
    "\n",
    "#using GridSearchCV\n",
    "model = GridSearchCV(LogisticRegression(penalty='l2'),tuned_parameters,scoring='roc_auc',cv=3)\n",
    "model.fit(tfidf_sent_train_standardized,train_data_labels.values)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 228,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "LogisticRegression(C=0.01, class_weight=None, dual=False, fit_intercept=True,\n",
      "          intercept_scaling=1, max_iter=100, multi_class='warn',\n",
      "          n_jobs=None, penalty='l2', random_state=None, solver='warn',\n",
      "          tol=0.0001, verbose=0, warm_start=False)\n"
     ]
    }
   ],
   "source": [
    "print(model.best_estimator_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 229,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LogisticRegression(C=0.01, class_weight=None, dual=False, fit_intercept=True,\n",
       "          intercept_scaling=1, max_iter=100, multi_class='warn',\n",
       "          n_jobs=None, penalty='l2', random_state=None, solver='warn',\n",
       "          tol=0.0001, verbose=0, warm_start=False)"
      ]
     },
     "execution_count": 229,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "l2_model = LogisticRegression(C=0.01,penalty='l2')\n",
    "l2_model.fit(tfidf_sent_train_standardized,train_data_labels.values)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 230,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train accuracy is 87.79%\n",
      "Test accuracy is 87.14%\n"
     ]
    }
   ],
   "source": [
    "print(f\"Train accuracy is {round(l2_model.score(tfidf_sent_train_standardized,train_data_labels.values)*100,2)}%\")\n",
    "print(f\"Test accuracy is {round(l2_model.score(tfidf_sent_test_standardized,test_data_labels.values)*100,2)}%\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 231,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "predictions = l2_model.predict(tfidf_sent_test_standardized)\n",
    "prediction_probability_test = l2_model.predict_proba(tfidf_sent_test_standardized)\n",
    "prediction_probability_train = l2_model.predict_proba(tfidf_sent_train_standardized)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 232,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAdkAAAFlCAYAAACnT5IMAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\ngg==\n",
      "text/plain": [
       "<Figure size 576x396 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Code for drawing seaborn heatmaps\n",
    "class_names = ['positive','negative']\n",
    "heatmap_dataframe = pd.DataFrame(confusion_matrix(test_data_labels.values, predictions), index=class_names, columns=class_names )\n",
    "heatmap = sns.heatmap(heatmap_dataframe, annot=True, fmt=\"d\")\n",
    "\n",
    "# Setting tick labels for heatmap\n",
    "heatmap.xaxis.set_ticklabels(heatmap.xaxis.get_ticklabels())\n",
    "heatmap.yaxis.set_ticklabels(heatmap.yaxis.get_ticklabels())\n",
    "plt.xlabel('True label')\n",
    "plt.ylabel('Predicted label')\n",
    "plt.title(\"Confusion Matrix\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 233,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The AUC value for test data is 0.8799096802736899\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAlgAAAFlCAYAAAA3apYyAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<Figure size 576x396 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fpr_train, tpr_train, threshold_train = roc_curve(train_data_labels, prediction_probability_train [:,1],pos_label='positive')\n",
    "fpr_test, tpr_test, threshold_test = roc_curve(test_data_labels, prediction_probability_test[:,1],pos_label='positive')\n",
    "print(f\"The AUC value for test data is {roc_auc_score(test_data_labels, prediction_probability_test[:,1])}\")\n",
    "\n",
    "plt.plot(fpr_train, tpr_train,'g', label = 'Train Data')\n",
    "plt.plot(fpr_test, tpr_test,'y', label = 'Test Data')\n",
    "plt.ylim(0,1)\n",
    "plt.legend(bbox_to_anchor=(1.05, 1), loc='upper left', borderaxespad=0.)\n",
    "plt.grid(True)\n",
    "plt.title(\"ROC Curve for Train and Test Data\")\n",
    "plt.xlabel(\"FPR\")\n",
    "plt.ylabel(\"TPR\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "nOvsBNf2LPpv"
   },
   "source": [
    "# [6] Conclusions"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "collapsed": true,
    "id": "T7VxD0q2LPpw"
   },
   "source": [
    "\n",
    "1. Among all the NLP techniques, we have obtained good accuracy with tf-idf model\n",
    "\n",
    "2. In order to test for multi-collinearity , we have tested using pertubation test. We don't have multi-collinearity (as we have seen that 99.8% hasn\n",
    "'t changed after pertubation)\n",
    "\n",
    "3. Important features for positive reviews are , i.e., when it has these words then they are predicted as POSITIVE review{'delici' 'great' 'best' 'excel' 'perfect' 'love' 'nice' 'amaz' 'favorit'\n",
    " 'wonder'}\n",
    "Similarly, When it has negative reviews , important features which obtained are {'stale' 'wast' 'money' 'unfortun' 'horribl' 'return' 'aw' 'terribl'\n",
    " 'worst' 'disappoint'}\n",
    "\n",
    "4. We have also seen that we have got sparsity for L1 more than L2 ( By L1 regularization, you essentially make the vector x smaller (sparse), as most of its components are useless (zeros), and at the same time, the remaining non-zero components are very “useful”.) https://medium.com/mlreview/l1-norm-regularization-and-sparsity-explained-for-dummies-5b0e4be3938a\n",
    "\n",
    "There is table below which has all the detailed model evaluations.. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "-----> Even though we haven't tested using cross-validation dataset, By observing test and train accuracies we can co-relate whether our model is good or is it tending to overfit/underfit.\n",
    "\n",
    "-----> Here, in only B.O.W, we can see such phenomena where our there is slight more difference between train and test accuracy. But, typically it also might not be case of overfitting as our test accuracy is somewhat on higher side.\n",
    "\n",
    "-----> Among all, we have got high roc-auc on Tf-idf model. So, we would say, it is best among others."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 234,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "+-----------------+---------------+-------------------+---------+----------------+---------------+\n",
      "| vectorizer_type | model-Penalty | Hyperparameter(C) | roc-auc | train_accuracy | test_accuracy |\n",
      "+-----------------+---------------+-------------------+---------+----------------+---------------+\n",
      "|   Bag of words  |       L1      |        0.01       |   0.93  |     97.18%     |     93.18%    |\n",
      "|   Bag of words  |       L2      |         1         |   0.93  |     91.1%      |     89.7%     |\n",
      "|      TF-IDF     |       L2      |         1         |  0.956  |     93.4%      |     91.2%     |\n",
      "|      AvgW2V     |       L1      |         1         |  0.907  |     89.21%     |     88.34%    |\n",
      "|      AvgW2V     |       L2      |         1         |  0.907  |     89.17%     |     88.3%     |\n",
      "|    Tf_idfW2V    |       L1      |         1         |  0.879  |     87.86%     |     87.2%     |\n",
      "|    Tf_idfW2V    |       L2      |        0.01       |  0.879  |     87.79%     |     87.14%    |\n",
      "+-----------------+---------------+-------------------+---------+----------------+---------------+\n"
     ]
    }
   ],
   "source": [
    "from prettytable import PrettyTable\n",
    "# reference : http://zetcode.com/python/prettytable/\n",
    "pretty_table = PrettyTable()\n",
    "pretty_table.field_names = [\"vectorizer_type\",\"model-Penalty\",\"Hyperparameter(C)\",\"roc-auc\",\"train_accuracy\",\"test_accuracy\"]\n",
    "pretty_table.add_row([\"Bag of words\",\"L1\",\"0.01\",\"0.93\",\"97.18%\",\"93.18%\"])\n",
    "pretty_table.add_row([\"Bag of words\",\"L2\",\"1\",\"0.93\",\"91.1%\",\"89.7%\"])\n",
    "pretty_table.add_row([\"TF-IDF\",\"L2\",\"1\",\"0.956\",\"93.4%\",\"91.2%\"])\n",
    "pretty_table.add_row([\"AvgW2V\",\"L1\",\"1\",\"0.907\",\"89.21%\",\"88.34%\"])\n",
    "pretty_table.add_row([\"AvgW2V\",\"L2\",\"1\",\"0.907\",\"89.17%\",\"88.3%\"])\n",
    "pretty_table.add_row([\"Tf_idfW2V\",\"L1\",\"1\",\"0.879\",\"87.86%\",\"87.2%\"])\n",
    "pretty_table.add_row([\"Tf_idfW2V\",\"L2\",\"0.01\",\"0.879\",\"87.79%\",\"87.14%\"])\n",
    "print(pretty_table)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "colab": {
   "name": "05 Amazon Fine Food Reviews Analysis_Logistic Regression.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": "3.6.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}