04 Amazon Fine Food Reviews Analysis_NaiveBayes.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "UIn1GsFNJkjH"
   },
   "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": "5Uf26aoLJkjJ"
   },
   "source": [
    "# [1]. Reading Data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "yo-LaJg3JkjK"
   },
   "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",
    "id": "dLYq3BkfJkjL"
   },
   "outputs": [
    {
     "ename": "ImportError",
     "evalue": "cannot import name 'cross_validation'",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mImportError\u001b[0m                               Traceback (most recent call last)",
      "\u001b[1;32m<ipython-input-1-d4db3f78992f>\u001b[0m in \u001b[0;36m<module>\u001b[1;34m()\u001b[0m\n\u001b[0;32m     13\u001b[0m \u001b[1;32mfrom\u001b[0m \u001b[0msklearn\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mfeature_extraction\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mtext\u001b[0m \u001b[1;32mimport\u001b[0m \u001b[0mTfidfTransformer\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m     14\u001b[0m \u001b[1;32mfrom\u001b[0m \u001b[0msklearn\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mfeature_extraction\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mtext\u001b[0m \u001b[1;32mimport\u001b[0m \u001b[0mTfidfVectorizer\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m---> 15\u001b[1;33m \u001b[1;32mfrom\u001b[0m \u001b[0msklearn\u001b[0m \u001b[1;32mimport\u001b[0m \u001b[0mcross_validation\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m     16\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m     17\u001b[0m \u001b[1;32mfrom\u001b[0m \u001b[0msklearn\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mfeature_extraction\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mtext\u001b[0m \u001b[1;32mimport\u001b[0m \u001b[0mCountVectorizer\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;31mImportError\u001b[0m: cannot import name 'cross_validation'"
     ]
    }
   ],
   "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",
    "from sklearn import cross_validation\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\n",
    "from nltk.stem.porter import PorterStemmer\n",
    "from sklearn.metrics import confusion_matrix\n",
    "from scipy.sparse import csr_matrix\n",
    "from scipy import sparse\n",
    "from sklearn import metrics\n",
    "from sklearn import preprocessing\n",
    "from sklearn.metrics import roc_curve, auc\n",
    "from nltk.stem.porter import PorterStemmer\n",
    "from sklearn import model_selection\n",
    "from sklearn.decomposition import TruncatedSVD\n",
    "import seaborn as sns\n",
    "from sklearn.naive_bayes import MultinomialNB\n",
    "from sklearn.model_selection import cross_val_score\n",
    "from sklearn.metrics import accuracy_score\n",
    "from sklearn.model_selection import cross_validate\n",
    "from collections import Counter\n",
    "from sklearn.metrics import precision_score\n",
    "from sklearn.metrics import f1_score\n",
    "from sklearn.metrics import recall_score\n",
    "from sklearn.model_selection import train_test_split\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": "code",
   "execution_count": 2,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "rXVxnx_NJkjO",
    "outputId": "98628980-4d6e-4c53-913f-2d207932d2d0"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of data points in our data (525814, 10)\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>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",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>B001E4KFG0</td>\n",
       "      <td>A3SGXH7AUHU8GW</td>\n",
       "      <td>delmartian</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1303862400</td>\n",
       "      <td>Good Quality Dog Food</td>\n",
       "      <td>I have bought several of the Vitality canned d...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>B00813GRG4</td>\n",
       "      <td>A1D87F6ZCVE5NK</td>\n",
       "      <td>dll pa</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1346976000</td>\n",
       "      <td>Not as Advertised</td>\n",
       "      <td>Product arrived labeled as Jumbo Salted Peanut...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3</td>\n",
       "      <td>B000LQOCH0</td>\n",
       "      <td>ABXLMWJIXXAIN</td>\n",
       "      <td>Natalia Corres \"Natalia Corres\"</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1219017600</td>\n",
       "      <td>\"Delight\" says it all</td>\n",
       "      <td>This is a confection that has been around a fe...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Id   ProductId          UserId                      ProfileName  \\\n",
       "0   1  B001E4KFG0  A3SGXH7AUHU8GW                       delmartian   \n",
       "1   2  B00813GRG4  A1D87F6ZCVE5NK                           dll pa   \n",
       "2   3  B000LQOCH0   ABXLMWJIXXAIN  Natalia Corres \"Natalia Corres\"   \n",
       "\n",
       "   HelpfulnessNumerator  HelpfulnessDenominator  Score        Time  \\\n",
       "0                     1                       1      1  1303862400   \n",
       "1                     0                       0      0  1346976000   \n",
       "2                     1                       1      1  1219017600   \n",
       "\n",
       "                 Summary                                               Text  \n",
       "0  Good Quality Dog Food  I have bought several of the Vitality canned d...  \n",
       "1      Not as Advertised  Product arrived labeled as Jumbo Salted Peanut...  \n",
       "2  \"Delight\" says it all  This is a confection that has been around a fe...  "
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# using SQLite Table to read data.\n",
    "# con = sqlite3.connect('database.sqlite') \n",
    "con = sqlite3.connect('C:\\\\Users\\\\rahul\\\\Desktop\\\\AI COURSE\\\\Amazon Fine food\\\\amazon-fine-food-reviews (1)\\\\database.sqlite')\n",
    "\n",
    "# filtering only positive and negative reviews i.e. \n",
    "# not taking into consideration those reviews with Score=3\n",
    "# SELECT * FROM Reviews WHERE Score != 3 LIMIT 500000, will give top 500000 data points\n",
    "# you can change the number to any other number based on your computing power\n",
    "\n",
    "# filtered_data = pd.read_sql_query(\"\"\" SELECT * FROM Reviews WHERE Score != 3 LIMIT 500000\"\"\", con) \n",
    "# for tsne assignment you can take 5k data points\n",
    "\n",
    "filtered_data = pd.read_sql_query(\"\"\" SELECT * FROM Reviews WHERE Score != 3\"\"\", con) \n",
    "\n",
    "# Give reviews with Score>3 a positive rating(1), and reviews with a score<3 a negative rating(0).\n",
    "def partition(x):\n",
    "    if x < 3:\n",
    "        return 0\n",
    "    return 1\n",
    "\n",
    "#changing reviews with score less than 3 to be positive and vice-versa\n",
    "actualScore = filtered_data['Score']\n",
    "positiveNegative = actualScore.map(partition) \n",
    "filtered_data['Score'] = positiveNegative\n",
    "print(\"Number of data points in our data\", filtered_data.shape)\n",
    "filtered_data.head(3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "ZLasp3QxJkjS"
   },
   "outputs": [],
   "source": [
    "display = pd.read_sql_query(\"\"\"\n",
    "SELECT UserId, ProductId, ProfileName, Time, Score, Text, COUNT(*)\n",
    "FROM Reviews\n",
    "GROUP BY UserId\n",
    "HAVING COUNT(*)>1\n",
    "\"\"\", con)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "pHzk_lUkJkjW",
    "outputId": "3461f626-dc72-4b0d-af37-c3625e7e81b2",
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(80668, 7)\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>UserId</th>\n",
       "      <th>ProductId</th>\n",
       "      <th>ProfileName</th>\n",
       "      <th>Time</th>\n",
       "      <th>Score</th>\n",
       "      <th>Text</th>\n",
       "      <th>COUNT(*)</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>#oc-R115TNMSPFT9I7</td>\n",
       "      <td>B007Y59HVM</td>\n",
       "      <td>Breyton</td>\n",
       "      <td>1331510400</td>\n",
       "      <td>2</td>\n",
       "      <td>Overall its just OK when considering the price...</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>#oc-R11D9D7SHXIJB9</td>\n",
       "      <td>B005HG9ET0</td>\n",
       "      <td>Louis E. Emory \"hoppy\"</td>\n",
       "      <td>1342396800</td>\n",
       "      <td>5</td>\n",
       "      <td>My wife has recurring extreme muscle spasms, u...</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>#oc-R11DNU2NBKQ23Z</td>\n",
       "      <td>B007Y59HVM</td>\n",
       "      <td>Kim Cieszykowski</td>\n",
       "      <td>1348531200</td>\n",
       "      <td>1</td>\n",
       "      <td>This coffee is horrible and unfortunately not ...</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>#oc-R11O5J5ZVQE25C</td>\n",
       "      <td>B005HG9ET0</td>\n",
       "      <td>Penguin Chick</td>\n",
       "      <td>1346889600</td>\n",
       "      <td>5</td>\n",
       "      <td>This will be the bottle that you grab from the...</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>#oc-R12KPBODL2B5ZD</td>\n",
       "      <td>B007OSBE1U</td>\n",
       "      <td>Christopher P. Presta</td>\n",
       "      <td>1348617600</td>\n",
       "      <td>1</td>\n",
       "      <td>I didnt like this coffee. Instead of telling y...</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "               UserId   ProductId             ProfileName        Time  Score  \\\n",
       "0  #oc-R115TNMSPFT9I7  B007Y59HVM                 Breyton  1331510400      2   \n",
       "1  #oc-R11D9D7SHXIJB9  B005HG9ET0  Louis E. Emory \"hoppy\"  1342396800      5   \n",
       "2  #oc-R11DNU2NBKQ23Z  B007Y59HVM        Kim Cieszykowski  1348531200      1   \n",
       "3  #oc-R11O5J5ZVQE25C  B005HG9ET0           Penguin Chick  1346889600      5   \n",
       "4  #oc-R12KPBODL2B5ZD  B007OSBE1U   Christopher P. Presta  1348617600      1   \n",
       "\n",
       "                                                Text  COUNT(*)  \n",
       "0  Overall its just OK when considering the price...         2  \n",
       "1  My wife has recurring extreme muscle spasms, u...         3  \n",
       "2  This coffee is horrible and unfortunately not ...         2  \n",
       "3  This will be the bottle that you grab from the...         3  \n",
       "4  I didnt like this coffee. Instead of telling y...         2  "
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print(display.shape)\n",
    "display.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "HUsn2lg8Jkjb",
    "outputId": "76e2e98e-b728-46ba-d619-c2008134a3b7"
   },
   "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>UserId</th>\n",
       "      <th>ProductId</th>\n",
       "      <th>ProfileName</th>\n",
       "      <th>Time</th>\n",
       "      <th>Score</th>\n",
       "      <th>Text</th>\n",
       "      <th>COUNT(*)</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>80638</th>\n",
       "      <td>AZY10LLTJ71NX</td>\n",
       "      <td>B006P7E5ZI</td>\n",
       "      <td>undertheshrine \"undertheshrine\"</td>\n",
       "      <td>1334707200</td>\n",
       "      <td>5</td>\n",
       "      <td>I was recommended to try green tea extract to ...</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              UserId   ProductId                      ProfileName        Time  \\\n",
       "80638  AZY10LLTJ71NX  B006P7E5ZI  undertheshrine \"undertheshrine\"  1334707200   \n",
       "\n",
       "       Score                                               Text  COUNT(*)  \n",
       "80638      5  I was recommended to try green tea extract to ...         5  "
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "display[display['UserId']=='AZY10LLTJ71NX']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "CyPhzHZKJkjf",
    "outputId": "28b60f13-e2e6-4960-e5ba-3467e16dbd03"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "393063"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "display['COUNT(*)'].sum()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "y22fy_WtJkji"
   },
   "source": [
    "#  [2] Exploratory Data Analysis"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "ZqpW3tWPJkjl"
   },
   "source": [
    "## [2.1] Data Cleaning: Deduplication\n",
    "\n",
    "It is observed (as shown in the table below) that the reviews data had many duplicate entries. Hence it was necessary to remove duplicates in order to get unbiased results for the analysis of the data.  Following is an example:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "fmlu_oooJkjl",
    "outputId": "9ae82f65-ed65-425a-ad4d-b96091e37940",
    "scrolled": false
   },
   "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>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",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>78445</td>\n",
       "      <td>B000HDL1RQ</td>\n",
       "      <td>AR5J8UI46CURR</td>\n",
       "      <td>Geetha Krishnan</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>5</td>\n",
       "      <td>1199577600</td>\n",
       "      <td>LOACKER QUADRATINI VANILLA WAFERS</td>\n",
       "      <td>DELICIOUS WAFERS. I FIND THAT EUROPEAN WAFERS ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>138317</td>\n",
       "      <td>B000HDOPYC</td>\n",
       "      <td>AR5J8UI46CURR</td>\n",
       "      <td>Geetha Krishnan</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>5</td>\n",
       "      <td>1199577600</td>\n",
       "      <td>LOACKER QUADRATINI VANILLA WAFERS</td>\n",
       "      <td>DELICIOUS WAFERS. I FIND THAT EUROPEAN WAFERS ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>138277</td>\n",
       "      <td>B000HDOPYM</td>\n",
       "      <td>AR5J8UI46CURR</td>\n",
       "      <td>Geetha Krishnan</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>5</td>\n",
       "      <td>1199577600</td>\n",
       "      <td>LOACKER QUADRATINI VANILLA WAFERS</td>\n",
       "      <td>DELICIOUS WAFERS. I FIND THAT EUROPEAN WAFERS ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>73791</td>\n",
       "      <td>B000HDOPZG</td>\n",
       "      <td>AR5J8UI46CURR</td>\n",
       "      <td>Geetha Krishnan</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>5</td>\n",
       "      <td>1199577600</td>\n",
       "      <td>LOACKER QUADRATINI VANILLA WAFERS</td>\n",
       "      <td>DELICIOUS WAFERS. I FIND THAT EUROPEAN WAFERS ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>155049</td>\n",
       "      <td>B000PAQ75C</td>\n",
       "      <td>AR5J8UI46CURR</td>\n",
       "      <td>Geetha Krishnan</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>5</td>\n",
       "      <td>1199577600</td>\n",
       "      <td>LOACKER QUADRATINI VANILLA WAFERS</td>\n",
       "      <td>DELICIOUS WAFERS. I FIND THAT EUROPEAN WAFERS ...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       Id   ProductId         UserId      ProfileName  HelpfulnessNumerator  \\\n",
       "0   78445  B000HDL1RQ  AR5J8UI46CURR  Geetha Krishnan                     2   \n",
       "1  138317  B000HDOPYC  AR5J8UI46CURR  Geetha Krishnan                     2   \n",
       "2  138277  B000HDOPYM  AR5J8UI46CURR  Geetha Krishnan                     2   \n",
       "3   73791  B000HDOPZG  AR5J8UI46CURR  Geetha Krishnan                     2   \n",
       "4  155049  B000PAQ75C  AR5J8UI46CURR  Geetha Krishnan                     2   \n",
       "\n",
       "   HelpfulnessDenominator  Score        Time  \\\n",
       "0                       2      5  1199577600   \n",
       "1                       2      5  1199577600   \n",
       "2                       2      5  1199577600   \n",
       "3                       2      5  1199577600   \n",
       "4                       2      5  1199577600   \n",
       "\n",
       "                             Summary  \\\n",
       "0  LOACKER QUADRATINI VANILLA WAFERS   \n",
       "1  LOACKER QUADRATINI VANILLA WAFERS   \n",
       "2  LOACKER QUADRATINI VANILLA WAFERS   \n",
       "3  LOACKER QUADRATINI VANILLA WAFERS   \n",
       "4  LOACKER QUADRATINI VANILLA WAFERS   \n",
       "\n",
       "                                                Text  \n",
       "0  DELICIOUS WAFERS. I FIND THAT EUROPEAN WAFERS ...  \n",
       "1  DELICIOUS WAFERS. I FIND THAT EUROPEAN WAFERS ...  \n",
       "2  DELICIOUS WAFERS. I FIND THAT EUROPEAN WAFERS ...  \n",
       "3  DELICIOUS WAFERS. I FIND THAT EUROPEAN WAFERS ...  \n",
       "4  DELICIOUS WAFERS. I FIND THAT EUROPEAN WAFERS ...  "
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "display= pd.read_sql_query(\"\"\"\n",
    "SELECT *\n",
    "FROM Reviews\n",
    "WHERE Score != 3 AND UserId=\"AR5J8UI46CURR\"\n",
    "ORDER BY ProductID\n",
    "\"\"\", con)\n",
    "display.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "pvogJHEyJkjq"
   },
   "source": [
    "As it can be seen above that same user has multiple reviews with same values for HelpfulnessNumerator, HelpfulnessDenominator, Score, Time, Summary and Text and on doing analysis it was found that <br>\n",
    "<br> \n",
    "ProductId=B000HDOPZG was Loacker Quadratini Vanilla Wafer Cookies, 8.82-Ounce Packages (Pack of 8)<br>\n",
    "<br> \n",
    "ProductId=B000HDL1RQ was Loacker Quadratini Lemon Wafer Cookies, 8.82-Ounce Packages (Pack of 8) and so on<br>\n",
    "\n",
    "It was inferred after analysis that reviews with same parameters other than ProductId belonged to the same product just having different flavour or quantity. Hence in order to reduce redundancy it was decided to eliminate the rows having same parameters.<br>\n",
    "\n",
    "The method used for the same was that we first sort the data according to ProductId and then just keep the first similar product review and delelte the others. for eg. in the above just the review for ProductId=B000HDL1RQ remains. This method ensures that there is only one representative for each product and deduplication without sorting would lead to possibility of different representatives still existing for the same product."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "LAWOrfBaJkjr"
   },
   "outputs": [],
   "source": [
    "#Sorting data according to ProductId in ascending order\n",
    "sorted_data=filtered_data.sort_values('ProductId', axis=0, ascending=True, inplace=False, kind='quicksort', na_position='last')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "dfaQob28Jkju",
    "outputId": "21104e4f-cd18-4353-a54a-5b4638fc552b"
   },
   "outputs": [],
   "source": [
    "#Deduplication of entries\n",
    "final=sorted_data.drop_duplicates(subset={\"UserId\",\"ProfileName\",\"Time\",\"Text\"}, keep='first', inplace=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "9QVxT6oxJkjx",
    "outputId": "f292b9fe-6065-4979-830d-c79bc6785a3d"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "69.25890143662969"
      ]
     },
     "execution_count": 50,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Checking to see how much % of data still remains\n",
    "(final['Id'].size*1.0)/(filtered_data['Id'].size*1.0)*100"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "qVaxONHLJkj1"
   },
   "source": [
    "<b>Observation:-</b> It was also seen that in two rows given below the value of HelpfulnessNumerator is greater than HelpfulnessDenominator which is not practically possible hence these two rows too are removed from calcualtions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "vLoqfOjcJkj1",
    "outputId": "db3448f8-fbda-4205-a7cc-5048e0f1fd19"
   },
   "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>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",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>64422</td>\n",
       "      <td>B000MIDROQ</td>\n",
       "      <td>A161DK06JJMCYF</td>\n",
       "      <td>J. E. Stephens \"Jeanne\"</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>5</td>\n",
       "      <td>1224892800</td>\n",
       "      <td>Bought This for My Son at College</td>\n",
       "      <td>My son loves spaghetti so I didn't hesitate or...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>44737</td>\n",
       "      <td>B001EQ55RW</td>\n",
       "      <td>A2V0I904FH7ABY</td>\n",
       "      <td>Ram</td>\n",
       "      <td>3</td>\n",
       "      <td>2</td>\n",
       "      <td>4</td>\n",
       "      <td>1212883200</td>\n",
       "      <td>Pure cocoa taste with crunchy almonds inside</td>\n",
       "      <td>It was almost a 'love at first bite' - the per...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      Id   ProductId          UserId              ProfileName  \\\n",
       "0  64422  B000MIDROQ  A161DK06JJMCYF  J. E. Stephens \"Jeanne\"   \n",
       "1  44737  B001EQ55RW  A2V0I904FH7ABY                      Ram   \n",
       "\n",
       "   HelpfulnessNumerator  HelpfulnessDenominator  Score        Time  \\\n",
       "0                     3                       1      5  1224892800   \n",
       "1                     3                       2      4  1212883200   \n",
       "\n",
       "                                        Summary  \\\n",
       "0             Bought This for My Son at College   \n",
       "1  Pure cocoa taste with crunchy almonds inside   \n",
       "\n",
       "                                                Text  \n",
       "0  My son loves spaghetti so I didn't hesitate or...  \n",
       "1  It was almost a 'love at first bite' - the per...  "
      ]
     },
     "execution_count": 51,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "display= pd.read_sql_query(\"\"\"\n",
    "SELECT *\n",
    "FROM Reviews\n",
    "WHERE Score != 3 AND Id=44737 OR Id=64422\n",
    "ORDER BY ProductID\n",
    "\"\"\", con)\n",
    "\n",
    "display.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "f_185ozFJkj9"
   },
   "outputs": [],
   "source": [
    "final=final[final.HelpfulnessNumerator<=final.HelpfulnessDenominator]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "CofRySJyJkkC",
    "outputId": "a06c52f6-66a9-429b-8a5e-e20259f0bb21"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(100000, 10)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "1    84328\n",
       "0    15672\n",
       "Name: Score, dtype: int64"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Before starting the next phase of preprocessing lets see the number of entries left\n",
    "final = sorted_data.sample(n=100000)\n",
    "final.shape\n",
    "\n",
    "print(final.shape)\n",
    "\n",
    "#How many positive and negative reviews are present in our dataset?\n",
    "final['Score'].value_counts()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "-fEWeywxJkkI"
   },
   "source": [
    "#  [3] Preprocessing"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "tzpbt9aOJkkJ"
   },
   "source": [
    "## [3.1].  Preprocessing Review Text\n",
    "\n",
    "Now that we have finished deduplication our data requires some preprocessing before we go on further with analysis and making the prediction model.\n",
    "\n",
    "Hence in the Preprocessing phase we do the following in the order below:-\n",
    "\n",
    "1. Begin by removing the html tags\n",
    "2. Remove any punctuations or limited set of special characters like , or . or # etc.\n",
    "3. Check if the word is made up of english letters and is not alpha-numeric\n",
    "4. Check to see if the length of the word is greater than 2 (as it was researched that there is no adjective in 2-letters)\n",
    "5. Convert the word to lowercase\n",
    "6. Remove Stopwords\n",
    "7. Finally Snowball Stemming the word (it was obsereved to be better than Porter Stemming)<br>\n",
    "\n",
    "After which we collect the words used to describe positive and negative reviews"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "YqxBeLk8JkkL",
    "outputId": "1819632d-0129-4a13-d7ff-4e7bbf129dff"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "I love this stuff.  I'd been drinking Twinings English Breakfast for years.  When my Mom asked for Christmas gift ideas last year, I mentioned loose tea and told her that was my favorite.  I opened gifts on Christmas morning to find my first bag of Taylors of Harrogate Yorkshire Tea -- she hadn't been able to find Twinings loose leaf.<br /><br />I make it using an individual french press mug set (a Christmas gift from my brother a couple of years ago!), and it's the best tea I've ever had.  I liked it so much I didn't even think twice about ordering the box of 6...and I may be asking for more by the time Christmas rolls around again!\n",
      "==================================================\n",
      "I own an espresso machine and love coffee. However the machine is a pain to use when I'm in a rush to work in the morning, which is most mornings.  I was thinking about getting a Keurig pod machine to make my mornings easier, however I didn't want to have yet another appliance sitting on my counter.  Then I remembered that my espresso maker came with an ESE (easy serve espresso) pod adapter so I decided to give these pods a try based on the good reviews.<br /><br />Although clean up is a lot easier, the coffee pods are a bit weak in flavor.  I make Americanos in the morning, and one pod in 8oz of water is not quite strong enough for me.  Two pods is perfect.  I will no doubt need to use 4 pods for my travel mug to get the bold flavor that I desire.  Overall, the coffee is good and cleanup is easy.\n",
      "==================================================\n",
      "This tea has the perfect amount of mint.  Not too strong, not too bitter.  I liked this tea so much, I've started buying the tea bags.\n",
      "==================================================\n",
      "I've had several cats, most of which didn't have much of a hairball problem.  However, the cat I have now, has hocked up some epic hairballs. She's the best cat I've ever had, except for this one nasty problem.<br /><br />So, I bought a small bag of this hairball control from my vet, tried it, and it does seem to have had an effect on my cat's massive hairball problem.  In fact, after about a month of mixing this with her other food, I've noticed that she's only thrown up once or twice, and that's certainly not as much as she did before.<br /><br />I found this bag on Amazon for quite a bit less than what my vet charged, so I'll continue to order it here. Based on what I've observed with my cat so far, I would highly recommend this for those of you who have cats with the same problem.\n",
      "==================================================\n"
     ]
    }
   ],
   "source": [
    "# printing some random reviews\n",
    "sent_0 = final['Text'].values[0]\n",
    "print(sent_0)\n",
    "print(\"=\"*50)\n",
    "\n",
    "sent_1000 = final['Text'].values[1000]\n",
    "print(sent_1000)\n",
    "print(\"=\"*50)\n",
    "\n",
    "sent_1500 = final['Text'].values[1500]\n",
    "print(sent_1500)\n",
    "print(\"=\"*50)\n",
    "\n",
    "sent_4900 = final['Text'].values[4900]\n",
    "print(sent_4900)\n",
    "print(\"=\"*50)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "lWiJ2_9gJkkP",
    "outputId": "08234806-3657-4d8f-c6d2-020dc9cda61c"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "I love this stuff.  I'd been drinking Twinings English Breakfast for years.  When my Mom asked for Christmas gift ideas last year, I mentioned loose tea and told her that was my favorite.  I opened gifts on Christmas morning to find my first bag of Taylors of Harrogate Yorkshire Tea -- she hadn't been able to find Twinings loose leaf.<br /><br />I make it using an individual french press mug set (a Christmas gift from my brother a couple of years ago!), and it's the best tea I've ever had.  I liked it so much I didn't even think twice about ordering the box of 6...and I may be asking for more by the time Christmas rolls around again!\n"
     ]
    }
   ],
   "source": [
    "# remove urls from text python: https://stackoverflow.com/a/40823105/4084039\n",
    "sent_0 = re.sub(r\"http\\S+\", \"\", sent_0)\n",
    "sent_1000 = re.sub(r\"http\\S+\", \"\", sent_1000)\n",
    "sent_150 = re.sub(r\"http\\S+\", \"\", sent_1500)\n",
    "sent_4900 = re.sub(r\"http\\S+\", \"\", sent_4900)\n",
    "\n",
    "print(sent_0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "cpLtZ-lAJkkR",
    "outputId": "7e0bf159-4db4-4ffe-d72e-6d5952d453f1"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "I love this stuff.  I'd been drinking Twinings English Breakfast for years.  When my Mom asked for Christmas gift ideas last year, I mentioned loose tea and told her that was my favorite.  I opened gifts on Christmas morning to find my first bag of Taylors of Harrogate Yorkshire Tea -- she hadn't been able to find Twinings loose leaf.I make it using an individual french press mug set (a Christmas gift from my brother a couple of years ago!), and it's the best tea I've ever had.  I liked it so much I didn't even think twice about ordering the box of 6...and I may be asking for more by the time Christmas rolls around again!\n",
      "==================================================\n",
      "I own an espresso machine and love coffee. However the machine is a pain to use when I'm in a rush to work in the morning, which is most mornings.  I was thinking about getting a Keurig pod machine to make my mornings easier, however I didn't want to have yet another appliance sitting on my counter.  Then I remembered that my espresso maker came with an ESE (easy serve espresso) pod adapter so I decided to give these pods a try based on the good reviews.Although clean up is a lot easier, the coffee pods are a bit weak in flavor.  I make Americanos in the morning, and one pod in 8oz of water is not quite strong enough for me.  Two pods is perfect.  I will no doubt need to use 4 pods for my travel mug to get the bold flavor that I desire.  Overall, the coffee is good and cleanup is easy.\n",
      "==================================================\n",
      "This tea has the perfect amount of mint.  Not too strong, not too bitter.  I liked this tea so much, I've started buying the tea bags.\n",
      "==================================================\n",
      "I've had several cats, most of which didn't have much of a hairball problem.  However, the cat I have now, has hocked up some epic hairballs. She's the best cat I've ever had, except for this one nasty problem.So, I bought a small bag of this hairball control from my vet, tried it, and it does seem to have had an effect on my cat's massive hairball problem.  In fact, after about a month of mixing this with her other food, I've noticed that she's only thrown up once or twice, and that's certainly not as much as she did before.I found this bag on Amazon for quite a bit less than what my vet charged, so I'll continue to order it here. Based on what I've observed with my cat so far, I would highly recommend this for those of you who have cats with the same problem.\n"
     ]
    }
   ],
   "source": [
    "# https://stackoverflow.com/questions/16206380/python-beautifulsoup-how-to-remove-all-tags-from-an-element\n",
    "from bs4 import BeautifulSoup\n",
    "\n",
    "soup = BeautifulSoup(sent_0, 'lxml')\n",
    "text = soup.get_text()\n",
    "print(text)\n",
    "print(\"=\"*50)\n",
    "\n",
    "soup = BeautifulSoup(sent_1000, 'lxml')\n",
    "text = soup.get_text()\n",
    "print(text)\n",
    "print(\"=\"*50)\n",
    "\n",
    "soup = BeautifulSoup(sent_1500, 'lxml')\n",
    "text = soup.get_text()\n",
    "print(text)\n",
    "print(\"=\"*50)\n",
    "\n",
    "soup = BeautifulSoup(sent_4900, 'lxml')\n",
    "text = soup.get_text()\n",
    "print(text)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "HYBbtiCMJkkW"
   },
   "outputs": [],
   "source": [
    "# https://stackoverflow.com/a/47091490/4084039\n",
    "import re\n",
    "\n",
    "def decontracted(phrase):\n",
    "    # specific\n",
    "    phrase = re.sub(r\"won't\", \"will not\", phrase)\n",
    "    phrase = re.sub(r\"can\\'t\", \"can not\", phrase)\n",
    "\n",
    "    # general\n",
    "    phrase = re.sub(r\"n\\'t\", \" not\", phrase)\n",
    "    phrase = re.sub(r\"\\'re\", \" are\", phrase)\n",
    "    phrase = re.sub(r\"\\'s\", \" is\", phrase)\n",
    "    phrase = re.sub(r\"\\'d\", \" would\", phrase)\n",
    "    phrase = re.sub(r\"\\'ll\", \" will\", phrase)\n",
    "    phrase = re.sub(r\"\\'t\", \" not\", phrase)\n",
    "    phrase = re.sub(r\"\\'ve\", \" have\", phrase)\n",
    "    phrase = re.sub(r\"\\'m\", \" am\", phrase)\n",
    "    return phrase"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "WPQSZurTJkkY",
    "outputId": "6d91fc9f-8f77-497c-95bb-96bfc9781196"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "I own an espresso machine and love coffee. However the machine is a pain to use when I am in a rush to work in the morning, which is most mornings.  I was thinking about getting a Keurig pod machine to make my mornings easier, however I did not want to have yet another appliance sitting on my counter.  Then I remembered that my espresso maker came with an ESE (easy serve espresso) pod adapter so I decided to give these pods a try based on the good reviews.<br /><br />Although clean up is a lot easier, the coffee pods are a bit weak in flavor.  I make Americanos in the morning, and one pod in 8oz of water is not quite strong enough for me.  Two pods is perfect.  I will no doubt need to use 4 pods for my travel mug to get the bold flavor that I desire.  Overall, the coffee is good and cleanup is easy.\n",
      "==================================================\n"
     ]
    }
   ],
   "source": [
    "sent_1000 = decontracted(sent_1000)\n",
    "print(sent_1000)\n",
    "print(\"=\"*50)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "6bakeyaQJkkb",
    "outputId": "f7cf0096-0b7b-4ec1-84df-01e838621d62"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "My two cats somehow never figured out how to use their teeth to bite into their food--given dry food, they'll just use their tongues to lick it up and swallow the kernels whole, with an occasional bite with the teeth to crunch on food that's already been delivered into their mouths.  For a while I despaired of getting them to eat wet food, as they would just sit there and lick up all the juices, without getting much of the meat.  Typical canned food is just packed too tight, and the texture is not amenable for breaking off little bits with just their tongues.  Finally, I tried out Spot's Stew, which solved the problem.  This food is much looser, I think due to the inclusion of peas and other veggies, and the meat part is just the right consistency where they can lick it up successfully.  At first the cats would eat around the veggies, but they now lick up every morsel, and even prefer the Spot's Stew over the Wellness pouches.  I also like that the food doesn't contain any fish, as one of my cats is allergic to something or other (his back legs are always itchy, and he'll pull out the hair), and the vet suggested it might be fish.  Anyways, thank goodness for Spot's Stew, otherwise my teeth-challenged cats would never have been able to enjoy wet food!\n"
     ]
    }
   ],
   "source": [
    "#remove words with numbers python: https://stackoverflow.com/a/18082370/4084039\n",
    "sent_0 = re.sub(\"\\S*\\d\\S*\", \"\", sent_0).strip()\n",
    "print(sent_0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "dMqqVqadJkkg",
    "outputId": "e4a6bd28-9d39-4dd2-9483-0406dd3dff50"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "I wish I had paid more attention to the reviews This product is pricey for what you get the little burned piece of cheese on the pan from some cheese dish It is so salty and so not worth it It s not a taste treat It is 1 carb of never mind Do not waste your money or taste buds on this \n"
     ]
    }
   ],
   "source": [
    "#remove spacial character: https://stackoverflow.com/a/5843547/4084039\n",
    "sent_1500 = re.sub('[^A-Za-z0-9]+', ' ', sent_1500)\n",
    "print(sent_1500)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "bvEoXB5NJkkk"
   },
   "outputs": [],
   "source": [
    "# https://gist.github.com/sebleier/554280\n",
    "# we are removing the words from the stop words list: 'no', 'nor', 'not'\n",
    "# <br /><br /> ==> after the above steps, we are getting \"br br\"\n",
    "# we are including them into stop words list\n",
    "# instead of <br /> if we have <br/> these tags would have revmoved in the 1st step\n",
    "\n",
    "stopwords= set(['br', 'the', 'i', 'me', 'my', 'myself', 'we', 'our', 'ours', 'ourselves', 'you', \"you're\", \"you've\",\\\n",
    "            \"you'll\", \"you'd\", 'your', 'yours', 'yourself', 'yourselves', 'he', 'him', 'his', 'himself', \\\n",
    "            'she', \"she's\", 'her', 'hers', 'herself', 'it', \"it's\", 'its', 'itself', 'they', 'them', 'their',\\\n",
    "            'theirs', 'themselves', 'what', 'which', 'who', 'whom', 'this', 'that', \"that'll\", 'these', 'those', \\\n",
    "            'am', 'is', 'are', 'was', 'were', 'be', 'been', 'being', 'have', 'has', 'had', 'having', 'do', 'does', \\\n",
    "            'did', 'doing', 'a', 'an', 'the', 'and', 'but', 'if', 'or', 'because', 'as', 'until', 'while', 'of', \\\n",
    "            'at', 'by', 'for', 'with', 'about', 'against', 'between', 'into', 'through', 'during', 'before', 'after',\\\n",
    "            'above', 'below', 'to', 'from', 'up', 'down', 'in', 'out', 'on', 'off', 'over', 'under', 'again', 'further',\\\n",
    "            'then', 'once', 'here', 'there', 'when', 'where', 'why', 'how', 'all', 'any', 'both', 'each', 'few', 'more',\\\n",
    "            'most', 'other', 'some', 'such', 'only', 'own', 'same', 'so', 'than', 'too', 'very', \\\n",
    "            's', 't', 'can', 'will', 'just', 'don', \"don't\", 'should', \"should've\", 'now', 'd', 'll', 'm', 'o', 're', \\\n",
    "            've', 'y', 'ain', 'aren', \"aren't\", 'couldn', \"couldn't\", 'didn', \"didn't\", 'doesn', \"doesn't\", 'hadn',\\\n",
    "            \"hadn't\", 'hasn', \"hasn't\", 'haven', \"haven't\", 'isn', \"isn't\", 'ma', 'mightn', \"mightn't\", 'mustn',\\\n",
    "            \"mustn't\", 'needn', \"needn't\", 'shan', \"shan't\", 'shouldn', \"shouldn't\", 'wasn', \"wasn't\", 'weren', \"weren't\", \\\n",
    "            'won', \"won't\", 'wouldn', \"wouldn't\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "0j3jfqWCJkko",
    "outputId": "c83c59c5-fe17-4ad2-d43f-0b0947a6e6ba"
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|████████████████████████████████████████████████████████████████████████| 100000/100000 [00:53<00:00, 1866.26it/s]\n",
      "100%|████████████████████████████████████████████████████████████████████████| 100000/100000 [00:38<00:00, 2582.66it/s]\n"
     ]
    }
   ],
   "source": [
    "# Combining all the above stundents \n",
    "from tqdm import tqdm\n",
    "preprocessed_reviews = []\n",
    "# tqdm is for printing the status bar\n",
    "for sentance in tqdm(final['Text'].values):\n",
    "    sentance = re.sub(r\"http\\S+\", \"\", sentance)\n",
    "    sentance = BeautifulSoup(sentance, 'lxml').get_text()\n",
    "    sentance = decontracted(sentance)\n",
    "    sentance = re.sub(\"\\S*\\d\\S*\", \"\", sentance).strip()\n",
    "    sentance = re.sub('[^A-Za-z]+', ' ', sentance)\n",
    "    # https://gist.github.com/sebleier/554280\n",
    "    sentance = ' '.join(e.lower() for e in sentance.split() if e.lower() not in stopwords)\n",
    "    preprocessed_reviews.append(sentance.strip())\n",
    "    \n",
    "preprocessed_summary = [] \n",
    "for sentence in tqdm(final['Summary'].values):\n",
    "    sentence = re.sub(r\"http\\S+\", \"\", sentence)\n",
    "    sentence = BeautifulSoup(sentence, 'lxml').get_text()\n",
    "    sentence = decontracted(sentence)\n",
    "    sentence = re.sub(\"\\S*\\d\\S*\", \"\", sentence).strip()\n",
    "    sentence = re.sub('[^A-Za-z]+', ' ', sentence)\n",
    "    # https://gist.github.com/sebleier/554280\n",
    "    sentence = ' '.join(e.lower() for e in sentence.split() if e.lower() not in stopwords)\n",
    "    preprocessed_summary.append(sentence.strip())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "nS_7EeV7Jkkr",
    "outputId": "4d4cd05e-0730-4e52-dc30-eb97014871dd"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "tea perfect amount mint not strong not bitter liked tea much started buying tea bags\n"
     ]
    }
   ],
   "source": [
    "print(preprocessed_reviews[1500])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "zak5yP5zJkkt"
   },
   "source": [
    "<h2><font color='red'>[3.2] Preprocessing Review Summary</font></h2>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "wZ3Epk1wJkku"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "great mint flavor\n"
     ]
    }
   ],
   "source": [
    "## Similartly you can do preprocessing for review summary also.\n",
    "print(preprocessed_summary[1500])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "## Take length of each review and use it as a feature\n",
    "review_len = [len(i) for i in preprocessed_reviews]\n",
    "#Add summary text and review text to extract as much information as possible\n",
    "# interlist element concatenation \n",
    "combined_text = [i +' '+ j for i, j in zip(preprocessed_reviews, preprocessed_summary)] \n",
    "\n",
    "len(review_len)\n",
    "\n",
    "final['combined_text'] = combined_text\n",
    "final['review_length'] = review_len"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "ChHwtw1iJkky"
   },
   "source": [
    "# [4] Featurization"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "qyWESNvzJkky"
   },
   "source": [
    "## [4.1] BAG OF WORDS"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "zS0niozAJkkz",
    "outputId": "e2a25e5d-2e1a-497a-f094-a7e249201001"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "some feature names  ['aa', 'aahhhs', 'aback', 'abandon', 'abates', 'abbott', 'abby', 'abdominal', 'abiding', 'ability']\n",
      "==================================================\n",
      "the type of count vectorizer  <class 'scipy.sparse.csr.csr_matrix'>\n",
      "the shape of out text BOW vectorizer  (4986, 12997)\n",
      "the number of unique words  12997\n"
     ]
    }
   ],
   "source": [
    "#BoW\n",
    "count_vect = CountVectorizer() #in scikit-learn\n",
    "count_vect.fit(preprocessed_reviews)\n",
    "print(\"some feature names \", count_vect.get_feature_names()[:10])\n",
    "print('='*50)\n",
    "\n",
    "final_counts = count_vect.transform(preprocessed_reviews)\n",
    "print(\"the type of count vectorizer \",type(final_counts))\n",
    "print(\"the shape of out text BOW vectorizer \",final_counts.get_shape())\n",
    "print(\"the number of unique words \", final_counts.get_shape()[1])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "LChtHtLjJkk1"
   },
   "source": [
    "## [4.2] Bi-Grams and n-Grams."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "Xgw2pSsoJkk3",
    "outputId": "bfcb230d-1f75-413a-e30d-c2aae84cd5c0"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "the type of count vectorizer  <class 'scipy.sparse.csr.csr_matrix'>\n",
      "the shape of out text BOW vectorizer  (4986, 3144)\n",
      "the number of unique words including both unigrams and bigrams  3144\n"
     ]
    }
   ],
   "source": [
    "#bi-gram, tri-gram and n-gram\n",
    "\n",
    "#removing stop words like \"not\" should be avoided before building n-grams\n",
    "# count_vect = CountVectorizer(ngram_range=(1,2))\n",
    "# please do read the CountVectorizer documentation http://scikit-learn.org/stable/modules/generated/sklearn.feature_extraction.text.CountVectorizer.html\n",
    "\n",
    "# you can choose these numebrs min_df=10, max_features=5000, of your choice\n",
    "count_vect = CountVectorizer(ngram_range=(1,2), min_df=10, max_features=5000)\n",
    "final_bigram_counts = count_vect.fit_transform(preprocessed_reviews)\n",
    "print(\"the type of count vectorizer \",type(final_bigram_counts))\n",
    "print(\"the shape of out text BOW vectorizer \",final_bigram_counts.get_shape())\n",
    "print(\"the number of unique words including both unigrams and bigrams \", final_bigram_counts.get_shape()[1])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "1fvAmu91Jkk6"
   },
   "source": [
    "## [4.3] TF-IDF"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "ag5-WCR8Jkk7",
    "outputId": "94a8ddc3-4bb7-4c49-d4e7-ed98560633b6"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "some sample features(unique words in the corpus) ['ability', 'able', 'able find', 'able get', 'absolute', 'absolutely', 'absolutely delicious', 'absolutely love', 'absolutely no', 'according']\n",
      "==================================================\n",
      "the type of count vectorizer  <class 'scipy.sparse.csr.csr_matrix'>\n",
      "the shape of out text TFIDF vectorizer  (4986, 3144)\n",
      "the number of unique words including both unigrams and bigrams  3144\n"
     ]
    }
   ],
   "source": [
    "tf_idf_vect = TfidfVectorizer(ngram_range=(1,2), min_df=10)\n",
    "tf_idf_vect.fit(preprocessed_reviews)\n",
    "print(\"some sample features(unique words in the corpus)\",tf_idf_vect.get_feature_names()[0:10])\n",
    "print('='*50)\n",
    "\n",
    "final_tf_idf = tf_idf_vect.transform(preprocessed_reviews)\n",
    "print(\"the type of count vectorizer \",type(final_tf_idf))\n",
    "print(\"the shape of out text TFIDF vectorizer \",final_tf_idf.get_shape())\n",
    "print(\"the number of unique words including both unigrams and bigrams \", final_tf_idf.get_shape()[1])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "_jVNRd5hJkk-"
   },
   "source": [
    "## [4.4] Word2Vec"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "8jGeY1r5Jkk_"
   },
   "outputs": [],
   "source": [
    "# Train your own Word2Vec model using your own text corpus\n",
    "i=0\n",
    "list_of_sentance=[]\n",
    "for sentance in preprocessed_reviews:\n",
    "    list_of_sentance.append(sentance.split())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "2X9M3ybMJklB",
    "outputId": "42dc6662-8d17-41ce-be7c-12e2db687f30",
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[('snack', 0.9951335191726685), ('calorie', 0.9946465492248535), ('wonderful', 0.9946032166481018), ('excellent', 0.9944332838058472), ('especially', 0.9941144585609436), ('baked', 0.9940600395202637), ('salted', 0.994047224521637), ('alternative', 0.9937226176261902), ('tasty', 0.9936816692352295), ('healthy', 0.9936649799346924)]\n",
      "==================================================\n",
      "[('varieties', 0.9994194507598877), ('become', 0.9992934465408325), ('popcorn', 0.9992750883102417), ('de', 0.9992610216140747), ('miss', 0.9992451071739197), ('melitta', 0.999218761920929), ('choice', 0.9992102384567261), ('american', 0.9991837739944458), ('beef', 0.9991780519485474), ('finish', 0.9991567134857178)]\n"
     ]
    }
   ],
   "source": [
    "# Using Google News Word2Vectors\n",
    "\n",
    "# in this project we are using a pretrained model by google\n",
    "# its 3.3G file, once you load this into your memory \n",
    "# it occupies ~9Gb, so please do this step only if you have >12G of ram\n",
    "# we will provide a pickle file wich contains a dict , \n",
    "# and it contains all our courpus words as keys and  model[word] as values\n",
    "# To use this code-snippet, download \"GoogleNews-vectors-negative300.bin\" \n",
    "# from https://drive.google.com/file/d/0B7XkCwpI5KDYNlNUTTlSS21pQmM/edit\n",
    "# it's 1.9GB in size.\n",
    "\n",
    "\n",
    "# http://kavita-ganesan.com/gensim-word2vec-tutorial-starter-code/#.W17SRFAzZPY\n",
    "# you can comment this whole cell\n",
    "# or change these varible according to your need\n",
    "\n",
    "is_your_ram_gt_16g=False\n",
    "want_to_use_google_w2v = False\n",
    "want_to_train_w2v = True\n",
    "\n",
    "if want_to_train_w2v:\n",
    "    # min_count = 5 considers only words that occured atleast 5 times\n",
    "    w2v_model=Word2Vec(list_of_sentance,min_count=5,size=50, workers=4)\n",
    "    print(w2v_model.wv.most_similar('great'))\n",
    "    print('='*50)\n",
    "    print(w2v_model.wv.most_similar('worst'))\n",
    "    \n",
    "elif want_to_use_google_w2v and is_your_ram_gt_16g:\n",
    "    if os.path.isfile('GoogleNews-vectors-negative300.bin'):\n",
    "        w2v_model=KeyedVectors.load_word2vec_format('GoogleNews-vectors-negative300.bin', binary=True)\n",
    "        print(w2v_model.wv.most_similar('great'))\n",
    "        print(w2v_model.wv.most_similar('worst'))\n",
    "    else:\n",
    "        print(\"you don't have gogole's word2vec file, keep want_to_train_w2v = True, to train your own w2v \")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "jIFbq6sDJklG",
    "outputId": "56320a9c-85c4-4f09-c139-0a64bfdc6ab2"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "number of words that occured minimum 5 times  3817\n",
      "sample words  ['product', 'available', 'course', 'total', 'pretty', 'stinky', 'right', 'nearby', 'used', 'ca', 'not', 'beat', 'great', 'received', 'shipment', 'could', 'hardly', 'wait', 'try', 'love', 'call', 'instead', 'removed', 'easily', 'daughter', 'designed', 'printed', 'use', 'car', 'windows', 'beautifully', 'shop', 'program', 'going', 'lot', 'fun', 'everywhere', 'like', 'tv', 'computer', 'really', 'good', 'idea', 'final', 'outstanding', 'window', 'everybody', 'asks', 'bought', 'made']\n"
     ]
    }
   ],
   "source": [
    "w2v_words = list(w2v_model.wv.vocab)\n",
    "print(\"number of words that occured minimum 5 times \",len(w2v_words))\n",
    "print(\"sample words \", w2v_words[0:50])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "bJ1C1s_eJklQ"
   },
   "source": [
    "## [4.4.1] Converting text into vectors using Avg W2V, TFIDF-W2V"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "33z1pU7zJklQ"
   },
   "source": [
    "#### [4.4.1.1] Avg W2v"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "Ce1_4JvaJklR",
    "outputId": "198ff07c-ab57-4395-bf04-4d43a9fa1205"
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|████████████████████████████████████████████████████████████████████████████| 4986/4986 [00:03<00:00, 1330.47it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "4986\n",
      "50\n"
     ]
    }
   ],
   "source": [
    "# average Word2Vec\n",
    "# compute average word2vec for each review.\n",
    "sent_vectors = []; # the avg-w2v for each sentence/review is stored in this list\n",
    "for sent in tqdm(list_of_sentance): # 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",
    "    sent_vectors.append(sent_vec)\n",
    "print(len(sent_vectors))\n",
    "print(len(sent_vectors[0]))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "-R_AYTtwJklT"
   },
   "source": [
    "#### [4.4.1.2] TFIDF weighted W2v"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "il5ldoFyJklU"
   },
   "outputs": [],
   "source": [
    "# S = [\"abc def pqr\", \"def def def abc\", \"pqr pqr def\"]\n",
    "model = TfidfVectorizer()\n",
    "tf_idf_matrix = model.fit_transform(preprocessed_reviews)\n",
    "# we are converting a dictionary with word as a key, and the idf as a value\n",
    "dictionary = dict(zip(model.get_feature_names(), list(model.idf_)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "VUZMBVxpJklX",
    "outputId": "e826680b-3c3f-4c22-9de3-c7850f5ff737",
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|█████████████████████████████████████████████████████████████████████████████| 4986/4986 [00:20<00:00, 245.63it/s]\n"
     ]
    }
   ],
   "source": [
    "# TF-IDF weighted Word2Vec\n",
    "tfidf_feat = model.get_feature_names() # tfidf words/col-names\n",
    "# final_tf_idf is the sparse matrix with row= sentence, col=word and cell_val = tfidf\n",
    "\n",
    "tfidf_sent_vectors = []; # the tfidf-w2v for each sentence/review is stored in this list\n",
    "row=0;\n",
    "for sent in tqdm(list_of_sentance): # 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_feat:\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_vectors.append(sent_vec)\n",
    "    row += 1"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "ZFN0k_hEJklb"
   },
   "source": [
    "# [5] Assignment 4: Apply Naive Bayes"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "O1ABSoejJklc"
   },
   "source": [
    "<ol>\n",
    "    <li><strong>Apply Multinomial NaiveBayes 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",
    "        </ul>\n",
    "    </li>\n",
    "    <br>\n",
    "    <li><strong>The hyper paramter tuning(find best Alpha)</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>Consider a wide range of alpha values for hyperparameter tuning, start as low as 0.00001</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>Feature importance</strong>\n",
    "        <ul>\n",
    "    <li>Find the top 10 features of positive class and top 10 features of negative class for both feature sets <font color='red'>Set 1</font> and <font color='red'>Set 2</font> using values of `feature_log_prob_` parameter of  <a href='https://scikit-learn.org/stable/modules/generated/sklearn.naive_bayes.MultinomialNB.html'>MultinomialNB</a> and print their corresponding feature names</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. Here on X-axis you will have alpha values, since they have a wide range, just to represent those alpha values on the graph, apply log function on those alpha values.\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": "YawkDx8TJklc"
   },
   "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": {
    "colab_type": "text",
    "id": "RGRFxjhgJklc"
   },
   "source": [
    "# Applying Multinomial Naive Bayes"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "FxGKHpZlJkle"
   },
   "source": [
    "## [5.1] Applying Naive Bayes on BOW,<font color='red'> SET 1</font>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "UR0f8SfZJkle"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "            Id   ProductId          UserId     ProfileName  \\\n",
      "147596  160143  B000F3WSFC  A3IJ4RQXTHQ56C              MJ   \n",
      "61747    67080  B0006348H2  A3EAP0PTGE5KBQ   wvrivergirl40   \n",
      "347537  375927  B0000DBN1O  A1IU7S4HCK1XK0  Joanna Daneman   \n",
      "\n",
      "        HelpfulnessNumerator  HelpfulnessDenominator  Score        Time  \\\n",
      "147596                     0                       0      1  1299196800   \n",
      "61747                      1                       1      1  1311811200   \n",
      "347537                     7                       7      1  1072656000   \n",
      "\n",
      "                                           Summary  \\\n",
      "147596           Makes a Lovely Cup of Tea, Indeed   \n",
      "61747   Pet will eat it if you mix with their food   \n",
      "347537                              Earl Grey, Hot   \n",
      "\n",
      "                                                     Text  \\\n",
      "147596  I love this stuff.  I'd been drinking Twinings...   \n",
      "61747   It grows really fast, but my dog and cat will ...   \n",
      "347537  This is a strongly aromatic version of Earl Gr...   \n",
      "\n",
      "                                            combined_text  review_length  \n",
      "147596  love stuff would drinking twinings english bre...            426  \n",
      "61747   grows really fast dog cat not eat cut put food...             46  \n",
      "347537  strongly aromatic version earl grey scented ci...            125  \n",
      "(70000, 49300)\n",
      "(70000, 49301)\n",
      "(30000, 49301)\n"
     ]
    }
   ],
   "source": [
    "# Please write all the code with proper documentation\n",
    "label = final['Score']\n",
    "\n",
    "# split the data set into train and test\n",
    "X_train, X_test, y_train, y_test = cross_validation.train_test_split(final, label, test_size=0.3, random_state=0)\n",
    "\n",
    "type(final)\n",
    "print(final.head(3))\n",
    "\n",
    "#Bag of words using Scikit learn\n",
    "count_vect = CountVectorizer()\n",
    "vocab = count_vect.fit(X_train['combined_text'].values)\n",
    "type(vocab)\n",
    "\n",
    "\n",
    "# #Using the vocab obtained in train to vectorize both train and test data\n",
    "X_train_bow = vocab.transform(X_train['combined_text'].values)\n",
    "X_test_bow = vocab.transform(X_test['combined_text'].values)\n",
    "\n",
    "print(X_train_bow.shape)\n",
    "\n",
    "X_train_len = X_train['review_length']\n",
    "X_test_len = X_test['review_length']\n",
    "X_train_len = X_train_len.values\n",
    "X_test_len = X_test_len.values\n",
    "\n",
    "X_train_len = X_train_len.reshape(X_train_len.size,1)\n",
    "X_test_len = X_test_len.reshape(X_test_len.size,1)\n",
    "\n",
    "\n",
    "X_train_len = sparse.coo_matrix(X_train_len)\n",
    "X_test_len = sparse.coo_matrix(X_test_len)\n",
    "\n",
    "X_trainf = sparse.vstack((X_train_bow.T,X_train_len.T)).T\n",
    "X_testf = sparse.vstack((X_test_bow.T,X_test_len.T)).T\n",
    "\n",
    "print(X_trainf.shape)\n",
    "print(X_testf.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.metrics import roc_curve\n",
    "from sklearn.metrics import roc_auc_score\n",
    "from sklearn.model_selection import cross_validate\n",
    "\n",
    "#function to find the optimal alpha in NB for 5 fold cross validation\n",
    "\n",
    "def find_optimal_alpha(x,y,n):\n",
    "    # creating a list of alpha for NB\n",
    "    alpha_list = [0.00001,.0001,.001,.01,.1,1,10,100,1000,10000]\n",
    "    \n",
    "    # empty list that will hold train scores\n",
    "    train_scores = []\n",
    "    cv_scores = []\n",
    "    \n",
    "\n",
    "# perform 5-fold cross validation \n",
    "\n",
    "#dropping the prior probabilities as we have higly imbalanced data\n",
    "#using roc_auc as our test metric\n",
    "    for alpha in alpha_list:\n",
    "        clf = MultinomialNB(alpha=alpha, fit_prior=False)\n",
    "        scores = cross_validate(clf, x, y, cv=n, scoring='roc_auc',  return_train_score=True)\n",
    "        train_scores.append(scores['train_score'].mean())\n",
    "        cv_scores.append(scores['test_score'].mean())\n",
    "\n",
    "# changing to misclassification error (error = 1 - accuracy)\n",
    "    MSE_cv = [1 - x for x in cv_scores]\n",
    "    MSE_train = [1 - x for x in train_scores]\n",
    "\n",
    "# determining best alpha\n",
    "    optimal_alpha = alpha_list[cv_scores.index(max(cv_scores))]\n",
    "    print('\\nThe optimal value of alpha is %f.' % optimal_alpha)\n",
    "    \n",
    "#taking log of the values for graph to maintain the scale\n",
    "    log_alpha = np.log(alpha_list)\n",
    "    \n",
    "    plt.figure(figsize=(16,8))\n",
    "# plot misclassification error vs k \n",
    "    plt.plot(log_alpha, cv_scores,label='CV_plot')\n",
    "    plt.plot(log_alpha, train_scores,label= 'Train_plot')\n",
    "\n",
    "    for xy in zip(np.round(log_alpha,3), np.round(cv_scores,3)):\n",
    "        plt.annotate('(%s, %s)' % xy, xy=xy, textcoords='data')\n",
    "\n",
    "    for xy in zip(np.round(log_alpha,3), np.round(train_scores,3)):\n",
    "        plt.annotate('(%s, %s)' % xy, xy=xy, textcoords='data')\n",
    "\n",
    "    plt.xlabel('Log-Value of alpha ')\n",
    "    plt.ylabel('AUC Score')\n",
    "    plt.legend()\n",
    "    plt.show()\n",
    "\n",
    "    print(\"the AUC score for each alpha value is : \", np.round(cv_scores,3))\n",
    "    return optimal_alpha\n",
    "#     # ============================== Naive Bayes with alpha = optimal_alpha ===============================================\n",
    "# instantiate learning model k = optimal_k\n",
    "\n",
    "#function to predict the label of test data and return the same as list \n",
    "def pred_test_data(x_tr,y_tr,x_tst,y_tst,optimal_alpha,model):\n",
    "    NB_optimal = MultinomialNB(alpha=optimal_alpha, fit_prior=False)\n",
    "\n",
    "    # fitting the model\n",
    "    NB_optimal.fit(x_tr, y_tr)\n",
    "\n",
    "    # predict the response\n",
    "    prob = NB_optimal.predict_proba(x_tst)[:,1]\n",
    "    \n",
    "    pred = NB_optimal.predict(x_tst)\n",
    "    # evaluate accuracy\n",
    "    roc = roc_auc_score(y_tst, prob) * 100\n",
    "    print('\\nThe ROC_AUC score of the NB classifier for alpha = %f is %f%%' % (optimal_alpha, roc))\n",
    "        \n",
    "    return pred,NB_optimal\n",
    "\n",
    "#Function to create Confusion matrix creation and plotting the same\n",
    "def confusion_heatmap(x,y):\n",
    "    cnf_mat = confusion_matrix(x,y)\n",
    "    print(\"Confusion matrix is : \" ,cnf_mat )\n",
    "    print(\"The True Positive value is:  \",(cnf_mat[1][1]))\n",
    "    print(\"The True Negative value is:  \",(cnf_mat[0][0]))\n",
    "    print(\"The False Positive value is:  \",(cnf_mat[1][0]))\n",
    "    print(\"The False Negative value is:  \",(cnf_mat[0][1]))\n",
    "    sns.heatmap(cnf_mat,cbar_kws={'label': 'colorbar title'},cmap='YlGnBu',linewidths=2, linecolor='yellow',annot=True,fmt='g')\n",
    "    \n",
    "    \n",
    "def metric_score(x,y):\n",
    "    print(\"The precision score for our model is: \", precision_score(x,y)) \n",
    "    print(\"The f1 score for our model is: \", f1_score(x,y))\n",
    "    print(\"The recall score for our model is: \", recall_score(x,y))\n",
    "    print(\"The accuracy of our model is: \", accuracy_score(x,y))\n",
    "    \n",
    "\n",
    "#Function to plot the the roc_auc curve for train and test data    \n",
    "def plot_roc_curve(xtrain,prob_train,xtest,prob_test,lin1,lin2,l1,l2):\n",
    "    train_auc = roc_auc_score(xtrain,prob_train)\n",
    "    fpr,tpr,_ = roc_curve(xtrain,prob_train)\n",
    "    plt.plot(fpr, tpr, linestyle=lin1, label=l1 )\n",
    "    test_auc = roc_auc_score(xtest,prob_test)\n",
    "    fpr,tpr,_ = roc_curve(xtest,prob_test)\n",
    "    plt.plot(fpr, tpr, linestyle=lin2, label=l2)\n",
    "    plt.xlabel('False Positive Rate')\n",
    "    plt.ylabel('True Positive Rate') \n",
    "    plt.legend()\n",
    "    plt.show()\n",
    "    \n",
    "#Function to generate wordcloud given a list of words\n",
    "def generate_wordcloud(text): # optionally add: stopwords=STOPWORDS and change the arg below\n",
    "    wordcloud = WordCloud(relative_scaling = 1.0 ).generate(text)\n",
    "    plt.imshow(wordcloud)\n",
    "    plt.axis(\"off\")\n",
    "    plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "The optimal value of alpha is 0.100000.\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x236c6f6d5c0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "the AUC score for each alpha value is :  [0.864 0.883 0.903 0.923 0.937 0.936 0.664 0.501 0.5   0.5  ]\n"
     ]
    }
   ],
   "source": [
    "#Encode the class labels and find the optimal alpha\n",
    "encoder = preprocessing.LabelEncoder()\n",
    "enc = encoder.fit(y_train)\n",
    "y_en_train= enc.transform(y_train)\n",
    "y_en_test = enc.transform(y_test)\n",
    "\n",
    "opt_alpha = find_optimal_alpha(X_trainf,y_en_train,5)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "u-ZSD3tlJklg"
   },
   "source": [
    "### [5.1.1] Top 10 important features of positive class from<font color='red'> SET 1</font>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "86fu_aO8Jklg"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "The ROC_AUC score of the NB classifier for alpha = 0.100000 is 94.210795%\n",
      "not great good like coffee love one taste tea product\n",
      "The wordcloud for top 10 positive features are:\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x236dac3fc50>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Please write all the code with proper documentation\n",
    "from wordcloud import WordCloud, STOPWORDS\n",
    "\n",
    "y_pred,NB_opt = pred_test_data(X_trainf,y_en_train,X_testf,y_en_test,opt_alpha,vocab)\n",
    "\n",
    "pos_class_prob_sorted = (abs(NB_opt.feature_log_prob_[1, :])).argsort()\n",
    "\n",
    "#Create the top words for each class based on the prob scores obtained for each feature NOTE: we are leaving out first prob score as it corresponds to review length\n",
    "\n",
    "pos_text = np.take(count_vect.get_feature_names(), pos_class_prob_sorted[1:11])\n",
    "\n",
    "posi_text = \" \".join(words for words in pos_text)\n",
    "\n",
    "print(posi_text)\n",
    "print(\"The wordcloud for top 10 positive features are:\")\n",
    "generate_wordcloud(posi_text)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "mquFtwF5Jklj"
   },
   "source": [
    "### [5.1.2] Top 10 important features of negative class from<font color='red'> SET 1</font>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 109,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "_RacUnTrJklj"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The wordcloud for top 10 negative features are:\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYEAAADKCAYAAABDsfw/AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMS4xLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvAOZPmwAAIABJREFUeJztnXd4I9W5/78z6l2yLfe+trd5d729L3VpG3oJhBYgARJCbkiA3PR2c28qN+WX3CRAaIGEGkpgCW1hF3YXtvdmr3u3JVm9jGZ+fwhrLWtGdSRb1vk8j5/H095zpLHP95T3vC/FcRwIBAKBkJ/QU10BAoFAIEwdRAQIBAIhjyEiQCAQCHkMEQECgUDIY4gIEAgEQh5DRIBAIBDyGCICBAKBkMcQESAQCIQ8hogAgUAg5DHSqa4AAFAURbYtEwgEQpJwHEela2NaiEAuI5UC552tBAAMDrHYf9A/xTUixIK8LwIhEmo6xA7K5ZHAyuVyvPlyMQDg1dc9uPXO0SmuESEW5H0RZhJkJDAN+NF3jFNdBUISkPeVPkpzOYpWngt940JQ0vhNCMsEYD95EL2vP51ymcbm5ai4+IaIcwHnGE7+348izuka5qP6yjsSsukd7MHwx+/CfvIgIGpnmIJ+9kIULTsbqvKahJ4IelxofewXYFwOEeuRGEQE0kCrpbB0sXyqq0FIEPK+0kOiVKP0nMthbF4GIPEOKC2VwThvKVivB4Nb/wU2IM4UnExrOFM3hQoVmz4H3az5CT+vLKlE1WW3wtXViu5Xn0DQ40q7TsriclRcdD2UJZVJPSdRadD4hW9hcNtmWPZ+CCB7kyPEOygN1q1WIIGOEGGaQN5Xesy67QEYm5cjGQGYSMGSdZh16zcgN5lFrZdUq0f9zV9LSgAmoqluwKyb74PcWJh2Xepvvi9pARiHlitRdt6VqL7yNtCy7HVWiAikwTkblFNdBUISkPeVOpqaxoied6rITWbU3/Q1qFJsKCdDyxWovfbutIVFZihA1eWfT8tG8dqLQNGStGwAgK6hGTXX3g1arkjbViKQflEanL0+Oy+JIA7kfaWGqrwm5jy7o+0ovIM9YFx2BD1uSFRqaGtnQ1M3B7RUFnW/RKlCzXV3o/2Z38M3OphW3crOuwqKotJJZzl4+rvg6moF47SDcTsh1eqhKq2CpqYJUrWW15ayuCLlehQuOwvmNRfwXvMM9sDd3Qb/mCU85aQsqYCmqiEkhlT0yEpdUYuqS29B54sPp1ynRCEikCI3XKtGU2P0HzhhekLeV2pUXf556JsWRp33DHSh++XHEXDYeJ+z7N8e/l2i0qDq0puhqWk6c06pRsPt38SRX349rfqFpqdCMG4nOl/4C7yDPXGfU5rLMevz90edL1p1HkZ2vptw+bRcgYY7/jNqlBT0uND96hNwdbXyPjd2bG/EsXH+MlRc8rmIc9r6uZCbzPBbhxOuTyoQEUgBvY7GpotUU12NMBQFNM+TYdECORbMl2H+XBkaG6QwGGhIaAqBAIcxO4vde/04fjKAXXv8+HCHD253ZhefdDoa525QYEGzDJderIJOR0OnpaFSUfB6OQwMBdHXH0TraQZtbQwOH/Xjgw99otdjur2vicyqk+Kcs5S46XoNKsol0GkpSKUU3G4Wu/b6ceBgAB/u8GHrR14wTHbrpq2dzSsAvtFBdL7wcMILqUGPC13//Cvmfu1nUdc01Y1wdZ1Ku65Brxudz/0fvMP9Cd3vHe7jPW+YvTgpETCvuSBKAFgmgI7n/gTvUG/CdmxHdkeJAACUrL8E3a8+kbCdVCAiEAeKAupqpGieL0PzvNDPxRfwNyiXbVLB2pv8XKepIn7PhQ+ZlMKG9Qr84SETSoqF5yKlUgoqlQSfuViFz1wcqrvXy+H1Nz34wj2WlMqOxVnrFLjjVi0uOE8JhYJ/EVGjoTCrTopZdVKsX3Nmmmb7Th/eed+L519yo6c3mHTZ2XhfDQv6MGphk35uYh0v36TCPXfpsGwJ/wKgTkfj3LOUOPcsJe67VweLlcWTT7vwf484MTSc/PeSCsVrL4o6x7Esul95PGlPGjbgB+v3gpZHrsuUrL8Ep5/+bVr1BICB915OWADGcfd1QF1eG3FOYS4FLVeA9cfvjEjVWhQsXhd1fnDLK0kJQLg+Pe1QV9ZFnNM3LYRMb0LAbk3aXqIQERBAq6Xw4+8Zcd2Vamg0ae/HEJViswSfv0mD22/RxGz8Y6FUUrj6CjW8Pg4Pftcm2qigrkaKl59NfZFuzSoF1qxS4LsPGrBtuw/f+oENx44H4j43nd/XZLZsLsGiBclNTRWYaHztKzrccasGNXP7xHVr50FdXsvr4249sD3lefzhne+iZMOmiHOq8hpoqhrg6uafNkkEd18HbEf2JP2c9eDOKBGgKBqq0irBaZyJFCxZF7Xm4beNwnpwZ9J1AYDRPR9EiQAoCsb5yzG8462UbCYC8Q4SQK+jcdtNmmnXoGw8V4mDn5TiW/frUxaAidz4WQ3+9YIZanX6n3PVCgXe21ycth0AoGlg7SpFwj3u6fq+JnPbTZqkBWAiOh2Nvz1aKMr7isXEufYwHIfhne+kbHN0zwdg/d6o86aFK1O2CQCW3VuRil+9u6ed97zcWJTA06HGeTKje7aCY1MbJTpaD/OOQHQN81KylyhkJJBjvP2eF39+1ImvfknHe31gMIiDhwPYstULq42F38fBaKRRVSHFV7+sg4RHNxYvkqP3VAXuvNeC519yJ10nlYrCh++UoL6W/8+pty+Ix55yYWAwCI+HA8dxMBporFiuwNzZMsxpkvJOG1138wiGhrIz9ZFp7rlTh//6Ab+LZTAI/OlRB7q6g7BYWXBcqOd/w7VqLFogBz2pq3bJhSr0nqrA2+95ccPnRxDMwFdkWrQ66tzAllfAOO0p2+QYBj2vP4PqK2+POG+YtxTWQ5+ktDbgOHUIYyf2p1Qfv3UYAYcNMl3kLnKZoSDus+bV50OmN0Wdt+zdllJdgNBUm/XQxyhcuiHivKq0GuryWrj7OlK2HQsiAgL09QdRWi88r/f3xwuj/M7/tTkzc+yT+dMjTtz9BS3kslDD+f42Hza/5cHmtzzo7hFuEf72Dxe+8009rrpMzXv9gf/Q48WX3Ui2I/OFW7W8AnDwcADf/4lNcLH3sb+F5pUVCgprVipwzlkK3Ht3SNw6Ohm8vy261yhENt+Xz5dcr3PDWgV+9N1oAWAY4MlnnPjNHxy87+3hx5yoq5Hiwa/rcf010e9s47lKfO0ePX79u9Qb5mRw93WmbcMj0JCpK+tSEgGhBd5EYX0eYJIISBLwz1dX1MW9JxWE1lpkhgKAiED2ifXPztdQsmzyDUQq9A8E8cI/3Vi3WoEnnnHhod8lFm/kdAeDO75kwaHDAfzg29GNUmNDaJE2WQ+dz/I0UDs+9uHKG0YS+j58Pg5btnqxZasXL/zTjeuuVqO1jUl63ns6vi+tlsKfflcQNQIbtbC47a5RbNse+7tu72Twpf+w4P1tXvzuV6aw8I/z4Nd1eOVfbrSezrzrkG90IG0bjNuJoMcFiUoTcV6Voo++dyS9OnEsz/vn8dufjLK4PK1yhQj6PLznxdioJwQRgRzlm98LLeamMv34mz84sKBZxjsiWLs6OREwF9GYPzd6jvvr37Kl1MAePBzAwcNjST83XXnwPj3KSiMVwOvlcNUNwzh4OP6C9zjPvuCGQU/j5z+J7LXKZRS+800Dbrsr89FQE/GYSQTvyAA0VbMizskLUltL8ltHxKhS0kg1et7z8x94KCPlURkMI0EWhnMUpzM1ARjnx/9j531+7arkdtWWl0UvMvQPBHH8ROIN3Ezmi5+P3p367R/akhKAcR5+zImdn0Q3xJddMj33QAjBuKNHrjJdaj1dvoXmmQidwaBXRATylM4uBj/7dfRccrJRNmWy6KFzgYnOuPdKLkBRIVfcyTz2VGrRKjkOuPTa6N2jkxeOpzt8i8u0XAmJIvnYTlMRenlqyNz/U479+RDE5CiP/73Q5i4hurqjFzQVCgq338wfnyWfWNoi/hBeaNewSpU7ossx/KMgSpJ8b5fj0hgOEwCQNYG8xmpL/x9oaDiIU60MGhsi/5S+9y09+gaCeOmV5F1OZwpn80QtzdQCbvM8GXbtyY1UmUJhktlsx8XIAF0vP5YRu37LUEbsAkQE8hqnU5xe1NPPuvDD70TO6cplFB79YwEu36TCT39px8lT+bdGsGpFdGN34mRmvoeaamkOiQD/upPQCCGXcJw6NNVVSBoyHZTHiBV64LGnXBgU2NR12SYVdrxXgnWr8y+Mc/O8aK+ptvbM9HbLStLfPZ4t+OLkB31ecOzM2BiYa5CRACFt7A4W93zdiueeLOJdpKRp4LUXzDh5KoAnn3Hj2RddGBmd2XO5Oh3NG9bjq1/SCe72TodcWoiX8vi8M86Z4xacaxARmAHcfIMGG89VYvVKBYoKp2Zw9+4WLwqrerCwWYYP/l3Ce09Towz/9QNDOHzCwGAonMRjTzkxPDKzRGFhc3ZzF/B5IYmNurIe7p7TadmQavVQV9RGnbefOJCW3Wzj7m3n3TVM0XTKsYOmCjIdlMPMbpLhz78rwO9+ZcKll6imTAAmcvBwAA98x5ZQuOPSEgm+db8ehz4pw6/+2xi1qSqXUeeQt06i8DXeydvgD7eQqbg4mcLd28F7XllSld2KiMDUtxqEpFEoKPzPj4z46J0SXHc1fxygqeSRx51YvGYA3/mhDX398cVAoaBwx61a7NtempUebTZI1tV2uuEbiY7Nz5dgJln4bHDBYNojjGzjPH2M97xpwYos1yR9yHRQjmEuovHC0+aEphvsDhYDA0GMWlhYrCz6B4JwuTm4XBzcbg7l5RJ8+YuZ8ed3uzn88WEn/vKYE5dtUuPRP8aPzKhQUNj2dgm+8OVRHDiU254iQpE9//iwE9t3ip89TWzX07Hj+1G8rizinKq0GnJDAfxjqQVJpKUy6OqjwyI724+BDeSGZ9M47t7TYFwOSDWR6zuGuUsw8P6rooXYyAZEBHKMV58zY85sYQE4ejyAf7/txSNPOOP2wtetVmRMBMZhGOClV9w4dMSPm2/Q4LNXq1FsFp72aaiX4s2Xi3HzF0fxznu5GxLA6+V3vTrdzuD1N/mDhE0nbId3wbz2QlBU5GSBec2F6N3895RsFi47i9czyHrok5TsTSXjYZ/Nq86POE/LFSheexEGtrwyRTVLHjIdlEMsmC/jFYCf/sIOU0UPTBU9WHveIH78s7GEpmGyyalWBt//yRhmt/SH6yoUtlqppPD8U0V493VxEtRMBUcEsqHNacqNflfAYUPrI/8DbtIGLmPzcjTe+V3BAGpC0DI5itdfEnW+5/Wn4Wg9nFZdp4qhbW/wZvwqXHYW5tz7U1HCTU+OtpoJiAjkEHfcyt9r/9VvU4snP9VZuL5wjwVnXTiI3fv4pwKWtMix8dzk48lMB4aHg3C5okcDfHsHpit+2yhvIyc3FKDm2rsgTyD5ChBqyKquuI332tjR5NNCTieGd7wDH084a4lShZrr7kZBy5qU7CqKylBy9qUwzF2cbhXjQkQgh7jw/OgGMZ14+IUFU++Nc/hoABdfMYQX/skfXuKaK6bfwncicFzos02mZWHmQgJngpFP3uPN/6s0l2HWbQ+iaOW5gs9SUimMzcvRcPs3oa2dHXU96M39kCJckEHPv57iXQOgpTKUbbwGBS1r4mYroyRSGOcvQ8XFN6Dp7u+j4bYHULT8nExVO4LcGJsSUGyWoJRnV+jbacybV5RPvQgAoXWDe+6zYmGzDE2NkT3ltTm80/jj3T6sXB7Z6Oea9xPHsuh59UnUfe6rkJsic+/SMjlKNnwGmqoGePq7wLjsYDwuSBQqaGtnQ1s/VzBOEBcMovvVJ7PxETKOd7gf3a8+georb+cNgle28RqUITSyCtitCHpc4NggaJkCtFwBmd4EmcEUtf6SLYgIpErmE4hFwBe3HwCOHEvdi2Y6NbD+AIc/PerEQz+LzNtaVioBTfNnBkuKLL8vAHj7XW9GdgdnG8btRMdz/4emu77He11bNwfaujkJ2+PYILpfexKuzpNiVXHKcbYfR+eLj6D2ursF75EbCyE3FmaxVolBpoNSpKcveuF1Vn3mNNXn52/FnM7UWrcHvqbHWeumjwgAQHtHtJsjTQMSOv3ec7bfFwB8uIPfTfCnPzTynp/OBOxWHP3V/Rja9kbUYnGijJ3YjxN/+D6O/vqBnAy0Fg9X50kcfehBDH34pjgurxyXlf0TZCSQIiOj0Y3KnCYZtFoq5YY5FqMW/q5wfQoNWVmpBPfcNf3i/fPNlw8MBhFg0v8+s/2+xnG7uai4PnffocXrb3oysl8gk3Aci+Gd78B6+BOYV54Hw/zlCSWC4YJBONqOoGeGTP/EggsyGN7xFiz7P0JByxoYm1ck1fvn2CA8fV1wth+D7chuBBy2DNY2BMWJFUoynUpQ1NRXIkk2XaTC3x6Nfrnf+JYNf33SKXp5NA2cOliOAlPk4K27J4iFK6N3dwqhUlF44yVzzAVKU0VPwvaKzZKEQkTEQ62m8NE7JaitiRS1197w4JYvpp8/N9vva5xvP6DHA1+Ldqe0jbG4/tYRfLwrvR6jRAIYDbRgJyGT0FIZNLWzUbT8LMiMRZAoVKClMrBMAN7BXvhGB+DqboOz/fiMWAROlcKlG6Aqq4bcVASZzhTaK0FRYP0+sH4fGJcdPssQ7CcOwN1zOqlRBMdxaQ+TiQikSIGJRuuhclCTXsHIKIv1GwcxMCi+n/5jfy7EFZ+JziebaKNdXibBY38qxIplsT1UkhGB3R+WorcviKf+7sJrb3hS8lZSqSg88ocCXHJh9Ge7814Lnn8p/QZkKt4XACjkFN5/s5h3f4fXy+HbP7Thqb+7BDOGCTF3jgw3XKvGZ69S41e/c+DhxzInZITpixgiQNYEUsRiZbH1o+jhfFEhjc3/NGckteBrb/DvNL3+mvhulEplKCRDPAFIFpoGNqxV4OH/V4DWg+V4+q+FqKlObIpKIadw7VVqbH2rhFcAunuCePVf4uyunYr3BYTWcu74soV3ykmppPDQz0zYsaUU1VWxvzOFnMKc2TJcf40ah3eVYfu7Jbj3bh2KecJVEwjJQEYCafLMY4W4+ILoBixR2jsZLFkTvdlEiJGuSkhE+L//+/Nu3HOfBZaeyqhryYwE9m4vRV2N+EtLP3/Ijp/9OrVNcLHI9vuayNnrFXj6r0Wix/5/8Ls2MhLIU8hIYBpw572WrKZO/Ob30l8oevQJJ77ydYtomcXEhGWBH/33WEYEAMj++5rI+9t8OHdT5nLFEgipQEQgTZxODhdfOZy18h59wok777XwhiSIh8fD4T8esOL+b9vCfvfpLigePiJug3repiH85g8OUW1OJNvvazInTgbw37+0w+NJX4FZNpTMZ8fHueVlRJheEBEQAYuVxVe+bsUxgaBhsUilMXj+JTc2XDCIf7zgTth90ufjsHzDAJ58xhVxftv29BqQW744ivM2DeHRBKKWChFgOGx+K+QFtP9g5kMKZ/t9TeaXv7GjeXk/fvoLOzo6k/e53/mJDz/5+RgWrOjHNTeN8IanIBAShawJiMyKZXKsXa3AiqUKLGmRwWCgIZVQ8Pk42B0s9h3w41QbgxMnGew76MeJk4G0dsOWFEvwo+8YsKBZhsICGkYDDYmEQv9AEN09DI6dCGDHJ368/a4Xdkfm3Qhn1Ulx0/Ua1NdJUV4mQVEhDYOBhkFPg+MAl5uFw8HB4WDx2mYPDh0J4MPtPlhtU5OSL9vvi48F82X46pd0qK+TorJcAp2OhkJBYdTCwuViMTjE4lRbAK1tDA4fC+DdLbkbYpsgLsRFdJpRrKxDmaoJerkZCokGHMchwHrhYqyw+QfQ7twLlkust1ysrMNsw1ooJBpIKBkY1o8A60Wv+1hSdggEwsxFDBEgO4ZF4qKKe8O/cxwLJ2MBywUhl6hgVtbCG3Qm1HDPN56DKk1zhJ0A64OEkkIuUUEp0RIBIBAIokFEQARk9JkYPMPeThyyvg0/e8a/XUrJE4oQKKMVqFDPDR9vGfhrhJ3QPbkZX59AIExPiAiIgFF+JhfrEdt7UQ03w/kTimJplJeBps5sAphsBwACLJkPJhAI4kG8g0RAKQkFY/MGHfAGU9+0M25n3BaBQCBkGiICIiChQnFhGDY9V71xO2LYIhAIhETIOe8gnawQJaoGVKrnQkYrQVMSMGwAvqATPe5jGPa2w8XE3lVbpmpEsbIeBnkJ5BIVKNAY8XZi0Hsa/Z6T4DhhH0CFRI15hrOhkGiglGihkGhAIfYCvZsZw9bByDC6Coka9dplUEq0MMhLErIDAG/2/j7uPXqZGc2m86CUaCGjFAhwXriYMey3bIYv6Ir7/EQ7ZeomFCqqPh2lcHAxYxj1dqHbfSQpWwQCQXzyykVUKdFinvFsFCvrYt4X5AJ4r/8RBDn+TThGeSlWma8VfN7F2HDY+i6s/j7e63qZGWuKr49X3Qj4RCAVO0BsEZBScswzno1ydXQ+VwAIcgxO2Xeiw7kvZhnx7CRji0AgZI68cRE1yEuwtPBSyOkzgb84cGBYH1guGB4RAECf+6SgABQr69BScHH42M964GHs4MDBKC8FAGikRiwvugIHrP/GoKctyoaf9aDLdTDinE5mhklehgDrRb8nOmWePxi9wBvLDgBBW0LIaCWWF10BvcwMINRIuxgLGDYAhUQNjdQECSXFHMM6qCQ6HBvbKmhrhfmqsJ2JtiSUDBppKP3juC0iAgRCbpMTIrCkYFOEAAx7O7Dfsjnc2FOgoJYaMc94FnrcR3htqKUGLCy4ADQlgYux4qjtA1h8PeA+ddsxyksxx7AeRnkpaEqCRaYLsYN5Do7ASIQdb9CJo7YPIs7VaZfAJC+DL+iOuiZELDsAkrIFAAtNG8MNd4D14r2BRyOmtZQSLc4uvQ0AUKNdhLHAIPrcJ3htTbRzbGwr+j2nwraUEi1m69eiTN0EAChXzxa0QyAQpj85sTCskGjCvx+xbcGe0dcievscOLgYK3aNvIwx/yCvjbmGsyClQjHjPx5+EaO+7rAAAIDNP4BdI/+EzR/K0kVTEiwwnZ+JjyM6papGmJW1AELTYR+PvBi1ruENOiNGNk36Nbx7F0pVjRF2+twnImx5g86IUZKQHQKBkBtM+//eas1CAKGG/v2Bx9DtOpy0jXrdMpiVNQBC0zB8/vdAaNpj5/ALsPlD8eInTolMZ1oKLgr/vmPoOTgDFt779lnegC8YytKllGgxR78u4rpSog3bimVnoi0+OwQCIXeY9iJgUpQDAOz+oZR98E0TNnONN/CxEFoUnu4EOQZORrjhBgAfe+Y7VEsNEddM8vKE7Uy0NdkOgUDIHab9moBKogMA2CfNzSeD8lMbAOCO4z6a6D3TEQkljYhhFI/JISjGG/N07RAIhNxh2o8E5J82MOmES5gY2yfAxo+fH2DPxLQfX0eYidCT5vInfk/p2CEQCLnDtB8JILyBSpz9DIlsyBI3A2z2CHKBhKa7xpm8qY76tE+Qrh0CgZA7THsRcDCjUEsNMCkqUrYx5h+EUhWKy2OQl8S9f+I9DJf5TFdiQYHGntHXUg41bQ8Mi2Inn1nU8FkUm+ZFnWc5Bu/u/skU1IhAiM20H8fbfCGXTYOsOGKvQDJY/L3h3xMRgYlRQXMJmpKEF3dTweLrFcUOgUDIHaa9CAx7OwCEGqZa7eKUbAx62sAh5OseT0i0ssLw7uFcYbwHDwC12paU7XiC9rCtdOwQCITcYdqLwERXxTrdkvBmpmTwBp3od58JwSCn1bz30ZQEzcZzwscuxpp0WVNBm2NX+Hezshb1uqVxnxESw3Fb6dohEAi5wbQXAQBgPvXWoUChpeAiLDJdGLVLVSFRY45hHZYWXspr4/jYh+Gol6vM16BIWROxSGyQF2NF0VXhqSAOLA5Z383ExxGdQU9bhMg16degQFEZkaAGCO0DmKVbjtXm67DQtFHQ1kQ7Swo/E2VLRivCtoTsEAiE3GDaLwwDod2piwsvCbtrlqmbUKpuQID1geNYSGl5OBa/0C5XP+vBXsvrWFZ4GdRSA5YVXoYA64WbGQMHNmIdgAOHI9Yt4RASucBh27uQ0orwzugVRVeCA8v7HQHAiLdL0NawtzNsp1hZh2JlHTiw8Ac9SdkhEGYCZmNT+HdfwAm7Kzc3kwqREyIw6uvGO31/BhDqzZao6lGmmg0ZrQBFUWA4P9zMGDqdBzDojY78Oc6YfxDv9j8Ms7IWZapGGOVl0MoKAFAY9LRh2NuBPs+JnPSKCXIM9oy+CiC0eWuB6XzoZIWQ0yrQdCjnwrD3FCy+Pgx522NmLptop1w9G4WKKuhkhVBKtGDYADzMKJyMBRZfX1QkVAJhJqHXlKOl8cbw8bDtOPaf+vsU1kh8ciafAIGQCxAX0ZlFbdl6NFaeCSQ53URAjHwCObEmQCAQCFNBoX7WVFch4xARIBAIBB5oWgajrmqqq5FxiAgQCAQCDyZdDWgqJ5ZN04KIAIFAIPDQWJkf7s9EBAgEAmEScqkGOnX8EDMzASICBAKBMIkCfT1yN55wchARIBAIhEkUGma+V9A4M3/VI0koioZBU4FCQyN06jIYtVWQShTguCCYoB9e/xh6R/ZidKwVHt/0jS1UU7oGOnUpNKpiKGTaTz8DhyDrh89vh9tngcszgs6B7QiymQ+XXVm8HAZNBVSKAqgURkglCtC0FMGgH/6AC26fBU7PIE73bQXLBjJeHz5C774ShYZZ0KnLoVEWQqUwRbx7l3cYNkcXeoZ389rIxY2GU4WElqOqZAW0qmIo5UYo5XpIJUpIJQqwHINgMIAg60eAccNib4fHZ4Xd1QuHZxAcx2akThQoqJQFKMgD19BxZvxmsY3LfxR17oP9v4Q/EJmveMOib0Ah16dUhs3ZhV3HHk3p2Vic1fIA5DJtxLlTPW+jo//DqHuN2irMr7sSamVhyuXtO/k3jIydSvn5caQSBeorzkGVeRloWhb/AQGGbSdxqvvfcHlTTy2aCGWFC9FUdWHUd50IFns7jnW+BrfkmBG+AAAgAElEQVR3FEBoMbG2bF3UfcluFis0NGBJ0828197e9YOk6ynEsjm3waSrzWgZ4zRWno8K8zLIpOIFHXR7R3G86w2MjrUmdD9NS1GonwWduhRlRS1QK0zI5rTP+/t+jgDjFs2eGJvF8nIkoFOXhv9oKIpGY+XGlAUAAIzaapQVLkT/aOZDKChl0fWsK9uAWRXnRAXVS4YgG4DF0Z5O1cKsWXAvFDJd/BvjYDY2oVBfj7a+LbzCly40JcHc2stQXpR62OwCfR1WzrsLR9tfxqD1qKj/4DOJCvMS1JatF92uWlmY1IhcJTeipfFzotcjl8nLNQGdejxYHIX5tZejpnRN2jab669CdcmqtO3EQyE3TDpDoaHyvLQEAAAs9tNgWSYtGwBQX36WKAIwDk1L0Vi5ERJa3FzPElqGlsYb0xKAcaQSBRY2XAcA8DOutO3NJKQSJRY1XI95tZdnxL7V0REehRFSI09FIJQ0pqlqI8pEaARCUJhdfTFKCxeKZI8f5aQRS1PVBaLYHbGdjH9THOZUX4JZFeeKUJtolsy+RVQhaK6/RuTFPwolBc3wTZpmzGcUMh2Wz70dxaa5GSujd3hvxmznC3k7HWQ2zkZN6Vre6y7vCKyODvj8Dnh8VkhoGZQKI/Sa8ri7COfXXQ6XZxgOd2bCUE8Ugcri5YKjGJZl4As4EGDc8DNucBwHrcoMlcIIvjnQ4bH0RKCx8nxUlayMeY8/4ITV0QmXdxhevx0cx0ImVUGtKIBBWwWtuiQix8NEjNoqLJl9M/YcfwIsl/6Ipdg0J+49voADnQM7EGDc4DgWEokcKrkBWlUJiiaEFx5nft0VGZm2ykXkMi2Wz70DKoUpwSc4+PwOeP12+AJOSGgZpFIF5FItVAoDhObtB61HRKtzvpKXIqBRFkXNC2478BC8/rGEnlcpTGhp/By0quKoazQlxar5d+NU91voGPhIlPpORC7TQqMyY1HD9dAoi8LnPT4r2vu2om90f1KeEw0V58HHOOHz21Oqj1JuxJoF9wj20gOMB0c7XsaQ9XhC9uQyDebXXsHbyBq11Thv2fdwtP0V9I6k2gOksHL+XYJXuwZ34mT3vxP6DimKRl3ZesyqOAcABQkt+/R3wlktD8S87g840dr7HnqH9yRll6JoFOjrUFqwEMWmuUlPYbq8Iwktep+79Du8f9PTLYqoGOSlCEzG6uhIWACAUIO7+/hjWD7ndmhUZt57asvWZ0QEAGBx4+egUhSEj0fGTuFQ2/Nggr6kbbX2ppc9bW7tZwQFwBdwYM/xx5Py7vEHXNh/6hnMqdmEyuLlvPc0VV+IkbFT8AWEcyIIUVrQDL26jPda58B2nOz+d8K2OI7F6b4P4As4MjbnnYvweRtNZMh6FEfaXwET9CZtm+NYjI61YXSsDcc6Xk2xhoSJ5OWawES8fntKyh5g3DjQ9qygT7uYbnCTmSgAALD/5NMpCYAYFBn4cz6zLIP9J59Oyb2TA4djna8Ljh6kEiUaJsR4TwY+981xTnW/lZLN3uG96BrcmdKzMw8KTVUXCl7tGtyBA63PpiQAkyF7MsQh70WgteedlP8gXZ7hOL39zPsfWx2d4DD1ez0m09a3Bfa01kU4HO14Bf4Av7dNWQoL8EZtVdgpIKo0jk3re2ztyY181Jmm2DQHek0577UByyGc6HozyzUixCOvRcDrH8OA5VBaNjoHtgv2wg2airRsxyPAeHDo9AsZLSMWQg2q129D58COtO0HGDdO973Pey0Vl9hYnluDlsNJ25tINnZd5wKx3KSPtpPpm+lIXotA/+jBtLefM0EfBgQ2iWXSNQ4Auoc+TnlBVwzKixbznu8c2AFOpKF67/AeQd97SVK7kSkUG4U9grqHPkmyZtFkKpRBrqBSFMCkqxG8ToRyepK3IsByDNp6t4hi63jXZt7zNaVrRN/kNM6A5ZBo9U8Fjcos2OsTc36c5YI42Pos77U5NZsStlNpXiq4K7x3ZC9szu6U6jeR9v6tadvIZebVXgahKdCTZBpo2pK3IuBwD4rWWxWyQ1G04JRJulgdnRmxmygFurqslSXUQCcT5EsfY2rO5uhKuk58jLl6RbGTi1CUJGYqRuLPP33JWxFwZzgo2TgaVVH8m1JArIYrVTIlbnwITbMo5fqEg77Feg9O90BK9Yq2MyiKnVxEqzILbqJ0eUfgncJpS0Js8lYEPD5bVspRK1KP6ikEyzJweYZEt5sM2mmSdSlRMYoVXVWsKKU+v1200WWuoVUJ/z2MiTDVRsgceSsCk0NJZwqFXLxgauMEGPeUu4Wq5Ebe88Es5wJQT9ozwQdFSSCXanivBRi3aHXmwKW0gW0mEApHwo/TM5zFmhCSJW9FIJkdwukwMbSDWLinOJmNXKYVnIZxZnmEotPw7/6dSKx3IHauArfXIqq9XEGn5t8bAJCRwHQnb0UgW7sNYwWbS5WpnnKI5ZqZ7axgibiJ0rTwOwgGxXVbFCMcdy4ikQi/BzF2BxMyR96KQLagaMlUV0F0aEr4M2V7Kz9NJSACWawvm6d7BWJ1dvJVGHOFvBUBoZDFhPTI/vcaf22EorJZp/wUgVg7uKd6/YoQm/wVgRi9QzGZib2gYIzPlK3vdZxEpp9ivYN0M7LlOmJ9/ljfcazpOMLUk7f/AYoUkoqnwkz0FvH6xwTXJUIJQLJHIq6+sbJ9iZkKEwCkksxFj80EUolSFDux/s7VCSeWIUwFeSsCMqk6K+VkyxU1u3CCDWuim7fEwhuIvwkpVt5fuYzfdTRV5Fn6uxILseobK4ZVdF5swnQib0Vgcq7eTDETRwIABJN7Z8IbKhaeBNxlWZZBgPHwXpPLtKJOYYktKpmEomjROkOxdgRrlfyJlwjTg7wVgcRzn6aHS6CxzHUcIoVaSJdEQzW4ffz++xQo0aYrJLQsp0RAKTeItibg8grvD4kVt4kw9eStCGQr7IF7hu6WtLv6slaWUE/dH3AlPNKK9R50AukmkyUUmkIkTyQu8x41fDmyUyVWp0CvKYdUohCtLIK45K0IKGS6lLJT8SGUYnHM2T1jI0sOWA5hzNXDe03sdYGmqgt4z5/qSTwdZKzw1mIlh68pXSOKHSD2OoZY1JWfJZqtAONB58B23msURWNRw/WilZUNOEERnnmu5XkrAgBQbJonip2Sgmbe86P2NlHsT1cGRvmzspUXtYhWBk1JBcV62HYiYTsOzyACjJv3mlpZKEpUVKHOQCrE8mgSA6XcIJgGMlWGbfw5oQGgQF8valmZRigBzkwc0eS1CBQZG9PutcqkapQWzOe9NmQ9lpbt6U7/6AHe8zUlq0EnlfVLmArzUsHFS6HFXj44jo0pGlXFK5Ku22TE9DgTyq0MJJtRjZ8K81LRN/ZZHZ2CDgMhcqcXHcuRYKaR1yJAU1LUlq5Ny0Zt2TrBBm+6LJ5milj/KHVl69O2L5OqUS/ilEW/QBpQACgrWiRaOeIgvCZQWrAgLcsyqTpmLuB06BnaJXhNzHeZaYRGjRploWh7K6YLeS0CAFBdsjLlqQCdujRj/0y5gtBO0bqydTBqq9OwTGFe7eWC3jajY61JW7TY2+O4tqbeUy0rzJ6IVJeujhkPKTYU5tVelrFpje7h3YKL9bMqzkZF0ZKMlCs2wtFlqYznDs82eS8CFCXBooYbkn5OLtVg4azrBP8ZYw3nZxKn+z/gPU9RErQ03pCiwFKYU3MJik38ieE5LoiT3YkvCk94Eh0DHwleTXWB2KSrxdzaS1N6NhW0qmLMqjg3pWfn112e0UaMZQNo631P4CqFeXWXYVbFOaJNRSkztBEtVvrWdGcPphuU8Cp4FitBURmrxMblP0roPo5j0Te6H10DO2LGxJdKlNjQ8o2YCeTtrj58cuxhwbSIiXJWywO8c5AW+2nsOfFEWrbFRKUowMp5d0ImFQ6ZsPfkU7CMtcUMJiaXabGm+Ssx7Vgdndhz4vG0vtsCfT2Wzr5V8LrDPYDWnncwMtYKoWkZk64G9eXnoEA/OdcyB74RBcsxeHf3T5KqZ03pGjRVXRjzHqujA22978VstNTKQpQXLUZNySqeqUv++r696wdJ1XUyG1ruTygkR2vPOxgda4XDMxjjnVIo0NdCry6HUVcDo7Y6/DdidXRg9/HH0qqrEEuabkGhgT+PtcXejhPdm5NIKUpBJlUJTjOlCsdxaatpXkZ28vissDo6UF60OHyOomhUFC1BRdESuLwjsDm74PaOwuu3Q0LLoJQbYNBUwKSribnoGWQDONz+UtoCkEt4fBYcbHsOi5tuFNwxvKTpZvgZF2yOTjg9Q/D5HWA5BlKJEmpFAfTaSujVZTE3L3l8Vhxsezbt79ZiPx3zuk5disVNN8Hrt6NrcGc4k5tcqoZMqoJGWcTrWebxWXGi6w20NN6YVv3G6R7aharilTGzdpl0tVg253b0jeyDyzuKAOMGyzKQSdXQqUtQoK+DUiALnM3Zje7BnVgw61pR6juRfSf/hmVzbo877dRQeT4aKs8HyzLw+G3w+R1ggj5IJDJIaQWkEgWUCqMoi+HJ0t6/VVAECvR1WD3/y7C7emF1dMLjs4EJekFTEkgkckglSsikKihkWug15VDIDXB7R7Hj8B+y/Cnik5ciQFE0jna8CqXcyNOTC2WiSi0jGIfDp1+Ea4ZuEIuFxX4aB1qfxaKG6wWnyORSDYpN81JyzfX4bNh9/HHRptks9nbedz8RpVwvuEeBj4Ntz3265sDfu04Wlg3gaMfLWDL71rjTJxM7NIkQYNw4fPoFZGpG2OEewN4TT2Bx000JeU3RtDSN/7vMYHV0YNh2EmZjk+A9ek1Fzu+Izss1AblMA45jcaD177A5u0SxyYHD0Y7XZrxbaCxGbCex98STGbG96/ij8PrjRwxNlAOt/xA17SHLMbC7+sAEfQnFM0oUi70dJzrfEM0eADBBH/aefAoen03Uuk5mzNWLXccehdOT6JTJ9OPw6RenugoZJy9FgKakkEoUoX+GE09h0Ho0bZsHW59D7/AeEWqX21gdHbA6OkSzx4FD99DHMaNUpgIT9GLPiSdEEW1/wBUhfmK7BncPfYJjHa+JMsXo9duw+/hjE8J+ZHZN0OUdwcdH/4LOwR0ZLSdTMEGv6Hmopxt5OR0EhBbL7K4+BFk/DrY+C5lUheqSVag0L0t4Q4jV0YHOge1J7VzNB8YX6mhaivKixagoWpLw7tQgG8Cw7Th6h/fAYm/PZDURZAM40PoPAIBMqsLq5nsSzi8wOtaGjoEPedcX+kb2ibYbfZye4d3oGd4NChSWzb09KfdblgtiyHIEnYM7shrzKVw+y+Bk15s42fUmlHIDKsxLUGyal3LsIo/PBqujHYOWIxgZOyVybaPZfuj3AEIxkKpLVqGkYH7S0XKZoBf2aRpCJm+9g/aceJy3kaEoCYzaKhQaZkGnLoNRWwUJLQcHFsGgH17/GHqG92B0rDWjQ+mZRoV5KYzaKqiVhVApTJBKFKBAI8j64Qs44fFZ4XQPor1/m+CW/UxDUTRMuhoU6OuhU5dBoyyEUm4Ex7FgWB+8fjtcniHYHF3oGd49JXUcRynXo9DQAL2mAhqlGTp1aXjxNMj6EWC8GLGdgMXRAYv99LRM9q6UG1FdsgJqZRGUcgMUMh0kEhloSoIgGwj/+PwOWOyn4f7UYSNW2OpsIJOqUFO6BnpNBZQyPaRSJaQSJWhaCobxIhD0wOd3wBeww+0dxYDlMFyeEWRi1CWGd1DeisDek0+ltOGIQCAQpgvERZRAmKGse+bzUJeHNkINbm3Fge+/PsU1IsxU8nJhmEAgEAghiAgQCARCHkNEgEAgEPIYIgIEAoGQxxARIBAIhDyGeAcRCISkMKkqUaxrhEFVDrXMCIVUC4b1IRD0wh90w+EbRvvox3D7LVNd1WnPhll3Qy03AQAGHSewr+elrNeBiACBQEiIjbPvF4zmKZOoIJOooIYJRlUFBuzHiAjkCEQECARCXCSUlFcAWI6B3TsEKS2HVKKAQqIBRdGwusULzkfILEQECElDyyQoXj8LBYurYF5bD5lOgaAnAE//GIY+PI3TT30iWlmqEh2K1tSj6vKFUBSoIVHJ4Le64Rt14dSfP4L1YC84Nv0N56oSHebcdw5UpYaIcoa2tWFoW5so5dByCcyr6jDr9tVQFmlASSXwjTrh6bfj0H+9Cb/tTM5mLhBM9yOJSplhfsSxN2DH0cG3MOxsiwhsR1MSaBVFYLnpVX+CMEQECElRsakZDbevgqIwMvcvLZNApldCP7sEDXesjmnj3Yv/iKAnEPMeRZEWjXeuRfnGOVGh+ZXFOiiLdVj2m6vh7rbi+O8/wMgnwpm10imn+uoWVF/dknY5hcurMe/+86EqiQxQp64wQl1hxIbn70DrX3eg89m94FgOAacvpXIyRZn+TEA8lgtiV9c/4PJH52tmuSDs3twNHZ2PEBEgJMyC716EsvNnZ7yc4rX1WPC9iyFRxv/zVFeZsOQXV6DvzWM48st3wAUTD7ecrXLqblyOxi+uiXkPLZOg6a51KFpeg73/+QoCY56Y92cTChQMqjNRYIedbbwCQMhNiAgQEmayAAQ9ARz77Rb4rR5IFFKoK40oXj8LhrkCyeU5wDvijDmtUn7RXMx/cCMoOrJbzrj88I04EfQGIDepoSjSRtxTftFcyPQK7Pv2awl9lkTLURbrop5LppyqyxfyCkDA7oV3yAGKpqAo1EBmCOXMLVhShYXfvxgB+/SJ+qmQaiGdkFPb5umZwtoQxIaIACEhFn7/4vDvjNOHj255Ej5LdNLs9md2o/HOtaj73LLwOWf7KLbf9re4ZUiUMjT/55l0jhzL4cgv3kbfm/yJX6RaBZb87DIYm0O9VPOaehSvm4WhD9umRTkAMPe+cyKOe18/giO/fIf3XvPaerT8eBOK1/HntZ0qlJNyLLj9JIT6TIJsFiPEhaIpFC6vCR/3vnmUVwDG6fh0XnscbV1hQuVUXbkw4rj1ke2CDTMQEqO933wVnr6x8Lm5XzsblDT2n3W2ytE1mCOOHW0jOPabLYL3D390Gq1/3RnT5tSQfr5kwvSFjAQIcVFXGiHTKcLHlr2xpwMCYx4420ehm5V40nCKplB1xaLwsbvHho7n9sZ9jnH5cPJPH2LRjzcBCC30lp7diP53+LO9ZascACjbOCfiuO3xnWDjeP10vrAPNde0QG6Kn5w9G9CUBPIEEsWni0pmgFnbALO2HkZVBSS0HH7GBR/jhM3Th+OD74ATISlLNsqhKSnM2lko08+FVmGGSmaAj3HCE7Bh0HESA/bj8AdDnShWhJSh6TLjReDtXT+Y6irkPJMbJO9g/MxO3mFHUiJQuLwmwnOm+5WD4JjE/kEGt7XCO+yE0hxKC1p15SLBxjlb5QBAydmN4d8Ddi+GP4pORTkZ1sdg6KPTqPxMc0J1EgsJJUWZYT6UUh0UMh0MyjIoZTrIJdECsLjy6rj23jz2PwmVq5Tq0FR8NsoN0Z9XKdNDKdPDoCpHkbYexwbfxogz/nc4leUUaeowv+xiqGSGiPNquQlquQmFmjrMKTkfbx//JThwYNipX/sh00GEuEhV8ojjeL1ZAOACyfVwTIsqIo4HtyaR9Y0Dhibcb5hbCqlazntrtsqRGVQRYjO6qzPhfQajKbqhpoNMqkJz2SVoMK9HlbEFemUJrwCIzfpZd/E2zJPRyAuwrOqzWFD+GVBU8s1WtspZVn19lABMhqYkWFZ9A2hKGh4RTCUzfiRASB/r4b5QetRPp4aLVtXB1SW8OEjLJChclngidACounxB+HfL3m54Bx1JPd/7xhFUX90CYHzKZyHan4nOA5ytcsrObYo4HnjvZMJlDO9oB8dyUZ5LmYQJ+tFp2cV7TS7Vokw/N3w87GxNe3HYqCrHytpbQE1Ybxh0nECHZVfEbmO5RI0S/WzML70IAFBhWIAKwwJ02/bjSP/maVOORl6AtfVfiDhncXfh1PAHsLrPTJ/qlMWoNi1FlbEFF8x5IK7dbEBEgBAXxumD7XAfjAtC3jGVm+aj68X9gr7yZRfMgUTFH2OGD4qmINWcWXPwjbqSruPkZ8ZdLqeiHCDkURTxXIyF9MmwgSCC3oDgKCMTMKwPxwb5vZaMqooIEeixHcCgI3FRm4yElmFh+aXhhpkDh8N9r6N37FDUvf6gG93WfSg3NMOkqgyfrzK2YMTZFrMeYpWTiAjMLd0ImpKEj3tsB3Ckf3PU+oLDO4Qj/ZtRaViY0kgjE0yPWhCmPZ0v7A//rqkpwLxvnAtKwv/nM/vL6yOOrQd7Y9qW6pQRx0yc3cR8MG5/xLFMr4y6J1vlAIBUE9mAx9shPZmgl0muYjlEtWkp1PKC8PGpoQ94G+aJ7Ol6Du6ALeLc3NILBO4Wt5x4jbVeWYIiTX342OEdwtGBf8dcYD41vDWmzWxCRICQEINbT0UcV1wyH6sf/RwMc0qgKtNDU2VC0apazPvGeRG97aAngOO/fT+mbemkHbusL/kGkPUHMfF/TqqOHolkqxwAkCgiywr6kyuL9SYvULkABQrVpiXhY7ffgnbLx3GfY1gfTgy+F3FOKdUJ3C1uOaW6uQJ3hyifFFepdWRb3NhJnZZd8DNTvx4AkOkgQqJwIT/3iR4/2tpCrPzT9YKPMG4/Dnz/dTjaRmKantwjn9yAJgKtkEa4szMuf9Q92SoHAIL+yEaAlkl478s3irT1EQunXdZ9EQHoYjHoOAFvwBG1eS3T5dQULEW//Yjg/SW6M67AgaAHQ45TgveOE+QYDDpPosrYklCdMgkZCRASZscdT2PPAy9HzG8zbj84lkPQG4B32AnL3m5sv/1pvHX2b/HeJf+H0d1dce0G7N6IWDmamoIYd/OjqTZFHLt7bFH3ZKscAPAORS44q8r0SZWj+NQNdaZRVxgZXLDDklzE2RNDkZvtCjV1GS/HqKoQLKdQUxchNqdHdyS8z6AzyTplCjISICRM1RULMfdrZ8IgdL96CMceei/GE4ljPzkU3pWsrU2+cdbWRu5Ktp8antJy3N2R3jOaShNG0JFQGYoC9YwdOeiVxWk97/RHft8GZSlGXe1TVo5GHvk35EoikY7bbwMHLsJzaSogIwFCwsz9jzMC0P/2cdEEAACsB/vCv8sMKqjLY/taT8Y4b0LQOg4YOzYwpeU4WiMbkckhJGKhqU0szEauQYGClFbEvzEGvoAz4lgu5fECy1I5ACCTRDoG+JjEPc5YLoggO/VrP0QECIkzocPS+fw+UU2P7omcNio+qyHxhymgeP2Z++2nhsAIxOPPVjk+ixu+kTMNSeGKmoRD8BS0VMa/KQeZ3GCmAsNO8s6ioxvnbJUDICK6KgAEWf41IiGICBBylskukOkydnQA9pND4eOqyxcmPCVSenYjFEVnktz0vCrsBpitcgBgcOuZKKOKAjWKVtbGL4QCSpIRphxCQqf/N8NykV5WfDazVU7ofKR3GMsm5wUW5KZeBMiaACEllv2vcPyYoDeAgMMHd7cVtiP96HvreNQcOR8f3/0PbHzvqwAAVake57/1FRz99bvo+ddh3vulmk9DPC84k/Dk6K+E7892Ocd/9z6qrzoTrG7Jzy5Hz2uHcfTX7/Ler2swY9n/Xh0RrG8m4Q3EjzkVj8mhLLzMWNQ92SoHiJ7+kUlUQIDfWYCPySOJqYCMBAgJ0/dv4XDLE5EoZVCatShYUoX6m1dg3VO3CG6qmgjHcmBcE6ZXKGDe/edh8U8vjQqhINMrsfbJmyMaZgBxG+ZY5ZgWVUSVU/mZ5pTLARC1q7ry0uaQrUlTQw23r8bKP1wXFoDgDNwnwIGDP5hexrTJ7qF+JtpetsoBQvsKJpJsxNV01y7EgIwECAlj2dcNU0tlVJ7cRFj+u2uw66svxM2Yte/br2Hxf18WMd1kXluPc167G74RJxjPpxm/zNqoBtt2uD/h+vCVs/y314Qzi42Xw/dZkynn1F8+QtOXIndQr/j9teHMYuDwaaa0M9NM/W8fh2fAjvqbVyRcTq5g9/ZH7K5NFq0icoFdKJ9xtsqZ7A2kkRdgGPGTDQGAQqqJCDUxVZCRACE+FFB9dQua//OCqEYx6A1E/AjtwtXWFka4lwphPdCL3fe9GBWjR6qRQ1NTENqhXKKLEoDhne3Yc/9LCX+kRMuZTLLldDy7F6efivYHl+mV0DWYoWs0RwjA8PbTOPKLdzB2fGYma58YTA0IhVhOBqMqckQ25u3jvS9b5TgmiYNOWZJwGZOFZqogIwFCTCiawoLvXoTST6NiMi4fOp7di6EPT8PVZeGNxS/TKaAqM6BgaRVqr1sSzkdQem4T2h7fGTMCKRDy5f/olicx69aVqL6qJWYGL8+gA6f+/GFSUTqzXU7rozswdmwQ875+DhRF/JvAWH8Qp5/6BO1P7wLHchg7yu96muuMujrQaN4QPi7RzUb7aOLZ1Ep0kXmuA0H+kaWY5di9A4Ll+BgXvIwjHMIimdFHgTq5SLuZgogAISY1n10SFgDfiBOffOV5eAZiL7wFHD4EHEOwnxzCyI52rH7kxnADa14dOwz1OIzLjxN/3Ib2v+9B8bp61Fy7JLSJSiGF3+qGz+JG6yPbYdnfk3BSmHjlzH/wfKgrjBHlDH14GsPbT6ddzvD20/hwbxfMq+sw67bVUBSqIVHKELB74eyw4OCPN8NvPbMT2291462zf5tyedMVm6cXdu8A9MrQfotq0xJ0WnbFjbUDAKX6uVBIE9tJLWY53dbY7tCD9hOoKQjl1FZINTBrZ2HYGX9KaGK4iamE4rj0U7alXQmKmvpKEKIoWFKFZQ9dFT5+54L/FwqgliTzv7kRFRfPAwAMbDmFgz96Q7Q6EjKPUVWBVbW3hI/39byYVihpChSW19yIAnVV+NyR/s3otu0XfGZlzc0wqSP3T8R7RqxyEsmStqLmxoiefbdtP472vykYQjHVZ60AAAMtSURBVOK8pvui9jMMOk5gX0/iU40AwHFc2tuNyZoAQRBNpTHiOBUBABDh8jgxdg8hP+HA4VDfaxGeNfPLLsaSymuipkhkEhWqjC1RDfOQ41TMxlzMchLh2MBbEQHqqowtWF7zuSh7KpkRjeYNYQGYDpvFyHQQQRiBfAHJQEloGOaeCbXgHXbGuJuQL3gCY9jb/TxW1NwUPlesa0SxrhFO3yiCrB9yqRpKmT4qto7N04ND/f+aVuU4fMM4MfQ+5pScGz5XoK7Gypqbw55FCqkmYpqpb+wwPAE7ZhWtSaiMTEFEgCCIf5LnjKJIA99Ictm4qi5fAEXhGe+X0V3Zz59LmJ5Y3N3wMc6oeX6tQjh20rCzFft7X06qB52tcjosH0eIwDh6Ho+hIecpHO5/A0Xa1N1YxYJMBxEEsR3pj0iOPvtL6xOOfwMAFZuaMfvLZzw0XJ2WiJANBMK2tr+gw/JJ3Fj/nsAYDvS+jD3dz6c0hZKtcvZ2vwAvEztvNcsx2Nf9IlguCJuH3/U0m5CFYUJMVKV6rH70xqhYQd4hBxynR+Dps4OW0qCVMigK1VCXGaDiicx54o/b0Pnc3mxVm0DIC8RYGCYiQIiLfnYJlvz8csiN/JEU4xEY82DL5X8RuVYEAoGIACFryI0qbHj+jqSSnXgGHejbfASdz+2LStBOIBDSh4gAIatItQoUr62HcUE5dPVFUBbrIC9QgwsEEfQGwLgD8I04MbitDZY9XXFzCxMIhPQgIkAgEAh5DNksRiAQCIS0ICJAIBAIeQwRAQKBQMhjiAgQCARCHkNEgEAgEPIYIgIEAoGQxxARIBAIhDyGiACBQCDkMdNisxiBQCAQpgYyEiAQCIQ8hogAgUAg5DFEBAgEAiGPISJAIBAIeQwRAQKBQMhjiAgQCARCHkNEgEAgEPIYIgIEAoGQxxARIBAIhDyGiACBQCDkMUQECAQCIY8hIkAgEAh5DBEBAoFAyGOICBAIBEIeQ0SAQCAQ8hgiAgQCgZDHEBEgEAiEPIaIAIFAIOQxRAQIBAIhjyEiQCAQCHkMEQECgUDIY4gIEAgEQh5DRIBAIBDymP8PXg2C3w837d4AAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x26120dcbe48>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Please write all the code with proper documentation\n",
    "neg_class_prob_sorted = (abs(NB_opt.feature_log_prob_[0, :])).argsort()\n",
    "\n",
    "#Create the top words for each class based on the prob scores obtained for each feature NOTE: we are leaving out first prob score as it corresponds to review length\n",
    "neg_text = np.take(count_vect.get_feature_names(), neg_class_prob_sorted[1:11])\n",
    "nega_text = \" \".join(words for words in neg_text)\n",
    "\n",
    "print(\"The wordcloud for top 10 negative features are:\")\n",
    "generate_wordcloud(nega_text)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 110,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Confusion matrix is :  [[ 3896   780]\n",
      " [ 1800 23524]]\n",
      "The True Positive value is:   23524\n",
      "The True Negative value is:   3896\n",
      "The False Positive value is:   1800\n",
      "The False Negative value is:   780\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x26103899c50>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "confusion_heatmap(y_en_test,y_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 111,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The precision score for our model is:  0.9679065174456879\n",
      "The f1 score for our model is:  0.948013218344483\n",
      "The recall score for our model is:  0.9289211814879166\n",
      "The accuracy of our model is:  0.914\n"
     ]
    }
   ],
   "source": [
    "#determining metrics precision,recall,f-1 scores for our model\n",
    "metric_score(y_en_test,y_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 112,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x2616ab5f438>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "ytrain_prob = NB_opt.predict_proba(X_trainf)\n",
    "ytrain_prob = ytrain_prob[:, 1]\n",
    "\n",
    "ytest_prob = NB_opt.predict_proba(X_testf)\n",
    "ytest_prob = ytest_prob[:,1]\n",
    "#get the class probabilities for train and test to plot the roc curve \n",
    "plot_roc_curve(y_en_train,ytrain_prob,y_en_test,ytest_prob,'-',':',\"train_roc_auc\",\"test_roc_auc\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Conclusion:\n",
    "1. Naive Bayes using 5-fold validation was performed on a sample dataset of AFF for Bag of words. We have chosen Multinomial NB becuase our feature vector BOW count based. \n",
    "2. We chose the standard 70-30 split i.e. 70% as train data and 30% as test data.\n",
    "3. The optimal alpha in was computed to be 0.1 based on AUC score as our metric.\n",
    "4. Confusion matrix was plotted for the same. We got True Negative value of 3896 and True Positive value of 23524. we also get a high false positive value of 1800 meaning lesser misclassification of the negative class.\n",
    "5. The accuracy,precision, recall and f1 scores are 91.4% , 96.79%, 92.89% and 94.80 % respectively. This means that we are performing remarkabely well in Postive class classification.\n",
    "6. Also NB helps us in getting the feature importance, we got the top 10 words that helped us in classifying a given point for both classes. A word cloud was generated for the same.\n",
    "7. It could be noted that the first highest prob belonged to review length(which means our feature engineering did help), hence we took the next 10 important features! \n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "6bp7BbwCJkll"
   },
   "source": [
    "## [5.2] Applying Naive Bayes on TFIDF,<font color='red'> SET 2</font>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 113,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "oL29fzPFJkll"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(70000, 1158145)\n"
     ]
    }
   ],
   "source": [
    "# Please write all the code with proper documentation\n",
    "X_train, X_test, y_train, y_test = cross_validation.train_test_split(final, label, test_size=0.3, random_state=0)\n",
    "#Converting text into vector using tfidf vectorizer\n",
    "tf_idf_vect = TfidfVectorizer(ngram_range=(1,2))\n",
    "vocab = tf_idf_vect.fit(X_train['combined_text'].values)\n",
    "X_train_tfidf = vocab.transform(X_train['combined_text'].values)\n",
    "X_test_tfidf = vocab.transform(X_test['combined_text'].values)\n",
    "\n",
    "X_trainf = sparse.vstack((X_train_tfidf.T,X_train_len.T)).T\n",
    "X_testf = sparse.vstack((X_test_tfidf.T,X_test_len.T)).T\n",
    "\n",
    "print(X_trainf.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 114,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "The optimal value of alpha is 0.010000.\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA8gAAAHjCAYAAAD/pjbeAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMS4xLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvAOZPmwAAIABJREFUeJzs3Xl4jFf7wPHvk32zJCGEIJYQEUmQWmpNI/YGbaxtbU2LUqUtjSpt/eqVllYppbWFl9qXqNq3ouUlNPYlQRBSWYmsk0me3x/DVGRTEhPcn+vKFTPPec7cczKJuefczzmKqqoIIYQQQgghhBAvOiNDByCEEEIIIYQQQpQGkiALIYQQQgghhBBIgiyEEEIIIYQQQgCSIAshhBBCCCGEEIAkyEIIIYQQQgghBCAJshBCCCGEEEIIAUiCLIQQQgghhBBCAJIgCyGEEEIIIYQQgCTIQgghhBBCCCEEACaGDuDfqlChgurs7GzoMIQQQgghhBBClIBjx47Fq6pa0RCP/cwlyM7OzoSFhRk6DCGEEEIIIYQQJUBRlKuGemwpsRZCCCGEEEIIIZAEWQghhBBCCCGEACRBFkIIIYQQQgghgGfwGmQhhBBCCCGEeNqysrKIjo4mIyPD0KE8NywsLHBycsLU1NTQoehJgiyEEEIIIYQQRYiOjqZMmTI4OzujKIqhw3nmqapKQkIC0dHR1KxZ09Dh6EmJtRBCCCGEEEIUISMjA3t7e0mOi4miKNjb25e6GXlJkIUQQgghhBDiEUhyXLxK43hKgiyEEEIIIYQQQiAJshBCCCGEEEIIAUiCLIQQQgghhBDPjL///pu+fftSu3Zt3Nzc6NKlC4qicOHChVztRo8ezTfffPOv+7exsSn0eFRUFL/88su/7vdZIatYCyGEEEIIIcS/8OWvZzh7M7lY+3SrUpbPX21QaBtVVenZsycDBw5k5cqVAISHh5Oens7KlSv5/PPPAcjJyWHt2rX88ccfxRoj/JMg9+/fv9j7Lg1kBrmEpKen07ZtW7Kzs/Mc279/P40bN8bExIS1a9fmOtapUyfKly9Pt27dCuw7JCSEihUr4uXlhZeXFwsWLCjy/LfffhtPT088PDwICAggJSWl0PgTEhLw8fHBxsaGkSNHFtguMTERPz8/XFxc8PPzIykpCYDNmzfrf0FLUmHjfPXqVXx9ffHw8KBdu3ZER0fnaZOWlkbXrl1xdXWlQYMGBAUF6Y8V9nMqyIQJE6hWrVqRn7xNnTqVOnXqUK9ePbZv3w6ARqOhTZs2aLXaR3qsp6mwcQZYvXo1bm5uNGjQoMA/lqtWrcLDw4MGDRowbtw4/f2ZmZn06dOHOnXq0KxZM6KiogDdXoMDBw6kYcOG1K9fn6lTpxYZ5+zZs6lTpw6KohAfH19guyVLluDi4oKLiwtLlizR39++fXv9a9gQihpngLVr16IoCmFhYfkev337NgEBAbi6ulK/fn0OHToEFPy7um/fPsqVK6f/ezJ58uQi43zWx1kIIYR4Vu3duxdTU1OGDRumv8/Ly4uZM2fqE2bQvY91dnamRo0a+fYTEhJC9+7d6dSpE/Xq1ePLL7/M00ZVVcaOHYu7uzsNGzZk1apVAAQFBXHgwAG8vLyYMWNGMT/DUkBV1Wfqq0mTJuqzYPbs2er333+f77ErV66oJ06cUN966y11zZo1uY7t2rVL3bRpk9q1a9cC+168eLE6YsSIfI8VdP6dO3f0/x4zZow6derUQuNPSUlRDxw4oM6dO7fAx1JVVR07dqy+r6lTp6rjxo1TVVVVc3JyVC8vLzU1NbXQx3lShY1zQECAGhISoqqqqu7evVt9880387RJTU1V9+zZo6qqqmZmZqqtWrVSt2zZoqpq4T+nghw6dEi9efOmam1tXWCbM2fOqB4eHmpGRoZ6+fJltVatWqpWq1VVVVW/+OILddmyZY/0WE9TYeN88eJF1cvLS01MTFRVVVVv3bqVp018fLxarVo1NTY2VlVVVR0wYIC6a9cuVVVVdc6cOerQoUNVVVXVFStWqL1791ZVVVWXL1+u9unTR1VV3c+pRo0a6pUrVwqN8/jx4+qVK1fUGjVqqHFxcfm2SUhIUGvWrKkmJCSoiYmJas2aNfWxh4SEqF999VWhj1GSChtnVVXV5ORktXXr1mqzZs3Uo0eP5ttmwIAB6vz581VV1b2mk5KSVFUt+Hd17969hf69yc+zPs5CCCHE4zh79qyhQ1Bnzpypjh49Ot9jbm5uanh4uKqqqjp06FB19uzZBfazePFitXLlymp8fLyalpamNmjQQP/e4v772LVr16rt27dXtVqt+vfff6vVqlVTb968+VjvHQqT37gCYaqB8k2ZQS4hy5cvp3v37vkec3Z2xsPDAyOjvMPv6+tLmTJlHvtxCzq/bNmygO4DkfT09CKXVLe2tqZVq1ZYWFgU2i40NJSBAwcCMHDgQDZu3Ajolmxv164dmzdvfpyn8cgKG+ezZ8/i6+sLgI+PD6GhoXnaWFlZ4ePjA4CZmRmNGzfWzzQX9nMqSPPmzXF0dCy0TWhoKH379sXc3JyaNWtSp04djhw5AkCPHj1Yvnz5Iz/e01LYOM+fP58RI0Zga2sLgIODQ542ly9fpm7dulSsWBHQzSCuW7cOyP0aCggIYPfu3aiqiqIopKamotVqSU9Px8zMTP86LkijRo1wdnYutM327dvx8/PDzs4OW1tb/Pz82LZtGwD+/v6sWLGi0PNLUmHjDDBx4kTGjRtX4O9lcnIy+/fv5+233wZ0r+ny5csDBf+uPo5nfZyFEEKI51G/fv1YuXIlWq2W0NBQevXqVWh7Pz8/7O3tsbS05LXXXuPgwYO5jh88eJB+/fphbGxMpUqVaNu2LUePHi3Jp1AqlFiCrCjKIkVRYhVFOV3AcUVRlFmKokQqinJSUZTGJRXL06bRaLh8+XKRbyCfxLp16/Tl0tevX3+kcwYPHkzlypU5f/4877//frHEcevWLX1C6OjoSGxsrP6Yt7c3Bw4cKJbHyU9R4+zp6alPwjZs2MDdu3dJSEgosL/bt2/z66+/6pPqknLjxg2qVaumv+3k5MSNGzcAcHd3L3V/eIoa54sXL3Lx4kVatmxJ8+bN9UnQg+rUqcP58+eJiopCq9WyceNG/ev2wfEwMTGhXLlyJCQkEBAQgLW1NY6OjlSvXp2PP/4YOzu7J34+hY2/ra0tmZmZhb5OSkpR4/zXX39x/fr1Qi+/uHz5MhUrVmTw4ME0atSIwMBAUlNTgcJ/Vw8dOoSnpyedO3fmzJkzxfJ8Sus4CyGEEM+yBg0acOzYsXyP9evXj9WrV7Nr1y48PDzynbR40MMTZg/f1k3kvnhKcgY5BOhUyPHOgMu9r3eBuSUYy1MVHx+vn7UpCa+++ipRUVGcPHmS9u3b62eFirJ48WJu3rxJ/fr19dcQlCQHBwdu3rxZYv0XNc7Tp0/n999/p1GjRvz+++9UrVoVE5P816XTarX069ePUaNGUatWrZIKGcj/j839P0jGxsaYmZlx9+7dEo3h3yhqnLVaLREREezbt48VK1YQGBjI7du3c7WxtbVl7ty59OnTh9atW+Ps7Kz/WRQ0HkeOHMHY2JibN29y5coVvv32Wy5fvvzEz6ew8YeSf90WpLBxzsnJYcyYMXz77beF9qHVajl+/DjDhw/nr7/+wtramuDg4ELPady4MVevXuXEiRO8//779OjR47Gfw4NK6zgLIYQQz7JXXnmFzMxM5s+fr7/v6NGj/P7779SuXRt7e3uCgoLo169fkX3t3LmTxMRE0tPT2bhxIy1btsx1vE2bNqxatYrs7Gzi4uLYv38/TZs2pUyZMqXqvWpxK7FVrFVV3a8oinMhTboDS+/VmB9WFKW8oiiOqqrGlFRMJS4nBzKTsURDRnoapOuShAmf/x+/bdsBQPj/HphRzdaAJlXfTi8zBbKz8t5/j72VMeSkQ3o677zZi08+GZe7bSHnGwN9enRl2oxZDO7bs+jnpEkDbWaBsVRyqEjM5fM4OlYmJuZvHCpW0LfNuBOPpZlxgec+qaLGuYqtFeuXLwIgJSWFdWvXUs5MzTeed4eOxKVmdUYPHZT3eEE/p6IU0N6pkj3XL1/UH4++FkUV+zL625mZGVioGZBe8EJNT1NR4+xUuSLNm76EqTaVmpVtqedSm4hTx3nJO3dRyKvtW/Nqe92CZD8vDMFY1b1GnapU5nrkWZzsbdBqtdy5cxs7SyN+WbqYTq+0xlSbikMZM1o2e4mwP3+nluMjzCKrOZB+B9Lz/olzcrBl3/6D/4x/1CXatWn1z+s2LRVLpeDfv5JiaWZCRkaG/vaECRP47bffAPj99985ffo07dq1A3TbO/j7+7Np0ya8vb315zg5OeHk5ESzZs0AXcn6/QS5UqVKxMTE4OjoSExMjP5T5QfL1rt06cJ7771HfHw8FSpUeKLn4+TkxL59+/S3o6Oj9fEDZGRkYGlp+USPIYQQQrxoFEVhw4YNjB49muDgYCwsLHB2dub7778HdLPI48ePp2fPot/nt2rVirfeeovIyEj69++f6z0FQM+ePfVVZoqi8M0331C5cmXs7e0xMTHB09OTQYMGMWbMmBJ5roailOTU+b0EebOqqu75HNsMBKuqevDe7d3AJ6qq5r806z3e3t5qQau3GtzdW/BtXQCqzbhLxPs2WJgUfK3voI3pdKtrQoCbaa7790Vpmf6nhs39rfI9L+ZuDo5ldJP/G85l8fUfGg4HWhd4vqqqXEpSqWNnpFuNbmcmANM7WLDhXBZHbmQztX3+1zSGhGsIu5nN7C75v5EduyMDeyuFoFbmBB/MJDFd5Rs/XV/f/plJVg4EtTIvcAyeVGHjHJ+Wg52lgpGiMGF3BsZGMNkn7/P8bE8G5+JzWNPLEqN8rs3O7+fkOjuF8yMLXqna5j/JpHya//WyZ2Kz6b8+nSOB1ty8q+K7NJWI920wNlJISMuh1eI0zo0ofBXsp62wcd4WqWXF6SyW9LAkPi2HRj+lEj7UGnur3AUqsak5OFgbkZSu4rMkldW9LKlrb8ycIxpOxWYzr5slK09nsf5cFqt7WfH1wUzOJ+SwyN+CtCx4aX4qKwMs8ahkjO/SVJb2sKRq2fyLYJy/v0vYu9ZUsMp7PDFdpcnPKRx/VzfGjX9O4di7NthZKqiqitOMFK6OtsHEqPDr9IudpZ1unKOii7z2v127dkyfPj3Pf2QArVu3ZsGCBdSrV48vvviC1NRUpk2bxtixY/WfKgcHB5OYmMg333zD33//TaVKlfSz9gEBAVy9ehVFUfD19WXp0qVUrVo13zicnZ0JCwvLN5lOTEykSZMmHD9+HNDNVB87dgw7OzvdODs5cfXq1QKrOoQQQojS5ty5c9SvX9/QYRSLkJAQwsLCmD17tqFDyXdcFUU5pqpq3jc6T4Eh35nk9+4z32xdUZR30ZVhU7169ZKM6cmY20BH3VY0HY4t46CDN+1fcs3T7OjZq/Qc/zNJdxV+vWLE52FWnFk+EYDWw7/j/NVbpKQpOP1oxMLxb9CxuRuT5m/G27U6/q09mDU3lE0bTmJibIxdWStCpvUF58oFnu/X1JWBw2eQnJqBqqp4utRj7ti+YG3JpYRdlLXLho4d88Tp/NpEklO1aLQ5bIwyYsf3I3Gr6Ujg1OUM69EK7/o1CGqeQu/PFrJwYRLVK1VmzZRAKKtL1vfunMvU4f5QO/8318WhsHHet+c44+dtQlEU2ni5M+ej3mCmS3K9Bv6H8CWfEh2bxJQvP8O1RiUar9D9Oox8vS2B/i0L/DnF305BXfwtdMy7jdW4ORv4ZUcYaVrd+Ae++jJfBHZl04GThJ2/xuR3utEA6J29DbfFhzAxMWLOxAEYt9Dtebd3z3G6tL8CHV8vsTF7HIWNc8cOKjtmrcdtyVmMjYyY9lEv7P10f8/ujzPAB5MWcSJSdw3qpPcCqHuvzdvtsnhr8hLqLLiOXVlbVk4eAlUrMKJ1BoOnLMP9vzGoKgzu44vHG37k5OQQOW8Sdv6TwNwsVyyzVu/lm+W7+DtFwWOxMV1aNGDB+DcIO3eVeRsPsmD8G9gBE03+5KUl2+/F0h27bi0AOHbuKs2b7MCk8zslMo4FU+HUWjo4/sHBqa/RPmg5WNo+0pk3b94kMDCQLVu2APDDDz/wxhtvoNFoqFWrFosXLwZ0WzL07t2bhQsXUr16ddasWQPoto2aO3cuJiYmWFpasnLlShRF0Y1zZGS+133PmjVLn1x7eHjQpUsXFixYQFhYGPPmzWPBggXY2dkxceJEXnrpJQAmTZqk7+vYsWM0b95ckmMhhBBClD4luUQ24AycLuDYT0C/B25fAByL6vNZ2ebp+PHj+W4rVNq88cYb+q13itPff/+tvvLKK8Xe78MMMc6//vqrOnPmzBLpu2fPnur58+dLpO8nUZpez6dOnVLHjBlTIn2PGjVKv/3UU6fVqMd/fl9908NMVae7qmrkHsPEcc9zO85CCCHEYyoN2zz9W9u2bVM9PT1zffXo0cPQYeVS2rZ5MuTH95uAkYqirASaAXfUZ/n644c0atQIHx8fsrOzMTY2NnQ4BVq2bFmJ9Hvt2rUiFxQqDoYY58JWEX4SGo2GHj16UK9evRLp/0mUptezu7s73333XYn1XdKrmBfI2JRG78zCJ9mWbNPfMP5vD2g6FNp/AWb5X25Rkp7bcRZCCCFeIB07dqRjPpWiomAldg2yoigrgHZABeAW8DlgCqCq6jxFt5zpbHQrXacBg9Uirj+GUn4NshBCFIesdNj1JfxvLti7wGs/QdUmho5KCCGEeKE9T9cglyYvzDXIqqoWurb4vanzESX1+EII8cwytYTOwVCvE2x8Dxb4QZux0OZjMDYt+nwhhBBCCPFYSnIfZCGEEE+iVjsY/ic07AW/B8NCP4i7aOiohBBCCCGeW5IgCyFEaWZZXldi3XspJF2Fn1rD4bm6fdeFEEIIIUSxkgRZCCGeBW7d4b3DULMtbAuC/3aH29cNHZUQQgghnpKEhAS8vLzw8vKicuXKVK1aVX9bo9E8Uh+DBw/mwoULxRbTrl276NGjR6Ft9uzZw+HDh4vtMUuabEIphBDPijKVoP8qOL4Etn0Kc1+GLtPAow8o+W0tL4QQQogSsTUI/j5VvH1Wbqhbg6QA9vb2hIeHA/DFF19gY2PDxx9/nKvN/a2KjIzynwddvHhx8cX7iPbs2UOFChVo3rz5U3/sxyEzyEII8SxRFGgyCIb/AQ5usGEorH4LUhMMHZkQQgghDCAyMhJ3d3eGDRtG48aNiYmJ4d1338Xb25sGDRowefJkfdtWrVoRHh6OVqulfPnyBAUF4enpSYsWLYiNjS3wMd58802GDx9O69atqVu3Llu3bs3TJj4+Hn9/fzw8PHj55Zc5ffo0ly5dYsGCBUybNg0vLy/+/PPPEhmD4iQzyEII8SyyqwmDt8CfP8DeKfBjc/D/QbfytRBCCCFKViEzvYZw9uxZFi9ezLx58wAIDg7Gzs4OrVaLj48PAQEBuLm55Trnzp07tG3bluDgYD788EMWLVpEUFBQgY9x/fp1fv/9dyIiImjfvj2RkZG5jk+cOJFmzZqxadMmduzYwaBBgwgLCyMwMJAKFSowevTo4n/iJUBmkIUQ4lllZAytRsM7e8HGAVb0gU3vQ+ZdQ0cmhBBCiKeodu3avPTSS/rbK1asoHHjxjRu3Jhz585x9uzZPOdYWlrSuXNnAJo0aUJUVFShj9G7d2+MjIyoV68e1apVIyIiItfxgwcP8tZbbwHQoUMHbt68SWpq6hM+s6dPEmQhhHjWVXaHd/ZAy9Fw/L8wtyVcPWToqIQQQgjxlFhbW+v/HRERwcyZM9mzZw8nT56kU6dOZGRk5DnHzMxM/29jY2O0Wm2hj6E8tN7Jw7dVVS309rNCEmQhhHgemJiD35cweKvuOuXFnWHnJNBmGjoyIYQQQjxFycnJlClThrJlyxITE8P27duLpd81a9agqioXL17k+vXruLi45Drepk0bli9fDuhWt3ZycsLa2poyZcpw9+6zU90m1yALIcTzpEYLGHYQtk+AP2ZCxC7dPsqVGxo6MiGEEEI8BY0bN8bNzQ13d3dq1apFy5Yti6XfOnXq0KZNG2JjY/n5559zzUADTJ48mcGDB+Ph4YGNjY1+xezu3bvTq1cv1q9fz5w5c3j55ZeLJZ6SojxrU9/e3t5qWFiYocMQQojS7+J2CB0J6Ung8ym0/EB33bIQQggh/rVz585Rv359Q4dhEG+++SYBAQFF7nn8OPIbV0VRjqmq6l3sD/YIpMRaCCGeV3U7wnuHoV5n2P0lLO4CiZcNHZUQQgghRKklJdZCCPE8s7aH3kvh5GrYMhbmtoKOU3R7KT+0uIYQQgghXmyTJ09m/fr1ue7r27cvy5YtM1BET5+UWAshxIviTjRsfA+u/A4uHXX7JpepZOiohBBCiGfCuXPncHV1zbN6s3h8qqpy/vx5KbEWQghhAOWc4K2N0OlrXZL8Y3M4s9HQUQkhhBDPBAsLCxISEp7Z7YtKG1VVSUhIwMLCwtCh5CIl1kII8SIxMoLmw6D2K7DhXVgzEC70gc7fgGV5Q0cnhBBClFpOTk5ER0cTFxdn6FCeGxYWFjg5ORk6jFwkQRZCiBdRxbrw9k448C38/g1EHYTuc6C2j6EjE0IIIUolU1NTatasaegwRAmTEmshhHhRGZtCuyAI3AmmVvDfHrBlHGjSDB2ZEEIIIYRBSIIshBAvuqpNYOh+aDYMjvwEP7WBG8cMHZUQQgghxFMnCbIQQggws4LOX+sW8cpKgwV+sHcqZGcZOjIhhBBCiKdGEmQhhBD/qO0Dw/+EhgHwezAs9IO4i4aOSgghhBDiqZAEWQghRG6W5eG1n6HXEki6Cj+1hsPzICfH0JEJIYQQQpQoSZCFEELkr0EPeO8Q1GwD2z7RLeJ1J9rQUQkhhBBClBhJkIUQQhSsTGXovxpenQnRYfDjy3BiFaiqoSMTQgghhCh2kiALIYQonKJAk0Ew/CA41IcN78LqAZCaYOjIhBBCCCGKlSTIQgghHo1dLRi8Bdp/ARe2wo/N4eJ2Q0clhBBCCFFsJEEWQgjx6IyModUYeHcvWFeEX3rDplGQmWLoyIQQQgghnpgkyEIIIf69yg11SXLLD+D4UpjXEq4eMnRUQgghhBBPRBJkIYQQj8fEHPwmw+CtukW7FneGnZ+DNtPQkQkhhBBCPBZJkIUQopRIT0+nbdu2ZGdnF9hm7dq1KIpCWFhYvsdv375NQEAArq6u1K9fn0OHdLO6iYmJ+Pn54eLigp+fH0lJSQDs27ePcuXK4eXlhZeXF5MnTy4yzitXrtCsWTNcXFzo06cPGscmMPwPaDwA/vge5r+C5tpfDB48mIYNG+Lp6cm+ffvy9OPv74+7u3uu+3744Qfq1atHgwYNGDduHACnTp1i0KBBRcYlhBBCCPGkJEEWQohSYtGiRbz22msYGxvne/zu3bvMmjWLZs2aFdjHBx98QKdOnTh//jwnTpygfv36AAQHB+Pr60tERAS+vr4EBwfrz2ndujXh4eGEh4czadKkIuP85JNPGDNmDBEREdja2rJw4UIwLwP+s6DfKkiJZf57reHWWU6dCGfnzp189NFH5OTk6PtYv349NjY2ufrdu3cvoaGhnDx5kjNnzvDxxx8D0LBhQ6Kjo7l27VqRsQkhhBBCPAlJkIUQopRYvnw53bt3L/D4xIkTGTduHBYWFvkeT05OZv/+/bz99tsAmJmZUb58eQBCQ0MZOHAgAAMHDmTjxo2PFaOqquzZs4eAgID8+6rXCd47xNmsKvhanITFXXAwSaV8+fL6We+UlBS+++47Pvvss1x9z507l6CgIMzNzQFwcHDQH3v11VdZuXLlY8UshBBCCPGoJEEWQhSpqNLf1atX4+bmRoMGDejfv3++bVatWoWHh0eu0lmAzMxM+vTpQ506dWjWrBlRUVEAZGVlMXDgQBo2bEj9+vWZOnVqkXHmKf3VaPK00Wg0/7r0d+LEiXh4eODl5UWHDh24efMmAJs3b+bzzz8vMq5HodFouHz5Ms7Ozvke/+uvv7h+/TrdunUrsI/Lly9TsWJFBg8eTKNGjQgMDCQ1NRWAW7du4ejoCICjoyOxsbH68w4dOoSnpyedO3fmzJkzhcaZkJBA+fLlMTExAcDJyYkbN27kbmRdAc/XPiL0rifamDNcmdKMY0cOcf3eDPDEiRP56KOPsLKyynXaxYsXOXDgAM2aNaNt27YcPXpUf8zb25sDBw4UGpsQQgghxJOSBFkIUaTCSn8jIiKYOnUqf/zxB2fOnOH777/P0yYhIYGxY8eye/duzpw5w61bt9i9ezcACxcuxNbWlsjISMaMGcMnn3wCwJo1a8jMzOTUqVMcO3aMn376SZ88FyTf0t+HzJ8/H9Bd1/qopb9jx47l5MmThIeH061bN/11ul27dmXTpk2kpaUVGtejiI+P18/2PiwnJ4cxY8bw7bffFtqHVqvl+PHjDB8+nL/++gtra+tcpdT5ady4MVevXuXEiRO8//779OjRo9D2qqrmuU9RlDz3DXn7bZzcW+C9qhyj9xrxcuUsTP6cQfgfu4mMjKRnz575xp+UlMThw4eZNm0avXv31j+eg4OD/oMJIYQQQoiSIgmyEKJIhZX+zp8/nxEjRmBrawvkLou97/Lly9StW5eKFSsC0L59e9atWwfkLv0NCAhg9+7dqKqKoiikpqai1WpJT0/HzMyMsmXLFhhjkaW/95w9exZfX199rI9S+vvg46ampuoTQkVRaNeuHZs3by4wrkdlaWlJRkaG/vaECRP0C2fdvXuX06dP065dO5ydnTl8+DD+/v55FupycnLCyclJf41yQEAAx48fB6BSpUrExMQAEBMTo/85lS1bVv+BQJcuXcjKyiI+Pr7AOCtUqMDt27fRarUAREdHU6VKlTztTExMmDFjBuGnzhAadpPbFtVx0Z7n0PTeHPvfHzg7O9OqVSsuXrxIu3bt9PG/9tprKIpC06ZNMTIy0seSkZGBpaXlvx5XIYQQQoh/w8TQAQjxJNLT0+nUqRN79uzJM7t59epcrDHYAAAgAElEQVRVhgwZQlxcHHZ2dixbtgwnJ6dcbdLS0ujVqxeXLl3C2NiYV199VT/jtn//fkaPHs3JkydZuXKlPvEqzLFjxxg0aBDp6el06dKFmTNn5pldS0pKYsiQIVy6dAkLCwsWLVqEu7s7GRkZtGnThszMTLRaLQEBAXz55ZcA7N69m7Fjx5KTk4ONjQ0hISHUqVOH2bNnY21tzeDBg59kGAtVVOnvxYsXAWjZsiXZ2dl88cUXdOrUKVebOnXqcP78eaKionBycmLjxo368ucbN25QrVo1QJdUlStXjoSEBAICAggNDcXR0ZG0tDRmzJiBnZ1dgXE+Uukv4OnpSWhoKH379uX69escO3aM69ev07RpUz777DPe/2AM2camZOeo3LidjkabQ6Y2m2+nfMmmdSuxtinLtyHr2HY6hkxtDlq7mixYs4WUKi+Rqc3Rt9d9193WZOe+7592OWi02WRqc6hXqQza7GwyMjKwsLBgypQpTJkyRR/3g0lru3btmD59Ot7e3rmeW+XKlalWrRoXLlygXr167N69Gzc3N0BXNr5kyRKCgoJYsmSJ/gOPv//+m0qVKqEoCkeOHCEnJwd7e3sAfH19Wbp0KVWrVtU/hqIo+Pj4sHbtWvr27ZurrwelpaWhqirW1tbs3L0bE9uquE3cgtv6dxnuGQ4eHYlyG0G3gP76MvcePXqwZ88e2rVrx8WLF9FoNFSoUEH/Ont4xWshhBBCiOImCbJ4phVW+vvxxx8zYMAABg4cyJ49exg/fjz//e9/823n4+ODRqPB19eXrVu30rlzZ6pXr05ISAjTp09/5HiGDx/Ozz//TPPmzenSpQvbtm2jc+fOudr85z//wcvLiw0bNnD+/HlGjBjB7t27MTc3Z8+ePdjY2JCVlUWrVq3o3LkzzZs3Z/jw4YSGhlK/fn1+/PFHvvrqK0JCQhgyZAgtW7Ys0QS5sNJf0JXFRkREsG/fPqKjo2ndujWnT5/OdY6trS1z586lT58+GBkZ8fLLL3P58mVAN/Orzc4hTaMlMytHl5gmpbP78F/czcxh8+EzJCQkMqxPNyxqeGHn6KRLOnMlmtkkxseTlKbhi01n0GTnEB9zk5jkTN5dGoYm+5/2GZl1OR27k/LVXTEr74DqUJcP15ziw60xxO37k40Wfmj3HCA2LoWWwXv+eaLWr2A54BWSDq1m8NivKN/6DQDSrySTcjKCi5vP6puaGCmYmRhhbmJ077sxZiZGmBkbYW6q+17W0lR/GxV+OxWDo1tTDh48SPv27R/553Pz5k0CAwPZsmULoNsm6Y033kCj0VCrVi0WL14MQFBQEL1792bhwoVUr16dNWvWALpto+bOnYuJiQmWlpasXLkSRVHIyckhMjIy3w8lvv76a/r27ctnn31Go0aN9IuCbdq0ibCwMCZPnkxsbCwdO3bEyMiIqlWr6n73KtaAwF2wfzrsnwbH94BGq+93yJAhDBkyBHd3d8zMzFiyZIn+A6a9e/fStWvXRx4XIYQQQojHIQmyeKYtX76cX375Jd9jZ8+eZcaMGQD4+Pjke22llZUVPj4+gG7F38aNGxMdHQ2gnzE1Mnq0KxFiYmJITk6mRYsWAAwYMICNGzfmSZDPnj3L+PHjAXB1dSUqKopbt25RqVIlfalrVlYWWVlZuUp5k5OTAbhz546+pNXKygpnZ2eOHDlC06ZNHynO/Gizc0jLyiZDk0161r0vje4rLuEuSXdTCQ2/QaY2h+Wzv+avP/agqvDxvFDiVRuq1GjI5N8uoNHmoJSvQuCsTZSvUV8/W6pLZMtj0/trMrU5rDv0K2lxptSfuI2r6ea0+Gw15lXro+ZkE30rAf8FJ0jcOQ/zKvXoM1+3UFNimZp89ON6rOu3zvc5qKrK7dt3WBt2FQszUzQ3LqJaludaYpo+WbUwNaKshRWdA4P0962YOJBXWzcm+twx9iZFcTdkKDk52eTcTsBq51dMWbBGl+DeS2YTulbl48B+bBo1DzMTI/bvSmcLZ1g4qYMuCTYxwtgo7zW5RbFZe5Jlt9ry/Y/zi0yQH1xYrEqVKvrkGMDLyyvfPZLt7e31130/aOTIkYwcOTLP/WfPnuX111/Pt6y5Vq1aHDlyJM/9/v7++Pv7A7rfnwsXLuQN3tgUfMaDSwecN7zL6f6RsDUI2n+OmZkly5Yty3NKZmYmYWFh+V7fLoQQQghRnCRBFs+sokp/PT09WbduHR988AEbNmzg7t27JCQk6MtHH3b79m1+/fVXPvjgg8eK58aNG7lKuAsr8V2/fj2tWrXiyJEjXL16lejoaCpVqkR2djZNmjQhMjKSESNG0KxZM7Kyc/h+zjw6d+6CuYUFVjZlWLxhO39ExpOuycamal1++OVX/LIdSLuX1GY8kOTmm/g+8D0jK5us7LwLLz0o4W4Go5YdQTExg8qdMXpdl/TP2HURjY0b5/fu5YrdSyiZd4mNjiJOKU/a7Qz9bGk5S1OyU29T3rECanoKESe20nvsdBxrVCcssRux147xVmBPwvf9xl8vt2Zsby9+zfQk5uolggZ5o2oyeGfdNaaN/IKGHh4M7vUqP/y0gBrVqulnaE2NFXpf7sDrbsn07duXYcPW4fHBYN57r02u55Kr9HfnTv7nUJbZI/wBf5itK2mPioqiW7dunD52CNAtRObi4gLAD9tW4tWwAW5VdNclh8ZcxbuRJ+WsTB/rdXPfhG712R8Rx+XwG6RmaLC2MHui/p6Uu7s73333Xck9gFMTGHoAdn0B/5sLl3ZDz3lQtUmepteuXSM4OFhfPi+EEEIIUVLk3YZ4ZhVV+jt9+nRGjhxJSEgIbdq0oWrVqgW+wdZqtfTr149Ro0ZRq1atR3p8VVXJylb1iWZ0UhqpmVqOXU0iXZPN0ahEYu9msvroddKzsnXJa1Y2Jo17snvuFByc62HrVAfbanWZsPEsFoczyMjKpmz/Gbgk32HOss/5JbEGxvY1iN0whXLdJmBepR6J/1tH1/7vYt95FAB3I9PISrzIAYt/tucxNzHC0swYK1NjLMyMsTTVfdmYm1DRxhzLe/dZmBrr21ma3bt979/325ibGPHV1U74NzPhFV8fzB8sGTYxwkjpwkcffcS2laMxNjZmybxZ9O2rm0X08vIiPDwcgH79+rH3xAkAZk79gr59ddsVZfh+yltvvcX0tztiZ2fHypUrqVXLiU71xjN48GDef/0VVFVl+LuBvO7XkpycHKKvXqG+c1UsLc1z/Uweu/S3CEFBQVy4cAEjIyNq1KjBvHnz9Mf27t37SFtQFaWshSnBr3sw8E4GP+y9TFBn1yfus9Qzs4Iu3+j2Tt44Ahb4Qdtx0Poj3UzzPS4uLvoPKIQQQgghSpKS35YdpZm3t7eaX/mgePEkJSXRqFEj/dY/EyZM4LfffgN0e8ZqsnP0yWt84h38Xm7Cb4dP6WdZ78+epmuy+emrsZiYW9LpnfH3ZldzSM/Skq7JZt/8L6ng1gK7Bm1Iy9KSrsnRz9Bm5/zz+6NNSeTWik+p+o4ueUo9+zsZ105h3yl3+aqFqZE+YbUwNeLw1P50+GwpZcqWxcrsnwT18NqfsLKyplWX15nxfm9mrN+Ppakxd+Nj+Py9N1m1408szYxZs2Q+dxLj+fzLybok18QYo8co8S3MX3/9xXffffdIyWRJO336NIsWLSrZ2c1HdOvWLfr3759v6fLjClp3ktVh11k7/GUaV7cttn5LvfQk2DIOTq2GKo2h509Qsa6hoxJCCCGEASiKckxVVe+iWxY/mUEWpUpOjsrdDC130rMK/UpOzyJVoyU2OZ1uM3ajwYS0sr6YBLQjXZNN7U+3kJV6ByPLMiiKEUn7l6LUasPrcw/lecyk/f9FmxBDhR5BLDt8TT9zev+7qoKVmTHOFaywMjPBwtSYBaO6M27hlnvtTO59N+KTffaMbqTQ5KVmfPT29wwaN5ROnX2wNDXGyswEcxMjkpPvYGVlhZmZGfPnz8epqx9Lx/gRFxeHqakp5cuXJz09nQ4zTjL8k0/o1KkJ376byku2mdStW5eFC7fTxNOdZrV0peLz/75Gy5Ytsbcxz/PcikujRo3w8fEhOzs73wXRnqYSL/39F65du1bk3sT/1oSu9dl/MY6xa07w26jWWJgadryfGktbeH0+uHaBzWPgp9bgNxleegcecR0AIYQQQognJTPIJaSw7YcK2z6oU6dOHD58mFatWhW4t2pISAhjx47Vb70ycuRIAgMDCz3/7bffJiwsDFVVqVu3LiEhIfoFoQoydepUFi5ciLGxMbNmzaJjx4552uzZs4ePP/4YjUZDkyZNWLhwIUZGxqxcu57/+/JzVBQwMmbIR1/gVL8x0TF/s+Crj3j90x9JTs/idrrmn8Q3LYu7mVoKe0maGiuUszSlrIUpNhYmnFrxNXVadKS2ZwuszHKXE58/tINfF36LkWKEe5NmvD8xmHI2VliZGfNW17Zs3nuIpLgYmrrXpV49VywszHON59GjR+nZsydJSUlYWFhQuXJlzpw5Q3x8PC1btsx3AaKwsDD9Nk+dO3fmhx9+QFEUfUnusGHDOHToEAMGDMDY2Bg3NzcWLlyIra0tJ0+eZODAgWRnZ5OTk0Pv3r2ZNGkSABs2bGDSpEkYGRlha2vLokWL9KXgjRs3ZseOHfrtcMSz70BEHG8tPMLQNrUY36W+ocN5+u7+DZveh4gdUKsddJ8D5ZyKOksIIYQQzwlDziBLglxC5syZg1arzXfBp6ioKJKTk5k+fTr+/v65EuTdu3eTlpbGTz/9VGiCHBYWxuzZs/McK+j85ORkypbVLSr04Ycf4uDgQFBQUIHxnzp9hn79+rF++z4io67zTt/u/Bh6gBRNzj8JbWoms4d2wmfMLNRyjpwM/ZkcqwqYNmhPdmY6iqkFiqKgib1CXOjX+tLjxK3fU6VpV6q4elHO0jTfr7L53FfeyhRLU+Nc+wobovR38+bNXL58mVGjRj21xyxIaSp9FsVr/PpTrDp67cUrtb5PVeFYCGyfAEYm0HU6NOwFSvFePiCEEEKI0kdKrJ9DhW0/VNj2Qb6+vrm2cPm3Hjw/O0flbsb9suQc7tyK43aahpNRsdikGjN16zmSHypdvp2m+x697xewa0SHWbqS5ERjO4Z9twrzqrrZLFNjBaucVHIUY0ztqlLO0pRmrdoR/msIw0cOy5XgXjtnzNQDVhwc/wrlLE3Z0TyLHTt28ONHox/7ed5niNLfbt26PZXHeRTx8fH83//9n6HDECXg0y6u7L8Yx8drTrDlRSq1vk9RwHsw1GoLG4bB+nfg/Gbo9j1Y5d2bWQghhBCiOEiCXAKK2n7o38jOUfMkseHXkli2cjUbtu7C1rEGbd76CNXaXn/82unjXL0YR+1Pt+TqK/6370m/HIZphWo4BPTk7MGoe4msCeUsTaloY06dijaUszRl93Etrp5N6Njdg3KWpsy+7IZvWwd699IluZb33qw7//IxnzY1x9vbmw8+WM55zW0+7FAP0JUFjxo/ntjYWH777Tccy+n2U33ppZeYOHHiE4/NfUOGDCm2vp41fn5+hg5BlJAyFqZ8/boHby78H9/uuMCErm6GDskw7GrB4K3wx0zY+x+4dhj8f4C6eS/5EEIIIYR4UpIgl4CHtx/KL8m9k57FpbgUtp2OIcLqXK77L588ycXIeBp+vp27mdo8/WenV6bcoJ9RTEyJPbGVZV9/wksjZlDO0hSHMuZYVi5LZjkLhvm65C5THroKGzMjvvn8E1q43ubdwNdylSvneg47ytOingO9vKsBsKqMOU621vok976VK1cyZswYMjMz6dChQ65tlHr27EnPnj3Zv38/EydOZNeuXQA4ODhw8+bNJx5nIZ53rVwq8Eaz6iw4eIVO7pVpUuMFnTk1MobWH0Kd9rBhKPzSG5oMgg5TwLzwtRSEEEIIIf4NSZCLUWKqhk/XnyIuPoFrsXdoGbyH5PQsru1cRPqlowBUGfyDvn38tducN72JrTYqVyJra21GeSszXm/ilO+1uA9+mRp1xN7ent0ftdP3u29fGtOPl+FDv/y3SBn8Vn+mTZvG0HfeLvC5ODk5cf36df3t6OhoqlSpkqddixYtOHDgAAA7duzg4sWLedq0adOGS5cuER8fT4UKFcjIyMDS0jJPOyFEXuO71GffhTjGrjnJlg9ewFLrBzl6wLv7YM9X8OcPcHmfbjuo6s0NHJgQQgghnheSIBcjY0UhMi6FcpY2KOTQuKo19uVsKNvq/yj/YGJrpfs+8fpKur/aiP59O+fqZ98+DdNPl+UL/wb5Pk5MTAyVyzkCujLm+vULX+VWVVUuXbpEnTp1UFWVX3/9FVdXV/35R44cYerUqbnO8ff3p3///nz44YfcvHmTiIgImjZtmqfv2NhYHBwcyMzM5Ouvv2bChAkAREZGUrt2bRRF4fjx42g0GuztddsSXbx4EXd390cYUSGEjbkJ0wI86L/gf0zffoHPur2gpdb3mZhDh/+Dep11s8mLO0PLD6DdeN0xIYQQQognIAlyMSpnZcquD9sC8HZYN7o73qV9+2Z52h09epRX7m0ftGvbFqb835ecOXMGgNatW3P+/HlSUlJwcnJi4cKFdOzYkUmTJuHt7Y2/vz+zZs1i06ZNmJiYYGdnR0hIiL7v/M738/Nj4MCBJCcno6oqnp6ezJ07F4BLly7pV7d+UIMGDejduzdubm6YmJgwZ84c/SJYXbp0YcGCBVSpUoVp06axefNmcnJyGD58OK+88goA69atY+nSpZiammJpacmqVav05dx79+6la9euxTfwQjznXq5TgTebV2fhH7pSa2/nF7TU+kE1Xobhf8L2T+HgDIjYCa/9DJXy/2BRCCGEEOJRyDZPJeRZ2X7nzTffZMaMGVSsWPGpPWabNm0IDQ3F1vYF3LpGiMeUmqml4/f7MTFS2PpBGyzNXuBS64dd2KrbNznjDrzyGbQYqbtuuQCF7VP/3XffsWDBAkxMTKhYsSKLFi2iRo0aefro1KkTMTExaLVaWrdurf8QMTExkT59+hAVFYWzszOrV6/G1taW0NBQJk6ciJGRESYmJnz//fe0atWq0Kd17Ngx/b7qXbp0YebMmXnWjUhKSmLIkCFcunQJCwsLFi1apK/QGTJkCJs3b8bBwYHTp0/rzzlx4gTDhg0jJSUFZ2dnli9fTtmyZTl16hTffvttrg9dhRBCCEMw5DZPefcZEsXiwe2HSrNly5Y91eQ4Li6ODz/8UJJjIf4la3MTvgnwICohjWnbLxg6nNKlXmd477BuZeudkyCkGyRFFdh80aJFvPbaa/luDdeoUSPCwsI4efIkAQEBjBs3Lt8+Vq9ezYkTJzh9+jRxcXGsWbMGgODgYHx9fYmIiMDX15fg4GBAtwXfiRMnCA8PZ9GiRQQGBhb5tIYPH87PP/9MREQEERERbNu2LU+b//znP3h5eXHy5EmWLl3KBx98oD82aNCgfM8JDAwkODiYU6dO0bNnT6ZNmwZAw4YNiY6O5tq1a0XGJoQQQjyvJEEuQUOGDHlqe/M+KypWrEiPHj0MHYYQz6SXa1dgQIsaLP7zCkeuJBo6nNLFugL0/i/0mAe3TsPclnBsCeRTJbV8+XK6d++ebzc+Pj5YWVkB0Lx5c6Kjo/Ntd//SFK1Wi0aj0c/shoaGMnDgQAAGDhzIxo0bAbCxsdG3SU1NLXAHgftiYmJITk6mRYsWKIrCgAED9H096OzZs/j6+gLg6upKVFQUt27dAnTVOnZ2ecvxL1y4QJs2bQDdVnHr1q3TH3v11VdZuXJlobEJIYQQzzNJkIUQ4hnySSdXnGwtGbf2BOma0l2h8tQpCnj1012bXKUR/DoKVvSFu7f0Tf7NPvULFy6kc+fOBR7v2LEjDg4OlClThoCAAABu3bqFo6NuEUVHR0diY2P17Tds2ICrqytdu3Zl0aJFhT72jRs3cHJy0t92cnLixo0bedp5enqyfv16AI4cOcLVq1cLTOrvc3d3Z9OmTQCsWbMm144F3t7e+p0JhBBCiBeRJMhCCPEMsTY34ZvXPYlKSOOb7ecNHU7pVL4aDNgEHafCpb3wY3M4q0sIH96nviDLli0jLCyMsWPHFthm+/btxMTEkJmZyZ49e4rss2fPnpw/f56NGzcyceLEQtvmtz5IfrPOQUFBJCUl4eXlxQ8//ECjRo1y7Uefn0WLFjFnzhyaNGnC3bt3MTMz0x+TfeqFEEK86CRBFkKIZ0yL2vYMbFGDxX9E8b/LCYYOp3QyMoIW78HQ/VC+Oqx+C9YPxVLJIiMjQ99swoQJeHl54eXlpb9v165dTJkyhU2bNmFuXvjWURYWFvj7+xMaGgpApUqViImJAXRl0g4ODnnOeXBv+II4OTnlmgkuaC/6smXLsnjxYsLDw1m6dClxcXHUrFmz0JhdXV3ZsWMHx44do1+/ftSuXVt/TPapF0II8aKTBFkIIZ5Bn3R2pbqdFWPXniRNozV0OKWXgysE7oK2n8CpNdiu6ka2VqtPkqdMmUJ4eDjh4eGAbgeCoUOHsmnTpnyTW4CUlBR9EqzVatmyZYt+b3l/f3+WLFkCwJIlS/TXOkdGRupnhR/eG/7+uQ9ydHSkTJkyHD58GFVVWbp0ab7XTd++fRuNRgPAggULaNOmTb5b9z3oftl3Tk4OX331FcOGDdMfk33qhRBCvOgkQRZCiGeQlZkJ0wI8uJaYxjfbZFXrQhmbgs+n0Hc53L5Gh6auHDx4MN+mY8eOJSUlhV69euHl5YW/v7/+2P1Z5tTUVPz9/fHw8MDT0xMHBwd9khkUFMTOnTtxcXFh586dBAUFAbq94d3d3fHy8mLEiBH6veHj4+PzLacGmDt3LoGBgdSpU4fatWvrr4eeN28e8+bNA+DcuXM0aNAAV1dXtm7dysyZM/Xn9+vXjxYtWnDhwgWcnJxYuHAhACtWrKBu3bq4urpSpUoVBg8erD9H9qkXQgjxopN9kIUQ4hn2xaYzhPwZxcp3m9O8lr2hwyndsrXwTS3+snyZ746Zlop96jdv3szly5cZNWqUoUMhMzOTtm3bcvDgwSKvYxZCCCFKkuyDLIQQ4rGM61SPGvZWjFt7ktRMKbUulLEJ1GpDo5xT+LRrVyr2qe/WrVupSI4Brl27RnBwsCTHQgghXmiSIAshxDNMV2rtyfWkNL7eJqtaF6m2LyRHM8S/lexT/xAXFxfatWtn6DCEEEIIg5IEWQghnnFNa9ox+OWaLD10lT8vFbwysgDq+Oq+R+42bBxCCCGEKJUkQRZCiOfA2I71qFnBWkqti1K+Oti7wCVJkIUQQgiRlyTIQgjxHLA0M2ZagAc3bqcTvFVKrQtVxxei/oCsjKLbCiGEEOKFIgmyEEI8J7yd7RjSsib/PXyVPyOl1LpAtX1Bmw7X/jR0JEIIIYQoZSRBFkKI58jHHXSl1mPXniRFSq3z59wSjM3kOmQhhBBC5CEJshBCPEcszYyZ3suDm3fSmbrlnKHDKZ3MrKF6C7i0x9CRCCGEEKKUkQRZCCGeM01q2BHYqibL/3eNP6TUOn+1X4HYs5B809CRCCGEEKIUKdEEWVGUToqiXFAUJVJRlKB8jtdQFGW3oignFUXZpyiKU0nGI4QQL4qPOtSjVkXdqtZSap2P+9s9Xdpr2DiEEEIIUaqUWIKsKIoxMAfoDLgB/RRFcXuo2XRgqaqqHsBkYGpJxSOEEC8SC1NjpgV4EnMnnf9IqXVeldzBppJs9ySEEEKIXEpyBrkpEKmq6mVVVTXASqD7Q23cgPvvTvbmc1wIIcRjalLDlsDWtfjlf9c4EBFn6HBKF0XRlVlf2gs52YaORgghhBClREkmyFWB6w/cjr5334NOAK/f+3dPoIyiKPYPd6QoyruKooQpihIWFydv8oQQ4lF96FeX2hWt+WTtSe5mZBk6nNKlti+kJ0JMuKEjEUIIIUQpUZIJspLPfepDtz8G2iqK8hfQFrgB5LlYTlXVn1VV9VZV1btixYrFH6kQQjynLEyNmd7Lk7+TM6TU+mG1fQAFImU1ayGEEELolGSCHA1Ue+C2E5BruVBVVW+qqvqaqqqNgAn37rtTgjEJIcQLp1F1W95pU4sVR66z/6JU4ehZVwBHT7kOWQghhBB6JZkgHwVcFEWpqSiKGdAX2PRgA0VRKiiKcj+G8cCiEoxHCCFeWGPa16WOgw2frDtJspRa/6OOL1w/Ahny2awQQgghSjBBVlVVC4wEtgPngNWqqp5RFGWyoij+95q1Ay4oinIRqARMKal4hBDiRXa/1PpWcgZTNkuptV5tX1Cz4cp+Q0cihBBCiFLApCQ7V1V1C7DlofsmPfDvtcDakoxBCCGEjle18gxtW5u5+y7RuWFl2tVzMHRIhletKZiVgcjdUP9VQ0cjhBBCCAMryRJrIYQQpczo9i64ONgwfv0pKbUGMDaFmm101yGrD68jKYQQQogXjSTIQgjxAjE30ZVax97N5KvNZw0dTulQ5xW4fQ0SLhk6EiGEEEIYmCTIQgjxgvGsVp6hbWqxOiyavedjDR2O4dX21X2X1ayFEEKIF54kyEII8QL6oL0LdSvZELT+JHfSX/BSa7uaYFsTLsl+yEIIIcSLThJkIYR4Ad0vtY5P0fB/Umqt2+7pygHQagwdiRBCCCEMSBJkIYR4QXk4lWd429qsPRbNnvO3DB2OYdX2haxUuH7Y0JEIIYQQwoAkQRZCiBfY+751/p+9O4+Purr3P/4+M9lJAgQSICQZYCYKyC6SuCKZul6LdReN195fteqttre3tT/aWmtXbattf7fX28XWXiuo4MLiUlesKwHDKiJbwpKQAIGwJSFkO78/JqQs2YD5ziQzr+fjkcck3znfcz6Mirz5nkVnDkrRd1/6VPvroniq9fALJVdM4LgnAAAQtQjIABDFjp5q/aNXPgt3OeETnyJl57NRFwAAUZcjsjkAACAASURBVI6ADABRbmxWX/37xV69tHy73l4bxVOtfQXSjk+lGnb2BgAgWhGQAQC6ryBXIwen6HvzoniqddtxT+xmDQBAtCIgAwAUF+PSozeMV3Vtg370cpROtR48TkoayDpkAACiGAEZACBJGjO0r/59mk8vrdiut6JxqrXLJXkLAk+QW1rCXQ0AAAgDAjIAoM2903waNSRV35v3qfbVReGZwD6/VLdb2rE63JUAAIAwICADANoEplqP097aBj20MAqnWnsLAq/sZg0AQFQiIAMAjnFWZl/dW+DT/JUVevOzHeEuJ7SSM6TBY6VNbNQFAEA0IiADAE7wtWk+jR6Squ/NW6O9tVE21drrl8qKpMMHw10JAAAIMQIyAOAEse7Artb76hr0ULTtau0tkFqapC0fhrsSAAAQYgRkAEC7Rmem6r6CXC1YWaHX10TRVOucfCk2ieOeAACIQgRkAECH/n2aV2dlpuqB+Z+qOlqmWsfES8MuZKMuAACiEAEZANChI1Ot9x9q1A+jaVdrn1+qLpWqN4e7EgAAEEIEZABAp0YNSdXXC3L18qoKvb6mMtzlhIbXH3jlKTIAAFGFgAwA6NLdF3s1Zmiqvj9vjfbUHA53Oc4b4JX65XDcEwAAUYaADADo0pGp1gfqG/VgNEy1NibwFHnz+1JzY7irAQAAIUJABgB0y8jBqfqGP1evrq7Ua59GwVRrn19qOCiVLQ13JQAAIEQIyACAbrt7qldjh/bVD+ZHwVTr4RdJxs06ZAAAoggBGQDQbTFulx67cbwO1jfpwQURPtU6oa+UPYXzkAEAiCIEZADASTljUIq+8YVcvfpppV5ZXRHucpzl9UuVq6Ta3eGuBAAAhAABGQBw0u66aITGZ/XVgws+0+5InmrtK5BkpZJ3w10JAAAIAQIyAOCkxbTual1T36QfzF8ja224S3LGkAlSYn/WIQMAECUIyACAU5I7KEXfvOQM/X3NDr2yOkJ3tXa5pRHTpJJFUqT+JQAAAGhDQAYAnLI7Lxyu8dn99OCCNao6GKFTrX1+qWantDPCNyUDAAAEZADAqYtxu/TYDeNU29CsB+Z/GplTrb0FgVemWQMAEPEIyACA0+LLSNF/XnKG3vhspxauisBdrVMzpYzRHPcEAEAUICADAE7bnReO0ITsfvrhws+062B9uMsJPm+BtG2x1FAb7koAAICDCMgAgNPmdhk9esN41TU064F5Ebirtc8vNTdIWz4KdyUAAMBBBGQAQFD4MpL17UvP0JtrI3Cqdc55Ukwi65ABAIhwBGQAQNB85YIRmpTTTw8u+Ey7DkTQVOvYBGnY+axDBgAgwhGQAQBB43YZ/eqG8apvbNb3Im2qtdcv7dko7dsW7koAAIBDCMgAgKDypifr25eeqbc/36n5K7eHu5zg8fkDrzxFBgAgYhGQAQBB938uGK6zPf310MK1kTPVeuAZUmoW65ABAIhgBGQAQNC5XUa/un5c61TrTyNjqrUxkq9AKn1fam4KdzUAAMABBGQAgCNGpCfr/svO1Nuf79K8FREy1dpbIB3eL21fFu5KAACAAwjIAADH/Nv5wzXZ018PLfxMOyNhqvWIiyXjYpo1AAARioAMAHDMkV2tG5pb9N2XImCqdWJ/aejZbNQFAECEIiADABw1fGAffeeykVq0bpdeXB4BU629fqliuVRXHe5KAABAkBGQAQCO+/J5wzRlWJp+9PJn2rG/l0+19vkl2yKV/iPclQAAgCAjIAMAHOdyGf3y+nFqbG7Rd19a3bunWmdOkhL6sg4ZAIAIREAGAITEsIF99H8vH6l311fp+WXl4S7n1LljApt1bVok9eagDwAATkBABgCEzO3nDtOU4Wn6yctrVbn/ULjLOXVev3SwQqpaF+5KAABAEBGQAQAh43IZ/er6cWpqsZr5Yi/e1drnD7yymzUAABGFgAwACCnPgD6aecVIvbehSs8X99Kp1n2zpIFnsg4ZAIAIQ0AGAITcbfke5Y9I009eWauKfb10qrXPL239WGrspfUDAIATEJABACHnchn98rrxarZWM1/qpVOtvX6pqV7a+lG4KwEAAEFCQAYAhEXOgCTNvGKk3t9QpTmflIW7nJPnOU9yxwd2swYAABGBgAwACJvCPI/OHTFAP331c23vbVOt45Ikz7lSCQEZAIBIQUAGAISNy2X0y+vHqcVazXxxde+bau31S1WfS/u3h7sSAAAQBARkAEBYZacl6btXjtIHG3frud421frIcU88RQYAICIQkAEAYXfrlByd5x2gn76yVuV768JdTvdljJZShnDcEwAAEYKADAAIO5fL6BfXjZMkzXyxF+1qbYzkLZBK3pVamsNdDQAAOE0EZABAj5CdlqTv/csofbhpt55Zui3c5XSft0Cq3ydVrAh3JQAA4DQRkAEAPcYtU3J0gW+gfv7q5yqr7iVTrb0Fkoy0iWnWAAD0dgRkAECPYYzRI9eNlTFG//fF1Wpp6QVTrZPSpMyJrEMGACACEJABAD1KVv8kfe/KUfq4ZI9m95ap1j6/VF4sHdoX7koAAMBpICADAHqcGVOydWHuQD38Wi+Zau31S7ZZ2vxeuCsBAACngYAMAOhxAlOtx8lljL7zQi+Yap01WYpPZR0yAAC9HAEZANAjDe2XqAf+ZZQWl+7R7CVbw11O59yx0vCLpJJFUm85ogoAAJyAgAwA6LFuOidbF52Rrp+/tk7b9vTwqdbeAml/mbR7Y7grAQAAp4iADADosYwxeuTasYpxGd3/wqqePdXa5w+8liwKbx0AAOCUORqQjTGXG2PWG2M2GWNmtvN+jjHmXWPMCmPMamPMlU7WAwDofTL7JeqBq0ZpyeZqPV3Ug6da9x8mpXk57gkAgF7MsYBsjHFLelzSFZJGS5phjBl9XLMHJM211k6UdLOk/3GqHgBA73Xj5GxNPSNdj/x9nbbuqQ13OR3z+aUtH0pNh8NdCQAAOAVOPkGeImmTtbbUWtsg6TlJVx/XxkpKbf2+r6QKB+sBAPRSgV2txyrGbXR/T97V2uuXGuukbYvDXQkAADgFTgbkoZLKjvq5vPXa0R6SVGiMKZf0mqT72uvIGPNVY0yxMaa4qqrKiVoBAD3ckL6J+sFVo7V0c7WeWrwl3OW0b9gFkiuW454AAOilnAzIpp1rx/+V/wxJ/2utzZJ0paSnjTEn1GSt/ZO1drK1dnJ6eroDpQIAeoMbzs7StDPT9YvX12nL7h441To+WcrJZ6MuAAB6KScDcrmk7KN+ztKJU6i/ImmuJFlrF0tKkDTQwZoAAL2YMUYPXztOsW6XvtNTp1r7/NLONdLBHeGuBAAAnCQnA/InknKNMcONMXEKbMK18Lg22yT5JckYM0qBgMwcagBAhwb3TdAPv3iWlm6p1v9+vCXc5ZzIy3FPAAD0Vo4FZGttk6R7Jb0h6XMFdqv+zBjzY2PM9NZm35J0pzFmlaRnJX3ZWtsDHwcAAHqS6yYNVcHIDP3yjXXa3NOmWg8aI/XJYB0yAAC9kOlteXTy5Mm2uLg43GUAAMJs54F6XfLr93TGoBTNuetcuV3tbX0RJi/dJW16S/r2Jsnl5GQtAAAijzFmmbV2cjjG5v/aAIBeaVBqYKp18da9+utHm8NdzrF8fqluj1S5MtyVAACAk0BABgD0WtdOGir/yAz96o31Kq2qCXc5/zRiWuC1hGnWAAD0JgRkAECvZYzRz68dq4RYt+5/YbWae8qu1snp0uBxUsm74a4EAACcBAIyAKBXG5SaoIemj9aynjbV2ueXypZI9QfCXQkAAOgmAjIAoNf70oSh+sKoQfrVG+tV0lOmWnv9UkuTtOWDcFcCAAC6iYAMAOj1jDH6+TVjAlOtn1/VM6ZaZ+dJcckc9wQAQC9CQAYARISM1AT9aPpZWr5tn/7yYWm4y5Fi4qRhF7JRFwAAvQgBGQAQMa6ekKlLRw/So29u0KZdPWCqtc8v7d0i7SkJdyUAAKAbCMgAgIhhjNFPrxmjpDi3vt0Tplp7CwKvJYvCWwcAAOgWAjIAIKJkpASmWq8s26cnPgjzVOsBXqn/MNYhAwDQSxCQAQARZ/r4TF121iD9+q0N2rTr4DHvHTp0SFOnTlVzc/MJ9/3617/W6NGjNW7cOPn9fm3durXd/i+//HKNHz9eZ511lu6+++62vqqrq3XJJZcoNzdXl1xyifbu3St5/Vrw2psaN3asJkyYoMmTJ+vDDz/s8tewbNkyjR07Vj6fT1//+tdl7YlPw3/1q19pwoQJmjBhgsaMGSO3263q6mqtX7++7fqECROUmpqq3/72t5Kkb3/721q0iCfaAAC0h4AMAIg4xhj99Etj1SfOrW89v1pNzS1t7z355JO69tpr5Xa7T7hv4sSJKi4u1urVq3X99dfrO9/5Trv9z507V6tWrdKaNWtUVVWl559/XpL0yCOPyO/3a+PGjfL7/XrkkUckn1/+nCatWvC4Vq5cqSeffFJ33HFHl7+Ge+65R3/605+0ceNGbdy4Ua+//voJbe6//36tXLlSK1eu1MMPP6ypU6cqLS1NZ555Ztv1ZcuWKSkpSddcc40k6b777gvUBQAATkBABgBEpPSUeP346jFaVbZPT3ywue367NmzdfXVV7d7z7Rp05SUlCRJys/PV3l5ebvtUlNTJUlNTU1qaGiQMUaStGDBAt1+++2SpNtvv13z58+Xhl2o5IRYmdLAU9va2tq29h2prKzUgQMHdO6558oYo3/9138N9NWJZ599VjNmzDjh+jvvvCOv1yuPxyNJ8ng82rNnj3bs2NFpfwAARCMCMgAgYl01boiuGDNYv3lrgzbuPKiGhgaVlpZq2LBhXd77l7/8RVdccUWH71922WXKyMhQSkqKrr/+eknSzp07NWTIEEnSkCFDtGvXLikhVcrO07wX5mjkyJH6l3/5Fz355JOdjr19+3ZlZWW1/ZyVlaXt27d32L6urk6vv/66rrvuuhPee+65504IzpMmTdJHH33UaQ0AAESjLgOyMSbJGPMDY8wTrT/nGmOucr40AABOjzFGP/nSGCUnxOhbz6/Sjp271K9fvy7vmzVrloqLi3X//fd32OaNN95QZWWlDh8+3PWaXu80XTO4QuuK39f8+fP1gx/8oNPm7a037uyp88svv6zzzz9faWlpx1xvaGjQwoULdcMNNxxzPSMjQxUVFZ3XDABAFOrOE+S/Sjos6dzWn8sl/dSxigAACKKByfH68dVnaXX5fs1ZsVP19fVt733/+99v28jqiLfffls/+9nPtHDhQsXHx3fad0JCgqZPn64FCxZIkgYNGqTKykpJgWnSGRkZgYZef+C19B+66KKLVFJSot27d3fYb1ZW1jHTu8vLy5WZmdlh+/aeEkvS3//+d02aNEmDBg065np9fb0SExM7/bUBABCNuhOQvdbaX0pqlCRr7SFJnS+eAgCgB7lqXKauHDtYTyzZpYbGpraQ/LOf/axtMytJWrFihe666y4tXLjwn+H2ODU1NW0huKmpSa+99ppGjhwpSZo+fbqeeuopSdJTTz3VttZ5U22ybGKatOkdLV++XA0NDRowYIAktd17tCFDhiglJUVFRUWy1upvf/tbh+um9+/fr/fee6/d9ztal7xhwwaNGTOm4w8MAIAo1Z2A3GCMSZRkJckY41XgiTIAAL3G964cpaaWFmWNyevwmKX7779fNTU1uuGGGzRhwgRNnz697b0jT5lra2s1ffp0jRs3TuPHj1dGRobuvvtuSdLMmTP11ltvKTc3V2+99ZZmzpwpSXpx3jyN+e99mnDv/+prX/ua5syZI2OMdu/e3e50akn6/e9/rzvuuEM+n09er7dtPfQf/vAH/eEPf2hrN2/ePF166aXq06fPMffX1dXprbfe0rXXXnvM9cbGRm3atEmTJ08+mY8PAICoYDr6H3NbA2MukfSApNGS3pR0vqQvW2v/4Xh17Zg8ebItLi4Ox9AAgF7ujqc+0cdLijVx/0eaPXtWaAdf+Yw0/x7prg+kIeMkSa+88opKS0v19a9/PWRlzJs3T8uXL9dPfvKTkI0JAMDJMMYss9aG5W9yYzp70wR2BFkn6VpJ+QpMrf6GtbbjhVMAAPRQt+Z79PbnuzRg0CE1Nze3exayY7wFgdeSd9oC8lVXhX7Py6amJn3rW98K+bgAAPQGnU6xtoHHy/OttXusta9aa18hHAMAequpuenKTkvUzsHnhjYcS1LKYGnQGGnTO6Ed9zg33HBDt3byBgAgGnVnDXKRMeYcxysBAMBhLpfRLVM8WrK5Wht2Hgx9Ad4CaVuRdLgm9GMDAIAudScgT5O02BhTYoxZbYz51Biz2unCAABwwo2TsxTndml20dbQD+7zSy2N0pb2NwkDAADh1Z2AfIUkr6QCSV+UdFXrKwAAvc6A5HhdOXawXlq+XbWHm0I7eM65UmxSYB0yAADocboMyNbarZL6KRCKvyipX+s1AAB6pcJ8jw4ebtLCVRWhHTgmXhp2QdjXIQMAgPZ1GZCNMd+QNFtSRuvXLGPMfU4XBgCAU8729NfIwSl6evHWDs8hdozXL1WXSHu3hHZcAADQpe5Msf6KpDxr7YPW2gcVOO7pTmfLAgDAOcYYFeZ7tLbygFaU7Qvt4D5/4JWnyAAA9DjdCchGUvNRPze3XgMAoNf60sSh6hPn1qxQb9Y1wCf1zZZKFoV2XAAA0KXuBOS/SlpijHnIGPOQpCJJf3G0KgAAHJYcH6NrJg3VK6srtbe2IXQDGxM47mnz+1JzY+jGBQAAXerOJl2/lvRvkqol7ZX0b9ba3zpdGAAATivM96ihqUXPLysL7cA+v3T4gFReHNpxAQBAp7qzSVe+pI3W2v+y1v4/SZuMMXnOlwYAgLNGDk7VOcP6a/aSbWppCeFmXcOnSsbNcU8AAPQw3Zli/XtJNUf9XNt6DQCAXq8w36Ote+r04abdoRs0sZ+UNZmNugAA6GG6tUmXPeoMDGtti6QY50oCACB0Lh8zWAP6xIV+sy6vX6pYIdXuCe24AACgQ90JyKXGmK8bY2Jbv74hqdTpwgAACIX4GLduPCdbb3++UxX7DoVuYJ9fkpVK3w3dmAAAoFPdCch3SzpP0vbWrzxJX3WyKAAAQumWKTmykp5bui10g2ZOlBL7c9wTAAA9SHd2sd5lrb3ZWpvR+nWLtXZXKIoDACAUstOSdPEZ6Xr2kzI1NreEZlCXWxpxcSAg2xBuEAYAADrUYUA2xtxpjMlt/d4YY540xuw3xqw2xkwKXYkAADivMN+jqoOH9eZnO0M3qNcvHayUdq0N3ZgAAKBDnT1B/oakLa3fz5A0XtIISf8p6f85WxYAAKF18ZkZGtovMbSbdXkLAq/sZg0AQI/QWUBustY2tn5/laS/WWv3WGvfltTH+dIAAAgdt8volrwcLS7do027arq+IRj6DpXSR3EeMgAAPURnAbnFGDPEGJMgyS/p7aPeS3S2LAAAQu+mc7IV6zaavSTET5G3LpYa6kI3JgAAaFdnAflBScUKTLNeaK39TJKMMVPFMU8AgAg0MDlel48ZoheWlauuoSk0g/oKpObD0taPQzMeAADoUIcB2Vr7iiSPpFHW2juPeqtY0k1OFwYAQDjclu/RwfomvbyqIjQDes6XYhKYZg0AQA/Q6TFP1toma+3e467VWmtDtDgLAIDQOmdYf50xKFmzikJ0JnJsouQ5j426AADoAbo8BxkAgGhijFFhvkefbt+vVWX7QjOo1y/tXi/tLw/NeAAAoF0EZAAAjnPNxKFKinPr6VAd+eTzB155igwAQFh1GJCNMZcZY65v5/qtxphLnC0LAIDwSUmI1ZcmDtXLqyq0r67B+QHTR0opmaxDBgAgzDp7gvwjSe+1c/0dST92phwAAHqGwjyPDje16IVlIZj2bExgN+vSf0jNIdo9GwAAnKCzgJxkra06/qK1doekPs6VBABA+I3OTNWknH6avWSbWlqs8wN6/VL9fqliufNjAQCAdnUWkBOMMTHHXzTGxEpKdK4kAAB6htvO9Wjz7lp9XLLH+cFGXCwZF+uQAQAIo84C8kuSnjDGtD0tbv3+D63vAQAQ0a4YM0T9k2I1KxSbdSWlSZmTWIcMAEAYdRaQH5C0U9JWY8wyY8xySVskVbW+BwBAREuIdevGydl66/Od2rG/3vkBfX5p+zLp0F7nxwIAACfoMCBba5ustTMlZUv6sqTbJeVYa2daaxtDVB8AAGF1S16Omlusnvtkm/ODef2SbQls1gUAAEKus2OerjXGXCvpCkm5knySJhtjUkJVHAAA4eYZ0EdTz0jXs0u3qbG5xdnBhp4txfdlHTIAAGHS2RTrLx73NV3StyWtNsYUhKA2AAB6hMJ8j3YeOKx3Pt/p7EDuGGnERVLJu5INwc7ZAADgGCfsUn2Etfbf2rtujPFImispz6miAADoSQpGZiizb4JmFW3T5WOGODuY1y99/rK0e4OUfqazYwEAgGN09gS5XdbarZJiHagFAIAeye0ymjElRx9u2q3SqhpnB/P5A69MswYAIOROOiAbY86UdNiBWgAA6LFumpKtGJfR7CUOb9bVL0cakMtxTwAAhEGHU6yNMS9LOn4BVJqkIZJuc7IoAAB6moyUBF02ZrBeWFau+y87UwmxbucG8/mlZU9JjfVSbIJz4wAAgGN0GJAlPXrcz1bSHkkbrbUNzpUEAEDPVJjn0aurK/XyqgrdMDnbuYG8fmnJH6RtH0te9sUEACBUOjsH+b3jvt631n4m6RxjzOMhrBEAgB4hf0SafBnJmlW01dmBhp0vueNYhwwAQIh1aw2yMWaCMeaXxpgtkn4qaZ2jVQEA0AMZY1SYl6NV5fv1afl+5waK6yPlnCuVLHJuDAAAcIIOA7Ix5gxjzIPGmM8l/bekMknGWjvNWvu7kFUIAEAPcu3ZWUqMdTv/FNnnl3atlQ5UODsOAABo09kT5HWS/JK+aK29oDUUN4emLAAAeqbUhFhdPSFTC1Zt1/5Djc4N5G097omnyAAAhExnAfk6STskvWuMecIY45dkQlMWAAA9V2G+R/WNLXpxWblzgww6S0oexDpkAABCqLNNuuZZa2+SNFLSPyR9U9IgY8zvjTGXhqg+AAB6nDFD+2pCdj/NXrJV1h5/ImKQGBPYwbr0XamFCVwAAIRCl5t0WWtrrbWzrbVXScqStFLSTMcrAwCgByvM96ikqlaLS/c4N4jXLx3aK1WsdG4MAADQplu7WB9hra221v7RWsuhjACAqHbVuCHqmxir2UXbnBvEO02SYR0yAAAhclIB+WQZYy43xqw3xmwyxpzw1NkY8xtjzMrWrw3GmH1O1gMAQLAkxLp14+QsvfHZDu06UO/MIH0GSkPGSyWsQwYAIBQcC8jGGLekxyVdIWm0pBnGmNFHt7HWftNaO8FaO0HS7yS95FQ9AAAE2y15HjW1WD33SZlzg/j8UtlSqd7Bc5cBAIAkZ58gT5G0yVpbaq1tkPScpKs7aT9D0rMO1gMAQFANH9hHF+YO1LNLt6mpucWZQbx+yTZLm993pn8AANDGyYA8VNLRf6Ve3nrtBMYYj6ThklhkBQDoVW7N86hyf70WrdvlzADZU6S4FI57AgAgBJwMyO2dmdzRWRg3S3rBWtvuORbGmK8aY4qNMcVVVVVBKxAAgNP1hVEZGpyaoKeLtjozgDtWGn5RYB2yU0dKAQAASc4G5HJJ2Uf9nCWpooO2N6uT6dXW2j9Zaydbayenp6cHsUQAAE5PjNulGVNy9MHG3dqyu9aZQXwF0r5t0p4SZ/oHAACSnA3In0jKNcYMN8bEKRCCFx7fyBhzpqT+khY7WAsAAI65eUq23C6jZ5Y6dOST1x94ZTdrAAAc5VhAttY2SbpX0huSPpc011r7mTHmx8aY6Uc1nSHpOWuZNwYA6J0GpSbo0tGDNLe4TPWN7a4WOj1pw6W0EaxDBgDAYTFOdm6tfU3Sa8dde/C4nx9ysgYAAELhtnyP/r5mh15dXanrzs4K/gBev7RyttR0WIqJD37/AADA0SnWAABEjXO9AzQivY9mLXFosy6fX2qsk7YVOdM/AAAgIAMAEAzGGN2a59GKbfu0Zvv+4A8w7ALJFcM6ZAAAHERABgAgSK6flKWEWJdmO/EUOT5Fys6XShYFv28AACCJgAwAQND0TYrV9PGZmr+iQgfqG4M/gK9A2vGpVLMr+H0DAAACMgAAwVSY79GhxmbNW749+J23HffEU2QAAJxAQAYAIIjGZfXTuKy+mlW0VUE/wXDwOClpIMc9AQDgEAIyAABBVpjv0cZdNVqyuTq4Hbtckrcg8AS5pSW4fQMAAAIyAADB9sVxmUpNiNGsIgc26/L5pbrd0o7Vwe8bAIAoR0AGACDIEuPcuv7sbL3x2Q5VHTwc3M69BYFXjnsCACDoCMgAADjg1vwcNTZbzS0uC27HyRnS4LHSJjbqAgAg2AjIAAA4wJuerPN9A/TMkm1qbgnyZl1ev1RWJB0+GNx+AQCIcgRkAAAcUpjn0fZ9h/TuuiCfW+zzSy1N0uYPgtsvAABRjoAMAIBDvjB6kDJS4jVrSZA368rOl2L7sA4ZAIAgIyADAOCQWLdLN0/J0XsbqrRtT13wOo6Jk4ZfyHnIAAAEGQEZAAAHzZiSLZcxmr00yE+RvQXS3s1SdWlw+wUAIIoRkAEAcNCQvon6wqgMPV9crsNNzcHr2OsPvJawmzUAAMFCQAYAwGGF+R5V1zbo75/uCF6nA7xSvxyOewIAIIgIyAAAOOx870ANH9hHTxcFcZq1MYGnyJvfl5obg9cvAABRjIAMAIDDXC6jW/NytGzrXn1eeSB4Hfv8UsNBqWxp8PoEACCKEZABAAiB68/OUnyMS7OC+RR5+EWScXPcEwAAQUJABgAgBPolxemqcZmav2K7DtYHaUp0Ql8pewrHPQEAECQEZAAAQuS2cz2qfo/bPgAAIABJREFUbWjW/BXbg9ep1y9VrpJqdwevTwAAohQBGQCAEBmf1VdjhqZqVtE2WWuD06mvQJKVSt4NTn8AAEQxAjIAACFijFFhnkfrdx5U8da9wel0yAQpMY11yAAABAEBGQCAEJo+IVMpCTHB26zL5Za806SSRVKwnkoDABClCMgAAIRQUlyMrpuUpdc+rdTumsPB6dTrl2p2SjvXBKc/AACiFAEZAIAQK8zPUWOz1dzisuB06J0WeGU3awAATgsBGQCAEPNlpCh/RJqeWbJNzS1BmBadmilljGYdMgAAp4mADABAGBTme1S+95De31AVnA69BdK2IqmhNjj9AQAQhQjIAACEwaWjBys9JV5PB2uzLp9fam6QtnwUnP4AAIhCBGQAAMIgLsalm8/J1rvrd6msuu70O8w5T4pJZJo1AACngYAMAECYzJiSIyPp2aXbTr+z2ARp2Pls1AUAwGkgIAMAECaZ/RLlHzVIc4vLdLip+fQ79PqlPRulfUEI3AAARCECMgAAYVSY79Humga9vmbH6Xfm8wdeeYoMAMApISADABBGF/oGyjMgSbOLgvDUd+AZUmoW65ABADhFBGQAAMLI5TK6ZUqOlm6p1vodB0+vM2MkX4FU+r7U3BScAgEAiCIEZAAAwuyGydmKi3FpVjCOfPL6pcP7pe3Fp98XAABRhoAMAECYpfWJ01Vjh2jeiu2qPXyaT35HTJWMi3XIAACcAgIyAAA9wK35HtUcbtL8ldtPr6PE/tLQyaxDBgDgFBCQAQDoASbl9NOoIal6evFWWWtPrzNvgbR9uVRXHZziAACIEgRkAAB6AGOMbsv3aN2Og1q+be/pdebzS7JS6T+CURoAAFGDgAwAQA9x9YRMJcfHaNbpHvmUOUlK6Ms0awAAThIBGQCAHqJPfIyunTRUr66uVHVtw6l35I6RRlwsbVokne50bQAAoggBGQCAHqQw36OG5hY9X1x2eh15/dLBCqlqXXAKAwAgChCQAQDoQc4YlKIpw9M0e8k2tbScxtNfnz/wynFPAAB0GwEZAIAepjDfo23VdXp/Y9Wpd9I3Sxp4JuuQAQA4CQRkAAB6mMvPGqyByXGnv1mXzy9t/VhqPBScwgAAiHAEZAAAepi4GJdunJytRet2avu+0wi3Xr/UVC9t/Sh4xQEAEMEIyAAA9EC35OXISnp2yWk8RfacJ7njA7tZAwCALhGQAQDogbL6J6ngzAw990mZGppaTq2TuKRASGYdMgAA3UJABgCghyrM92h3zWG9uXbHqXfi8weOetpfHrzCAACIUARkAAB6qIvOSFd2WqJmFW099U68BYHXEqZZAwDQFQIyAAA9lNtldMsUj4pKq7Vx58FT6yRjtJQyhPOQAQDoBgIyAAA92I2TsxTndmn2qW7WZUzgKXLpP6SW5qDWBgBApCEgAwDQgw1IjtcVYwfrxWXlqmtoOrVOvAVS/T6pYkVwiwMAIMIQkAEA6OFuy/fo4OEmLVhZcWodeAskGaZZAwDQBQIyAAA93Nme/ho5OEWzirbKWnvyHSSlSZkTOe4JAIAuEJABAOjhjDG6Nd+jzyoOaGXZvlPrxOeXyoulQ6d4PwAAUYCADABAL3DNxKHqE+fWrKJT3KzL65dss7T5veAWBgBABCEgAwDQCyTHx+iaSUP18uoK7a1tOPkOsiZL8amsQwYAoBMEZAAAeonCfI8amlr0wrLyk7/ZHSsNv0gqWSSdyjpmAACiAAEZAIBeYuTgVE329NfsJVvV0nIKIdfnl/aXSbs3Br84AAAiAAEZAIBepDDfoy176vRRye6Tv9nrD7yymzUAAO0iIAMA0ItcMXaw0vrE6enFW0/+5v4eaYCPdcgAAHSAgAwAQC8SH+PWjZOz9fbnO1W5/9DJd+AtkLZ8KDXWB784AAB6OQIyAAC9zK15ObKSnl1advI3e/1S0yFp2+Kg1wUAQG9HQAYAoJfJTkvSxWek67ml29TY3HJyNw+7QHLFBnazBgAAxyAgAwDQCxXme7Tr4GG9tXbnyd0Ynyzl5BOQAQBoh6MB2RhzuTFmvTFmkzFmZgdtbjTGrDXGfGaMecbJegAAiBQXn5mhof0SNavoFDbr8vmlnWukgzuCXxgAAL2YYwHZGOOW9LikKySNljTDGDP6uDa5kr4r6Xxr7VmS/sOpegAAiCRul9EteTn6uGSPSqpqTu7mtuOeeIoMAMDRnHyCPEXSJmttqbW2QdJzkq4+rs2dkh631u6VJGvtLgfrAQAgotw4OVuxbqPZRdtO7sZBY6Q+GRz3BADAcZwMyEMlHb29ZnnrtaOdIekMY8xHxpgiY8zl7XVkjPmqMabYGFNcVVXlULkAAPQu6SnxunzMEL2wrEyHGpq7f6PLFTjuqfRdqeUkN/kCACCCORmQTTvX7HE/x0jKlXSxpBmS/myM6XfCTdb+yVo72Vo7OT09PeiFAgDQWxXm5ehAfZNeXlVxcjf6/FLdHqlypTOFAQDQCzkZkMslZR/1c5ak4//vXS5pgbW20Vq7WdJ6BQIzAADohinD03TGoGTNWnKSm3WNmBZ4LWGaNQAARzgZkD+RlGuMGW6MiZN0s6SFx7WZL2maJBljBiow5brUwZoAAIgoxhgV5nu0uny/VpXt6/6NyenSkPHSJjbqAgDgCMcCsrW2SdK9kt6Q9Lmkudbaz4wxPzbGTG9t9oakPcaYtZLelXS/tXaPUzUBABCJrpk4VElx7pM/8snrl8qXSvUHnCkMAIBextFzkK21r1lrz7DWeq21P2u99qC1dmHr99Za+5/W2tHW2rHW2uecrAcAgEiUkhCrqycM1curK7S/rrH7N3oLpJYmafP7zhUHAEAv4mhABgAAoVGYn6P6xha9sLy8+zdl50lxyaxDBgCgFQEZAIAIcFZmX03K6afZRVtl7fGHRnQgJk4adqFUwjpkAAAkAjIAABGjMN+j0t21+rjkJLbz8PmlvVukPSWO1QUAQG9BQAYAIEJcOXaI+ifFntxmXd6CwCtPkQEAICADABApEmLdunFytt5cu1M7D9R376YBXqn/MGkT65ABACAgAwAQQW7Jy1Fzi9WzS7d1/yavX9rygdTU4FxhAAD0AgRkAAAiiGdAH110RrqeW1qmpuaW7t3k80sNNVLZEmeLAwCghyMgAwAQYQrzcrTjQL3e/nxX924YdqHkiuG4JwBA1CMgAwAQYQpGZiizb4JmL+nmZl0JqYEzkVmHDACIcgRkAAAiTIzbpRlTcvTBxt3avLu2ezd5C6Qdq6Wabj51BgAgAhGQAQCIQDdNyVaMy2h2d4988vkDryXvOlcUAAA9HAEZAIAIlJGSoMvOGqznl5WrvrG56xsGj5eSBrAOGQAQ1QjIAABEqMJ8j/YfatQrqyu7buxySSOmSSWLpJZu7n4NAECEISADABCh8kekyZeRrKdPZpp1bZW081NnCwMAoIciIAMAEKGMMbo1L0eryvZpzfb9Xd/gLQi8lixytjAAAHooAjIAABHs2klZSox1a1Z3niKnDJYGjeG4JwBA1CIgAwAQwfomxurqCZmav3K79h9q7PoGb4G0rUg6XON8cQAA9DAEZAAAIlxhvkf1jS16aXl51419fqmlUdryofOFAQDQwxCQAQCIcGOG9tX47H6avWSbrLWdN845V4pN4rgnAEBUIiADABAFCvNytGlXjYpKqztvGBMvDbuAdcgAgKhEQAYAIAp8cXym+ibGdm+zLq9fqi6R9m5xvC4AAHoSAjIAAFEgIdatG87O0huf7dCuA/WdN/b5A688RQYARBkCMgAAUeLWfI+aWqzmfFLWecMBPqlvDuchAwCiDgEZAIAoMXxgH12YO1DPLt2mpuaWjhsaI/kKpNL3pOZuHA0FAECEICADABBFbs3zqGJ/vRat29V5Q69fajgolX8SmsIAAOgBCMgAAESRL4zK0ODUBM1asq3zhsMvkoybdcgAgKhCQAYAIIrEuF26eUq23t9Qpa17ajtumNhPyprMecgAgKhCQAYAIMrMmJIjt8voma6eInv9UsVKqXZPaAoDACDMCMgAAESZQakJunT0IM0tLlN9Y3PHDX1+SVYqfTdktQEAEE4EZAAAolBhvkd76xr12qeVHTfKnCgl9ue4JwBA1CAgAwAQhc7zDtCIgX00q2hrx41cbmnExYGAbG2oSgMAIGwIyAAARCFjjG7N92j5tn36rGJ/xw29fulgpbRrbeiKAwAgTAjIAABEqesnZSkh1qVZRZ1s1uUtCLxy3BMAIAoQkAEAiFJ9k2L1xXGZWrByuw7WN3bQaKiUPorjngAAUYGADABAFCvM96iuoVnzVmzvuJHPL21dLDXUha4wAADCgIAMAEAUG5/dT+Oy+urpxVtlO9qIy1sgNR+Wtn4U2uIAAAgxAjIAAFGuMM+jjbtqtHRzdfsNPOdJMQmsQwYARDwCMgAAUe6L4zOVmhCjWUs62KwrNjEQklmHDACIcARkAACiXGKcW9efna3X11Sq6uDh9ht5/dLuDdK+stAWBwBACBGQAQCAbs3PUWOz1dziDgKwzx94LVkUuqIAAAgxAjIAAJA3PVnneQfomSXb1NzSzmZd6SOllEymWQMAIhoBGQAASAoc+bR93yH9Y/2uE980RvIVSKX/kJqbQl4bAAChQEAGAACSpEtGD1JGSrxmFW1tv4HXL9XvlyqWh7YwAABChIAMAAAkSbFul26ekqN/bKhSWXXdiQ1GXCwZF8c9AQAiFgEZAAC0mTElWy5jNLu9I5+S0qTMSaxDBgBELAIyAABoM6RvovwjMzS3uEyHm5pPbODzS9uXSYf2hr44AAAcRkAGAADHuO1cj6prG/T6mh0nvun1S7YlsFkXAAARhoAMAACOcb53oIYNSNLTi9vZrGvo2VJ8X9YhAwAiEgEZAAAcw+UyujXPo+Kte7Vux4Fj33THSCOmSiWLJNvOeckAAPRiBGQAAHCC68/OUlyMq/0jn3x+6cB2qWp96AsDAMBBBGQAAHCC/n3i9MVxmZq3fLtqDjcd+6a3IPDKbtYAgAhDQAYAAO0qzM9RbUOz5q3Yfuwb/XKkAbmsQwYARBwCMgAAaNeE7H46KzNVs4u2yh6/3tjnl7Z+JDXWh6c4AAAcQEAGAADtMsbotnyP1u04qGVbjzv32OuXmuqlbR+HpzgAABxAQAYAAB2aPiFTKQkxevr4zbqGnS+545hmDQCIKARkAADQoaS4GF03KUt//3SH9tQc/ucbcX2knHMDxz0BABAhCMgAAKBTt+blqKG5RXOLy499w+eXdq2VDlSEpzAAAIKMgAwAADqVOyhF+SPS9MzSrWpuOWqzLq8/8MpTZABAhCAgAwCALhXme1RWfUjvb6j658VBZ0nJg1mHDACIGARkAADQpUtHD9bA5HjNOnqzLmMkb4FU+q7U0hy+4gAACBICMgAA6FJcjEs3n5OtRet3qXxv3T/f8PmlQ3ulipXhKw4AgCAhIAMAgG6ZkZcjI+nZpdv+eXHENElGKmGaNQCg9yMgAwCAbhnaL1EFIwdpzidlamhqCVzsM0AaMp51yACAiEBABgAA3VaYn6PdNQ16/bMd/7zo80vln0j1+8NXGAAAQUBABgAA3XZRbrpy0pKO3azL65dss1T6XvgKAwAgCAjIAACg21wuo1vzcrR0c7U27DwYuJg9RYpL4TzkMKmurla/fv3U0NCgOXPmKCUlRQkJCUpMTNQ3vvGNdu+55ZZb2tqkpqZq4cKFbe9ddtllio2NVVxcnH72s59Jkvbt26fk5GQlJiYqISFBU6dO7bKuAwcOKCcnR7GxsUpOTtaHH37YYduGhgYlJSVp0KBBbdcee+wxJSUlKSEhQV6vV/X19ZKkzZs3KzMzU4mJiUpOTta8efMkSTU1Nerbt29bu2A7+nM+ory8XG63W+PGjevwvuuvv15xcXFKSEjQlClT2q639zlLUkxMTNs/mz59+nRZV0tLi8aPH6/Y2FglJiZq9uzZnbYfPHiwEhIS2n6eO3du278zgwYNUnl5uaTA55mbm9tWy29/+9u2e9LS0rR58+YuazsVx3/Oixcv1oABAxQfH6/4+Ph2/z366KOPlJaWpsTERPXr10+ffPJJ23t33nmn4uLiFBcXpzvvvLPt+n333df2axs4cKDWr1/faV2R/jnn5eUpISFBCQkJHf6+cfXVVys+Pl6JiYlKS0vTRx991PbewIEDZYw55r/hzpzM7w/GmFhjTJ0xZudR1/6z9Vq9MWaTMSa+9fowY0yFMeaQMabGGPOl1ut9jDH7jrTrlLW2V32dffbZFgAAhM+emsM29/uv2R/M//SfF5+ZYe1vxljb0hK+wqLUDTfcYK+55hprrbWvv/66ffPNN6211i5btsy6XC67ZcuWE+4pKytr+/573/ueHTBggLXW2gULFtiEhAS7f/9++95779mYmBh7+PBh29zcbCsrK6211tbW1to+ffrYJ554otO6brrpJjtq1ChrrbX33Xefzc7O7rDt9OnTrcfjsRkZGdZaaxsbG63b7bavv/66tdbaCy+80N5+++3WWmsnT55sL774Ymutta+++qrt379/Wz8XX3yxveeeezqt61Qd/TkfMX78eOvxeOzYsWPbveexxx6z/fv3t/v377fWWrtmzRprbcefs7XWut1uu27dum7X9dBDD9mBAwfa5uZm+8QTT9g+ffp02Pb++++3Ho/HxsfHt11LSkqyv/3tb6211n75y1+2F1xwQduv1+fztdWdmJhoGxsbrbXW3nHHHfYLX/hCt2s8Gcd/zn379rUPP/ywtdbayspKW1VVdcI9Q4cOtXfccYe11tpHH33UDh8+3Fpr7aZNm2xMTIzdtGmTLS0ttTExMba0tNQeOnTIGmPaPudzzjnHTp06tdO6Ivlz/uEPf2jT0tLsoUOH7M6dO21SUtIxv0cc8dhjj7V9/jfffPMx/03/8pe/tA888EDbf8Nd6er3B0nFtjX/SVogaYukna0/uyU1Sbq09ef3JP219fulkha1fn+FpOqj+lkk6XHbRd4Me+A92S8CMgAA4fcfz62wZz34uq2pD/xBzi59wtofplpbtTG8hUWh5ORk+8EHH7T7XkJCQltg7si9995rBw4caK219tJLL7WXXnpp23sDBgywf/zjH49pX1VVZRMTE+2f//znTvs9+t4jgaS5ufmEdkuXLrX9+/e3jz76aNsfrteuXWtjYmLa2vzud7+z6enp1lpr09PT7eOPP972XkxMjP3008Bf1syZM6etXbAd/zk//fTTNjs7237lK1/pMCBnZWXZX/ziFydc7+xzPtmAPHLkSHvvvfe2/RwbG2tXrFhxQrvKykqbkpJiFyxYcExwk9T2z6WoqMjGxcVZa60dM2aMvfvuu9va9e/f3z755JPWWmtLS0uP6SOYjv6cFyxYYFNSUrq8Jz4+3i5dutRaa21zc7MNPAMM/Ls9cuTItnZHPqva2lprjLEffvihbW5utqNGjbK33nprp2NE8ud85ZVXWr/f3/Zebm6u/Y//+I9O73/mmWdO+Gfzm9/8ptsBuavfH44EZEmTJVVL+s+jAvJISQ32n8H3a5J2tX6/S9I9R73XKOms1u9vONKusy9Hp1gbYy43xqxvfew9s533v2yMqTLGrGz9usPJegAAQHAU5ueo5nCTFqysCFzw+gOvHPcUUjU1Naqrq9MFF1xwwnt//etf1dLSomnTprV774033qjY2Fj9/ve/1zPPPCNJqqys1IgRI9rapKWltU09bWhoUGJiotLT0zV27Fh95Stf6bK2iRMnSpISEhLkdru1cePGE9pdc801+q//+i+53e62a2eeeaastfrb3/4mSfrzn/+sAwcOSJJ8Pp/+8pe/tP0am5qatHJl4BzuL33pS9q9e3endZ2K4z/npqYm3XvvvXr++ec7vW/37t169dVXlZycrH79+umpp56S1PnnbIzRxIkTlZSUpMLCwi5rq66u1qhRo9p+Tk5O1urVq09od+WVV+qee+5RWlraMdeTk5P1wAMPSJJ+/vOft025HT9+vF5++WXV19fr/fff1969e7V27VpJ0vDhw9XS0tLuP8/TcfznvHjxYiUmJiozM1NJSUk655xzjpnifsTgwYP18MMPS5JmzgxEjo0bN2rz5s0aOnRoW7vMzExt3rxZSUlJ+uY3v6kLLrhAsbGx2r59u5588slOa4vkz/mCCy7Q0qVLtXv3bq1fv16bN2/Wpk2bOu3j4YcfPmbJwKnU0J3fHyTNl3SfpJajrq2X5DLG3Nb68x2SUlu/39j6s4wxX5YUI2nCUX0N7Ko2xwKyMcYt6XEFHm2PljTDGDO6naZzrLUTWr/+7FQ9AAAgeCbl9NeoIamaVbQ18Lf0acOltBHHHPd06NAhTZ06Vc3NzZKkp556Srm5ucrNzW0LCserrq7WJZdcotzcXF1yySXau3dvl7V0p98jHn30URlj2kLU3r17dc0112jcuHGaMmWK1qxZc0z75uZmTZw4UVdddVXbtZtvvjnof2A9VRs2bFBMTMwJ11euXKm77rpLjz/+eLvvS4E1kY2Njbrrrrt09913S9KRJy7HcLkCf1yMi4vToUOHtGXLFm3YsKFt7e/JONLXEQ8++KD69et3QhB0uVz6n//5H913331KTk5WcnJy271z587VgQMHlJiYqJ/85CdKSkpSXFxcW43GGFVUVJx0bZ05/nOeMWOGzjvvPOXl5XV6n7VW+/bt04EDB/Sb3/xGd9xxh1paWjr9nJcsWaK6ujp98sknevHFF/W73/2uyzE66uuIOXPmaPv27frFL35xQts5c+boj3/8o5KSknTgwAEZYyQF/lIiIyND/fv310033aT09HTFxsa23ZeQkNBuQDwdx3/ODQ0N2rVrl5555hlVV1eroqJCX/3qV0+476WXXlJRUZGSkpL05ptvyuVyKT4+vt3Pxhijuro6/fWvf9U777yjxsZGeTweXXnllZ3WFsmf83e/+13l5eUpOztb5513noYOHXpMDce75557VFpaqpdeeimodR3/eRpjfiRpn7X2mAXfrU+Y75H038aYGkk1+meAvlFSqjHmkKQHJdVJami9r1GSNcYM6ayO9n/HDI4pkjZZa0slyRjznKSrJa11cEwAABACxhgV5ufo+/PWaPm2fTrb0z/wFHnlbKnpsBQTryeffFLXXnut3G63qqur9aMf/UjFxcUyxujss8/W9OnT1b9//2P6feSRR+T3+zVz5kw98sgjeuSRR9r9w+YR3e1XksrKyvTWW28pJyen7drPf/5zTZgwQfPmzdO6dev0ta99Tf+/vfsOr6rK+jj+XSmEhCAkRBCEEEDEgtICiA4aYLAyYEUsIKgjKo7oIA6IOPNioeiAWLAgSFFAEBkZC6JoREchhKJS1YCUEIIhoaSXu98/7iUkIUFEwo3y+zxPnpy6zzobTu5Zd++zz5Ilh5L8iRMncvbZZxe3XoL35nDcuHFMnjz5t1Th0SkqhIIsyM+G/CwoyCIvO5O09D1k7NtLyrYUigoLWfOht/XLAak/7+Hqv46g91VxtGtgrP5giq8w7w1+2fv8/pe2ZNKkSax6fzLhwR7WLF/Kqvcng3Ps3rWTkJwUVr33askSaNaoLuOfeozooNQSJZUuuHq1IP4z/VkCUv5Ebn4+RUWF7F23mJUbPvHF4WHB3Jls+GErQYGBeHzdGxvUq8PCV/6PtnXhk+newaten/8RP2wMIfE/z+OAWU/fD3gHTrrohr8Ttm8jKxY8V1zuj5/PIjmkghv8EhVQMmIrPrvDJafuoaiwkIT5EwD43+dLSN2z1xe3B+eg5ZmNmTr6gTJ1EEjXNk1JXPAc557ijW3e8yOoEVjA6q8/I2G+d0Cm3SnJBB/YVjyfsGUpAGc3PZ2501/mgtM9VKRGNSP+v7Npf1oBAAf278dSv2HFO3twviRsygvTSfv5Z4ICA3HO4XGOmjVCWfLGWKKAD1/7JwDLVm9g2ddfkbDgOcB4eeSA4uN0uXkokaSRsMCbsOfn5ZLy3WckBBz+ZcTB5O/XSk5No6iwkBX/edFbf7kphIdVp8bedXy3aB0dWzbhy88WF68v6d1JIwBIy9hPjzu/JXXVf6lBJus3bSrefvP362h/fgvGDhtIYX4uNfdvYOXCDVzaoQVT5i0qt9yDaoQEEP/+W7Rv4P0fcmD/fmz32lL7THnh9UP1jMPj8dbzp28+w6nAoimjAPh61XqWff1V8b6vPHaoN0bcTUOow57idfl5uexa+zkrAku8Wu83Sk5NK/7i8qCPP/64eDomJoZWrVqVu++4ceOYOnUqq1at4pRTTil3m6MRHh7O6tWrad++Pbm5uRQVFdGsWbOym3UHzjKzQsDwthpvds41dc5NBiYD+HoqxwA455KB5r7lhreL9VclyjRg75Fis/K+DTkezOx64HLn3MEm7r5AR+fcfSW26Q+MBn4GvgcedM5tL6esu4C7AKKjo9tt3bq17CYiIiJygmXlFdLxqSVcek49xt/YGjZ9CLP7QL+F0PQSLrzwQmbNmkVMTAyzZ88mPj6eV155BYCBAwcSFxfHTTfdVKrMFi1aEB8fT/369UlJSSEuLu6Io8sebbngHU145MiR9OrVi8TERKKiorjqqqsYPnx4cVfDZs2a8dVXXxWPMnvbbbcxYsQIxo8fz3vvvQd4k7JmzZrxww8/eFthnIOCHCjwJbH5Wb7pTG9iW2rat77stC/5JT8bl5+JJz8by88iwHN4d9KygkbtJ+3hcGpXDyAz30P0hEwuaRzIgj4Vj4D8cVIB3Zt5E8iRn+Yyflk+WY+cwrsbC+gzP4fUIeGs2eWh28xssoaHk5ThCAs2GtcOID3bQ5PnMhncsRqjulSn02uZXNw4iLHdq5c6Ru952azd7WH9oHDu/zCHBRsL2f5gzQpjenZZHqO/yCd1qHebdbsLObeRVyk+AAAX4klEQVRuEPtzPcRMzGRE5xCGXBjC1r0e6oRBeLUA+r2Tzf+2F5E02LvPD3uKaPlSFnmPHvuNe0VK1nNJdy7MJiHZw7f3hB+2z83zs9mx38PSAeF89GMBV83KIf/RcP77fVG59bw3Fwo90OCUAHZnemj2fCb/uCiERy8O4YZ52QDMuyGs1DH+9VkuLybmkzoknKlrCnhgUR6Zj1R8/l9uK+TPM7LJ9dXRwXou9Hho8XwWF8cE8nqvMNKyPXg8UDc8gDFf5jH6yzz2DfPu4/F4CH4ik6xHwqkedHw7pJas5/xCD7XGZrLqrhqcfWogzZ8/QJvTAplbpg42pRXRLNIICgjgoimZBAbA0gHhJKV7OOvFTL6/z/tvc+YLmWwcFM6+PA/tJ2ez9h5vuZ2nZpJTCIl3hZ889fxEFmlpacUjWW/dupXmzZvz9ttvc+utt7J3795So3CD9+9tv379+OCDD+jevfthZT777LOMHj2a1NRDX5x16tSJiy+++LAvOnv37s3atWtZv349999/PwsWLGD79kNpoJmtdM7Flph/ABjunKvnmz/HObfezGoCW4EnnHPjzawxkOacyzKzGcBFzrlmvn3OANY55444knVltiCX99VR2Wz8v8Bs51yemd0NTAe6HraTc68CrwLExsZWTkYvIiIiv0qNkCCubXs6c1Zs59Ee5xAZ0xkCgiFpCfkNO7F582ZiYmIASE5OplGjRsX7NmzYkOTk5MPKTE1NpX59b++3+vXrs3v37iPGcLTlLly4kNMb1KfVmY2972zesxnyd9Aqpg7vTJnAnyLSSFj9HVu3/sSO98ZRr2kdHnhsBuOubcOBL8fDztUw42rIzyKgIJszqqXxzZAmtKvn8Sa35bY7ls8FBOGCa1AYGEpeQCjZLoRMF8L+ompkFESSXlCPLBdCDtXJciHkWXWqhdYktOYphIefQu1atYiIiCQqojYhISE0mnYdowt6cle/2xg5ajQZue/w4U9BVB+TB8C4Jx/jL1dezo1976R9u7Y89MAgBl/Th6S3tmABAYRUC2H8hLH81C2OVkC7bXdT5xlvi/zfBt1DSr+BxC/6mL8//Kh3EBsc7dq0547Jr7EV+OH5LvS6YhDbbrwe57v9M2DkX/Zz6VXXEvxkOsHBwUyb/ArbL+jAd2vX0/f2gaxJ+AIrcbu4t2AGRYmTSe73BWDcfHVvNmzcBDi6xnWjz8svkGzGW7PfYsQ/H8fMiIyIZMHCN9np+z/wzBNjaHH2Cnb2n4+3EbP07Whxy2aFLZxW4erGM/7MvwN6c9+dpbv4Zm8cTkH2N6Te+QEAF13SncF/u5fe11/DqBsP0PWyHoSM/pmAgABGPDKM3X+9g45A++TbqPPv5ZgF8OADg0kfeB9ffb2Mm/t5h+RxztGxw4XcOW0au4C1c3rQ6YIO7Lr9sVLHv6u/h7cu+wshozcTGBDIuLFj2XXdNTjnoUmLVmzZ9E2p7dMSEnFv3k7KgAQAhg68j08mxwNwXsvzeeKd2ey0AJavSOTGW/oDRlhYGDPeeJ2dsW0BmDf/P5xadzzpdy49vKJ+Q+Obc47o6VfwtLuOe3318GD+ZNpMfBFHAfVPa8jjs98hObwG1/bpS8f27fjHkAf491NjeX20txdudKNo3lswl+Sa4VQHrt35CGdO8n65dd2111D9gSepDvRMeYhWkz8mwIyaNSNY8NYbJDdpzHdzetGpQ3uS+z1aKrY7bvUw56prCBm9mYCAQMY9NZrka3sB0OycNiStX11q+59XrMS9eQfJ/ZYB8NA995eo5/MYNW8WyQEBJKxYyY19B/jqOZTpM6eQ3M5bz2+/8y6nnjqBPbfHH3OdVqTprF68+uqrPPzww2RnZ9OyZUsAgoODmT59enFyfPHFF9O5c2eefPJJBg0aRFFRET179gSgVq1a7Nq1q3j6wIEDOOcIDAxk1KhRjBgxgs2bNzNgwIDDjv/qq6/SsmVLgoODCQkJKf4CctWqVVx++eVHcwrTzKwV3ov2fefceN/yS4EXzcwBacBFJfa5G/julwquzBbkTsC/nHOX+eaHAzjnRlewfSDeYbhrHanc2NhYl5iYeLzDFRERkWPwfeoBLp2wlOFXnMXAS5rBtB6Qs5edvebStWtXNm7cCMDTTz9NXl5e8SA1B58dHTJkSKnyateuzd7dO4tbVCOatCLju08qbHl9+s3F5OXm8Oi1rSA/k8fnJhIW6GFI17q+lttssjP30+XVXSy+NZRa1Y2YZw+QeFcNosIC2J/nGLwol9UpRZxXL5CNaUW89pdQtu/38EGSMen6+sRvdTzzeTrvDW4H1WpAcA1umZRAn7iz+ctF50G1sOLlVKuBCw4j01UjNSeIlJwAdmQaWzONLfsdW/Y6ftpXSH5h6S6z9U4JoVFEGI0iw2gUEUrDyDDffCinnVKdoMCKW49mz55dfCPqD1FRUZUyMNaxaNCgAVOmTOGKK6447mX7u57r1atHUlIS4eGHt1SfaK1bt+aWW25h6NChx71s1fMhf4R6Pta/D2VbkI8HM0sG7nDOLTrSdpXZgrwCaG5mTYBkoA9wc8kNzKy+cy7FN9sT2FCJ8YiIiMhxdma9mnSIiWRWwjb+2rkpAc26wpL/I3TD2+TuT4MvJ0B+Ng33rCB+TRLM3wr5Wez49CvizqgJk+aXesa2XtABUobXo37NAFIOeKgblA1T/lzh8RtmOOK3OdicDtVqsGN3OnFn14PwuhAcBtXCSdqZxZbMebSaFgRm7Mg02s4MIWHus5zW4HRevzccqoXhgsNo0uYSmjyxnDnPTGTh0jf44MVCcnNz2b8/j1vjT+ONN94AIHPutexueRsfR1/A9vRstmdksz09hx0Z2WxPzyYrvwjvazq9IsKCaRQZRvMGYXRpGVqcDDeMCOX02qFUDw6s4Ax/2U033cTixYvJz88vHqzqRKoqyXFmZiY9evSolOQY/F/PJbut+tv5559fKUkbqJ5L+iPUc1X5+2BmNYD3fik5hkpsQfYFciXwLN6XOU91zj1pZqPwvtdqoZmNxpsYF+J9v9U9zrmNRypTLcgiIiJVy7trkhk8Zw3Tb+/AJbV+hpcuBByNJhzgh795n51LLwyl3aQ0Vg1tAdVq0Hb0t6wc1YXIyAhfIuttfR06dSl1Imox7PZejHnjE9Iz8xk3/F4S1m/lhZn/YcbLE3yttd590vcdoF27dqxatQqAtm3bsnLlysNesVJSTExM8TPIe/fuLR4FefLkyXzxxRfMmDGDvMIikjNy2J6Rw+JPljB/2st0Gzzelwhns+75u6h74+MEhXuPE1YtsLjFt2GJluCDSXDN6hWPCCsiIqVVRgvy0arMFmSccx8AH5RZ9liJ6eHA8MqMQURERCrX5S1PIyq8Gm8s28ol/WLhwbXgKaTrD8Po8O46Vq9eTWRQEF1T7qDeP2cC0K9fPyLv9b7d8c477+Tuu28lNjaWgc1+pEOHDjw2dTHh4eEkJiZC06Zs++ZtQuucDlHNSx07MjKSrl27Uq9evUPl+pJjb7l3Ext76B7rmWeeYevWrez+OY3cwDDmL1zCP+7uS2FREQGBQZzbcyAXPLWE1AO5OAeevGySX76LgJBQ6qfsp2FEKGvmDOe0iHAm3dWNRhHeBDiyRrVjHrlXRESqjkpNkEVEROSPLyQokN6xjXj58ySS9+Zweu2GANRvFMP675MIDAoiPT2dzz77jF27dhW/jikjI4OIiAhee+214rIODhpz8DVPr7zyCmPHjmX58uUMGjTosGNXVG7t2rV5asILbE/P4d01yezIyGHd95tZ8PrbhNSuy1WvrMZVT2Lf1wsJaXUlDboMICowjzUT+nNvr5uJqduIRhFhzH3hcfJ7XEGDulG8+FAcADcvaUFOTg49zm9wQupXREROHCXIIiIi8pvd1CGalz5PYk7CNoZc2gKApUuXcsMNN1BUVMRHH31E9+7di1t3u3fvzqJFiw57HdO7775LfHw8ALfddhtxcXGMHTuWp59++rBj7ssp4LXZCzir3UXMX7uX7RnZBEe3ovPfJlDYpBO5BaUHwtr33zG07nk3K197hH6dYjiryenEF3xNdkYqr426nJ07ttF9Tl3G92lLQEAAK1euxHL20euqKyj5eNeFF17ISy+9dDyrT0REqgglyCIiIvKbNYoMo0uLusxZsZ37uzXHFRWyefNmHn74YeDYXvNUu05ddqXu5tONqSUGwMopfg54f24h+5Z/hSuC9R9soGb1IFxYHcIL93FVx8bFzwA3igzjmy8/4cuctkwc3Z+Y2f/ib92aExUVRY+z/0HPnj2JiW7IgQMHeOuttwgICMDj8TBkyBBmzpzJkiVLSsV43333MXHiRPbs2UOdOnUqsVZFROREU4IsIiIix0XfCxozYNoKFq9LpU2Uo3bt2sXryhsU1ONg257s4oR3e0Y2OQVFXDPpf2xPzyEtM4/M3AJun+ZtvQ0JCqChL+ltGx1Bo8hQluc3ICzI8cRjl1IrLJjHH0/wvj6qxznFx8nOzqbvuDEsXrz4sBg++ugjWrduzaeffkpSUhLdu3enc+fOzJgxgyuvvLJUUl9S3bp12blzpxJkEZE/GCXIIiIiclxcfOapNIwIZeayn7jwhrPIzc1l175ctmdks7MglOXLVrJ33jdsT8/m8/eW4zntHCbt+ax4/8AAIzCsNmRl0O2saE5xB5h0Wj3m39OJRhFhnFoz5LCBsGomn0N8fDy1wryjRO/YsYO4uLhS2yQlJbFlyxZatWpVvE3btm1JSEjg9ddfZ9iwYZgZZ5xxBk2aNGHjxo18/fXXfPHFF0yaNInMzEzy8/MJDw9nzJgxAOTm5hIaGlqJtSkiIv6gBFlERESOi8AA4+aO0YxbtInrpnxDcnoWHR//EAuqRlFOFClfxpPb+kZOjwglZ8tqHho6khaNG9Aw0vte4Pq1qjO8oA91itYy7PoejBkzhlt6X0e7xpEkJCTw0AsvMGPGjFLHvOyyy3jkkUfIyMgAYPHixYwePbrUNueddx67d+8uni/5mqfo6GiWLFlC586dSU1NZdOmTTRt2pQ333yzePtp06aRmJhYnBw759i1axcxMTGVVJMiIuIvSpBFRETkuLmpfTQJW9IJqxbI/gsuplt0NpdfdhGNIsJY0jKTp8cOZT/w7NgnGHB9R+DQ65gaxcYybNgwevfuzZQpU4iOjmbevHkAbNu2rdwW28jISEaOHEn79u0BeOyxx474mqeyRo4cSf/+/TnvvPNwzjF27FiioqKOeI4rV67kggsuIChIt1EiIn80Vt4zQVVZbGysKzmSpIiIiFRNq1evZvz48cycOfM3lzV06FD69u3L+eeffxwi+20GDx5Mz5496datm79DERH5QzKzlc65ir/drET66lNEREQqRZs2bejSpQtFRUUEBgb+prLKe82Tv7Rs2VLJsYjIH5RakEVERERERKTK8GcLcoA/DioiIiIiIiJS1ShBFhEREREREUEJsoiIiIiIiAigBFlEREREREQEUIIsIiIiIiIiAihBFhEREREREQGUIIuIiIiIiIgASpBFREREREREACXIIiIiIiIiIoASZBERERERERFACbKIiIiIiIgIoARZREREREREBFCCLCIiIiIiIgIoQRYREREREREBlCCLiIiIiIiIAEqQRURERERERAAw55y/Y/hVzOxnYKu/45BiUUCav4MQqSJ0PYgcoutBxEvXgsghR3s9NHbOnVrZwZTnd5cgS9ViZonOuVh/xyFSFeh6EDlE14OIl64FkUN+D9eDuliLiIiIiIiIoARZREREREREBFCCLL/dq/4OQKQK0fUgcoiuBxEvXQsih1T560HPIIuIiIiIiIigFmQRERERERERQAmyiIiIiIiICKAEWY6Rmd1gZuvMzGNmsWXWDTezH81sk5ld5q8YRU40M/uXmSWb2Rrfz5X+jknkRDOzy31//380s2H+jkfEn8zsJzP7zveZkOjveEROJDObama7zWxtiWWRZvaxmf3g+x3hzxjLowRZjtVa4FpgacmFZnYO0Ac4F7gcmGRmgSc+PBG/meCca+37+cDfwYicSL6/9y8CVwDnADf5PhdETmZdfJ8JVfrdryKVYBrefKCkYcAS51xzYIlvvkpRgizHxDm3wTm3qZxVvYA5zrk859wW4Eegw4mNTkRE/KQD8KNzbrNzLh+Yg/dzQURETjLOuaVAepnFvYDpvunpwNUnNKijoARZjrfTge0l5nf4lomcLO4zs2993YqqXLchkUqmzwCR0hyw2MxWmtld/g5GpAqo55xLAfD9ruvneA4T5O8ApOoys0+A08pZNcI5925Fu5WzTO8Skz+MI10XwEvA43j/zz8O/Bu4/cRFJ+J3+gwQKe0i59xOM6sLfGxmG32taiJSRSlBlgo55/58DLvtABqVmG8I7Dw+EYn439FeF2Y2GXivksMRqWr0GSBSgnNup+/3bjNbgPcxBCXIcjJLNbP6zrkUM6sP7PZ3QGWpi7UcbwuBPmYWYmZNgOZAgp9jEjkhfH/oD7oG72B2IieTFUBzM2tiZtXwDtq40M8xifiFmdUws5oHp4FL0eeCyELgNt/0bUBFvVL9Ri3IckzM7BrgeeBU4H0zW+Ocu8w5t87M5gLrgUJgkHOuyJ+xipxA48ysNd4upT8BA/0bjsiJ5ZwrNLP7gI+AQGCqc26dn8MS8Zd6wAIzA+899yzn3CL/hiRy4pjZbCAOiDKzHcA/gTHAXDO7A9gG3OC/CMtnzunRIBERERERERF1sRYRERERERFBCbKIiIiIiIgIoARZREREREREBFCCLCIiIiIiIgIoQRYREREREREBlCCLiMhJwMwyj3N5cWb2dZllQWaWWuZ92GX3+5eZPXQ8YynnGPeb2QYze/Mot48xsyO+m/VothEREfkj0HuQRUREfr2lQEMzi3HO/eRb9mdgrXMuxX9hAXAvcIVzbouf4xAREfndUQuyiIiclMyssZktMbNvfb+jfcubmdkyM1thZqPKa312znmAecCNJRb3AWb7yvirb/9vzGy+mYWVc/x4M4v1TUeZ2U++6UAze9q3/7dmNrCC+P9uZmt9Pw/4lr0MNAUWmtmDZbaPMbMvzGyV7+fCcsrsb2bvmtkiM9tkZv8ssTrQzCab2TozW2xmoUd7riIiIr8XSpBFRORk9QIwwzl3PvAm8Jxv+URgonOuPbDzCPvPxpsUY2YhwJXAfN+6d5xz7Z1zrYANwB2/Iq47gH2+47cH/mpmTUpuYGbtgAFAR+AC3zZtnHN3+2Lu4pybUKbc3UB351xbvIn9c5SvA3AL0Bq44WASDzQHXnTOnQvsBa47DucqIiJSpShBFhGRk1UnYJZveibwpxLL5/mmZ5Xd6SDn3Aog3MxaAFcAy5xzGb7VLX2ttd/hTTbP/RVxXQr0M7M1wHKgDt7ktKQ/AQucc1nOuUzgHaDzL5QbDEz2xTQPOKeC7T52zu1xzuX4yj1YL1ucc2t80yuBGN/0bzlXERGRKkXPIIuIiHi5I600syeBqwCcc619i+fgbUU+G1/3ap9pwNXOuW/MrD8QV06RhRz6orp6yUMBf3POfXSkcI4UawUeBFKBVr7j5lawXdl6ODifV2JZERDqm57GL5+riIjI74JakEVE5GT1Fb4u0nhbPr/0TS/jUPfhg+txzo1wzrUukRyDNym+FegKLCyxvCaQYmbBvrLL8xPQzjd9fYnlHwH3+PbFzM40sxpl9l0KXG1mYb511wBfHOFcAWoBKb7np/sCgRVs193MIn3PGF8N/O8Xyj2acxUREfldUAuyiIicDMLMbEeJ+fHA/cBUMxsK/Iz3mV6AB4A3zGwI8D6wr6JCnXPrzSwbWOmcyyqxaiTe7tFbge/wJpFlPQPMNbO+wKcllr+Gt/vyKjMzX2xXlznuKjObBiQc3Mc5t7qiOH0mAfPN7AbgMyCrgu2+xNvl/AxglnMu0cxijlDu0ZyriIjI74I5d8QeZSIiIicV3yjMOc45Z2Z9gJucc738HdeJ4OsiHeucu8/fsYiIiPiDWpBFRERKawe84Gu93Qvc7ud4RERE5ARRC7KIiIiIiIgIGqRLREREREREBFCCLCIiIiIiIgIoQRYREREREREBlCCLiIiIiIiIAEqQRURERERERAD4f3vv5J0Fn5AjAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x26124531208>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "the AUC score for each alpha value is :  [0.93  0.938 0.943 0.946 0.77  0.484 0.499 0.499 0.499 0.499]\n"
     ]
    }
   ],
   "source": [
    "#Plot MSE for each value of alpha\n",
    "opt_alpha = find_optimal_alpha(X_trainf,y_en_train,5)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "Dtpj7iMSJkln"
   },
   "source": [
    "### [5.2.1] Top 10 important features of positive class from<font color='red'> SET 2</font>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 115,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "g3flcTmmJklo"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "The ROC_AUC score of the NB classifier for alpha = 0.010000 is 96.783151%\n",
      "The wordcloud for top 10 positive features are:\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x26108145438>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Please write all the code with proper documentation\n",
    "y_pred,NB_opt = pred_test_data(X_trainf,y_en_train,X_testf,y_en_test,opt_alpha,tf_idf_vect)\n",
    "type(y_pred)\n",
    "\n",
    "pos_class_prob_sorted = (abs(NB_opt.feature_log_prob_[1, :])).argsort()\n",
    "\n",
    "pos_text = np.take(tf_idf_vect.get_feature_names(), pos_class_prob_sorted[1:11])\n",
    "\n",
    "posi_text = \" \".join(words for words in pos_text)\n",
    "\n",
    "#pass the list of top 10 words for each class to wordcloud\n",
    "print(\"The wordcloud for top 10 positive features are:\")\n",
    "generate_wordcloud(posi_text)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "UovIVRyEJklq"
   },
   "source": [
    "### [5.2.2] Top 10 important features of negative class from<font color='red'> SET 2</font>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 116,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "i3i3cShuJklr"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The wordcloud for top 10 negative features are:\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x26107869470>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Please write all the code with proper documentation\n",
    "neg_class_prob_sorted = (abs(NB_opt.feature_log_prob_[0, :])).argsort()\n",
    "\n",
    "neg_text = np.take(tf_idf_vect.get_feature_names(), neg_class_prob_sorted[1:11])\n",
    "nega_text = \" \".join(words for words in neg_text)\n",
    "\n",
    "print(\"The wordcloud for top 10 negative features are:\")\n",
    "generate_wordcloud(nega_text)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 117,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Confusion matrix is :  [[ 3466  1210]\n",
      " [  613 24711]]\n",
      "The True Positive value is:   24711\n",
      "The True Negative value is:   3466\n",
      "The False Positive value is:   613\n",
      "The False Negative value is:   1210\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x2616a617f28>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "confusion_heatmap(y_en_test,y_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 118,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The precision score for our model is:  0.9533197021719841\n",
      "The f1 score for our model is:  0.964425797638794\n",
      "The recall score for our model is:  0.975793713473385\n",
      "The accuracy of our model is:  0.9392333333333334\n"
     ]
    }
   ],
   "source": [
    "#determining metrics precision,recall,f-1 scores for our model\n",
    "\n",
    "metric_score(y_en_test,y_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 119,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x2616478bc88>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "ytrain_prob = NB_opt.predict_proba(X_trainf)\n",
    "ytrain_prob = ytrain_prob[:, 1]\n",
    "\n",
    "ytest_prob = NB_opt.predict_proba(X_testf)\n",
    "ytest_prob = ytest_prob[:,1]\n",
    "\n",
    "plot_roc_curve(y_en_train,ytrain_prob,y_en_test,ytest_prob,'-',':',\"Train ROC AUC\",\"Test ROC AUC\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Conclusion:\n",
    "1. Naive Bayes using 5-fold validation was performed on a sample dataset of AFF for Bag of words.We have chosen Multinomial NB becuase our feature vector TFIDF has real valued fractions. \n",
    "2. We chose the standard 70-30 split i.e. 70% as train data and 30% as test data.\n",
    "3. The optimal alpha in was computed to be 0.01 based on AUC score as our metric.\n",
    "4. Confusion matrix was plotted for the same. We got True Negative value of 3466 and True Positive value of 24711. we also get a low false negative value of 1210.\n",
    "5. The accuracy,precision, recall and f1 scores are 92.92 % , 95.33%, 97.57% and 96.44 % respectively. This means that we are performing remarkabely well in Postive class classification. We get a very high recall value which suggests that most of our classification were correct. \n",
    "6. Also NB helps us in getting the feature importance, we got the top 10 words that helped us in classifying a given point for both classes. A word cloud was generated for the same.\n",
    "7. It could be noted that the first highest prob belonged to review length, hence we took the next 10 important features! \n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "vst5Y7mmJklu"
   },
   "source": [
    "# [6] Conclusions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 120,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "De9HYtIGJklv"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "+---------------+-------+--------+\n",
      "|     METRIC    | B.O.W | TF-IDF |\n",
      "+---------------+-------+--------+\n",
      "|     Alpha     |  0.1  |  0.01  |\n",
      "| True Positive | 23524 | 24711  |\n",
      "| True Negative |  3896 |  3466  |\n",
      "|    Accuracy   |  91.4 | 93.92  |\n",
      "|   Precision   | 96.79 | 95.33  |\n",
      "|     Recall    | 92.89 | 97.57  |\n",
      "|   F1 Scores   |  94.8 | 96.44  |\n",
      "+---------------+-------+--------+\n"
     ]
    }
   ],
   "source": [
    "# Please compare all your models using Prettytable library\n",
    "\n",
    "from prettytable import PrettyTable\n",
    "    \n",
    "x = PrettyTable([\"METRIC\", \"B.O.W\", \"TF-IDF\"])\n",
    "\n",
    "x.add_row([\"Alpha\",.1, .01])\n",
    "x.add_row([\"True Positive\",23524, 24711])\n",
    "x.add_row([\"True Negative\",3896, 3466])\n",
    "x.add_row([\"Accuracy\",91.4, 93.92])\n",
    "x.add_row([\"Precision\",96.79, 95.33])\n",
    "x.add_row([\"Recall\",92.89, 97.57])\n",
    "x.add_row([\"F1 Scores\",94.80, 96.44])\n",
    "\n",
    "print(x)"
   ]
  }
 ],
 "metadata": {
  "colab": {
   "name": "04 Amazon Fine Food Reviews Analysis_NaiveBayes.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": "3.6.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}