602_Project_Amazon_Laptop_Review.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Welcome to DATA 602: Introduction to Data Analysis and Machine Learning Spring 2019 Term Project\n",
    "#### My experiments with word2vec among other things...\n",
    "\n",
    "We are going to use the following datasets:\n",
    "* [Amazon Laptop Review Dataset](http://sifaka.cs.uiuc.edu/~wang296/Data/LARA/Amazon/AmazonReviews.zip)\n",
    "\n",
    "Below, we install and import required libraries"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Requirement already satisfied: gensim in /opt/conda/lib/python3.6/site-packages (3.7.2)\n",
      "Requirement already satisfied: tensorflow in /opt/conda/lib/python3.6/site-packages (1.13.1)\n",
      "Requirement already satisfied: wordcloud in /opt/conda/lib/python3.6/site-packages (1.5.0)\n",
      "Requirement already satisfied: smart-open>=1.7.0 in /opt/conda/lib/python3.6/site-packages (from gensim) (1.8.3)\n",
      "Requirement already satisfied: numpy>=1.11.3 in /opt/conda/lib/python3.6/site-packages (from gensim) (1.13.3)\n",
      "Requirement already satisfied: scipy>=0.18.1 in /opt/conda/lib/python3.6/site-packages (from gensim) (1.1.0)\n",
      "Requirement already satisfied: six>=1.5.0 in /opt/conda/lib/python3.6/site-packages (from gensim) (1.11.0)\n",
      "Requirement already satisfied: absl-py>=0.1.6 in /opt/conda/lib/python3.6/site-packages (from tensorflow) (0.7.1)\n",
      "Requirement already satisfied: wheel>=0.26 in /opt/conda/lib/python3.6/site-packages (from tensorflow) (0.32.3)\n",
      "Requirement already satisfied: astor>=0.6.0 in /opt/conda/lib/python3.6/site-packages (from tensorflow) (0.7.1)\n",
      "Requirement already satisfied: tensorflow-estimator<1.14.0rc0,>=1.13.0 in /opt/conda/lib/python3.6/site-packages (from tensorflow) (1.13.0)\n",
      "Requirement already satisfied: protobuf>=3.6.1 in /opt/conda/lib/python3.6/site-packages (from tensorflow) (3.6.1)\n",
      "Requirement already satisfied: gast>=0.2.0 in /opt/conda/lib/python3.6/site-packages (from tensorflow) (0.2.2)\n",
      "Requirement already satisfied: tensorboard<1.14.0,>=1.13.0 in /opt/conda/lib/python3.6/site-packages (from tensorflow) (1.13.1)\n",
      "Requirement already satisfied: keras-applications>=1.0.6 in /opt/conda/lib/python3.6/site-packages (from tensorflow) (1.0.7)\n",
      "Requirement already satisfied: termcolor>=1.1.0 in /opt/conda/lib/python3.6/site-packages (from tensorflow) (1.1.0)\n",
      "Requirement already satisfied: keras-preprocessing>=1.0.5 in /opt/conda/lib/python3.6/site-packages (from tensorflow) (1.0.9)\n",
      "Requirement already satisfied: grpcio>=1.8.6 in /opt/conda/lib/python3.6/site-packages (from tensorflow) (1.20.1)\n",
      "Requirement already satisfied: pillow in /opt/conda/lib/python3.6/site-packages (from wordcloud) (5.1.0)\n",
      "Requirement already satisfied: boto>=2.32 in /opt/conda/lib/python3.6/site-packages (from smart-open>=1.7.0->gensim) (2.49.0)\n",
      "Requirement already satisfied: requests in /opt/conda/lib/python3.6/site-packages (from smart-open>=1.7.0->gensim) (2.20.1)\n",
      "Requirement already satisfied: boto3 in /opt/conda/lib/python3.6/site-packages (from smart-open>=1.7.0->gensim) (1.9.142)\n",
      "Requirement already satisfied: mock>=2.0.0 in /opt/conda/lib/python3.6/site-packages (from tensorflow-estimator<1.14.0rc0,>=1.13.0->tensorflow) (3.0.4)\n",
      "Requirement already satisfied: setuptools in /opt/conda/lib/python3.6/site-packages (from protobuf>=3.6.1->tensorflow) (40.6.2)\n",
      "Requirement already satisfied: werkzeug>=0.11.15 in /opt/conda/lib/python3.6/site-packages (from tensorboard<1.14.0,>=1.13.0->tensorflow) (0.15.2)\n",
      "Requirement already satisfied: markdown>=2.6.8 in /opt/conda/lib/python3.6/site-packages (from tensorboard<1.14.0,>=1.13.0->tensorflow) (3.1)\n",
      "Requirement already satisfied: h5py in /opt/conda/lib/python3.6/site-packages (from keras-applications>=1.0.6->tensorflow) (2.7.1)\n",
      "Requirement already satisfied: chardet<3.1.0,>=3.0.2 in /opt/conda/lib/python3.6/site-packages (from requests->smart-open>=1.7.0->gensim) (3.0.4)\n",
      "Requirement already satisfied: urllib3<1.25,>=1.21.1 in /opt/conda/lib/python3.6/site-packages (from requests->smart-open>=1.7.0->gensim) (1.23)\n",
      "Requirement already satisfied: idna<2.8,>=2.5 in /opt/conda/lib/python3.6/site-packages (from requests->smart-open>=1.7.0->gensim) (2.7)\n",
      "Requirement already satisfied: certifi>=2017.4.17 in /opt/conda/lib/python3.6/site-packages (from requests->smart-open>=1.7.0->gensim) (2018.11.29)\n",
      "Requirement already satisfied: s3transfer<0.3.0,>=0.2.0 in /opt/conda/lib/python3.6/site-packages (from boto3->smart-open>=1.7.0->gensim) (0.2.0)\n",
      "Requirement already satisfied: jmespath<1.0.0,>=0.7.1 in /opt/conda/lib/python3.6/site-packages (from boto3->smart-open>=1.7.0->gensim) (0.9.4)\n",
      "Requirement already satisfied: botocore<1.13.0,>=1.12.142 in /opt/conda/lib/python3.6/site-packages (from boto3->smart-open>=1.7.0->gensim) (1.12.142)\n",
      "Requirement already satisfied: python-dateutil<3.0.0,>=2.1; python_version >= \"2.7\" in /opt/conda/lib/python3.6/site-packages (from botocore<1.13.0,>=1.12.142->boto3->smart-open>=1.7.0->gensim) (2.7.5)\n",
      "Requirement already satisfied: docutils>=0.10 in /opt/conda/lib/python3.6/site-packages (from botocore<1.13.0,>=1.12.142->boto3->smart-open>=1.7.0->gensim) (0.14)\n",
      "Requirement already satisfied: xgboost in /opt/conda/lib/python3.6/site-packages (0.82)\n",
      "Requirement already satisfied: numpy in /opt/conda/lib/python3.6/site-packages (from xgboost) (1.13.3)\n",
      "Requirement already satisfied: scipy in /opt/conda/lib/python3.6/site-packages (from xgboost) (1.1.0)\n",
      "Requirement already satisfied: keras in /opt/conda/lib/python3.6/site-packages (2.2.4)\n",
      "Requirement already satisfied: h5py in /opt/conda/lib/python3.6/site-packages (from keras) (2.7.1)\n",
      "Requirement already satisfied: keras-applications>=1.0.6 in /opt/conda/lib/python3.6/site-packages (from keras) (1.0.7)\n",
      "Requirement already satisfied: six>=1.9.0 in /opt/conda/lib/python3.6/site-packages (from keras) (1.11.0)\n",
      "Requirement already satisfied: pyyaml in /opt/conda/lib/python3.6/site-packages (from keras) (3.13)\n",
      "Requirement already satisfied: numpy>=1.9.1 in /opt/conda/lib/python3.6/site-packages (from keras) (1.13.3)\n",
      "Requirement already satisfied: scipy>=0.14 in /opt/conda/lib/python3.6/site-packages (from keras) (1.1.0)\n",
      "Requirement already satisfied: keras-preprocessing>=1.0.5 in /opt/conda/lib/python3.6/site-packages (from keras) (1.0.9)\n",
      "Requirement already satisfied: nltk in /opt/conda/lib/python3.6/site-packages (3.4.1)\n",
      "Requirement already satisfied: six in /opt/conda/lib/python3.6/site-packages (from nltk) (1.11.0)\n",
      "Collecting string\n",
      "\u001b[31m  Could not find a version that satisfies the requirement string (from versions: )\u001b[0m\n",
      "\u001b[31mNo matching distribution found for string\u001b[0m\n",
      "Requirement already satisfied: tqdm in /opt/conda/lib/python3.6/site-packages (4.31.1)\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[nltk_data] Downloading package stopwords to /home/jovyan/nltk_data...\n",
      "[nltk_data]   Package stopwords is already up-to-date!\n",
      "[nltk_data] Downloading package punkt to /home/jovyan/nltk_data...\n",
      "[nltk_data]   Package punkt is already up-to-date!\n",
      "WARNING: Logging before flag parsing goes to stderr.\n",
      "W0506 18:14:10.503599 139801131669312 __init__.py:56] Some hub symbols are not available because TensorFlow version is less than 1.14\n",
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "!pip install gensim tensorflow wordcloud\n",
    "!pip install -q tensorflow-hub\n",
    "!pip install xgboost\n",
    "!pip install keras\n",
    "!pip install nltk\n",
    "!pip install string\n",
    "!pip install tqdm\n",
    "\n",
    "import gzip\n",
    "import gensim\n",
    "import os\n",
    "import sys\n",
    "import json\n",
    "import shutil\n",
    "import time\n",
    "import re\n",
    "import tarfile\n",
    "import zipfile\n",
    "\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import collections\n",
    "import math\n",
    "import random\n",
    "import datetime as dt\n",
    "\n",
    "import matplotlib.pyplot as plt \n",
    "import seaborn as sns\n",
    "%matplotlib inline\n",
    "from wordcloud import WordCloud, STOPWORDS, ImageColorGenerator\n",
    "\n",
    "import string\n",
    "import nltk\n",
    "nltk.download('stopwords')\n",
    "from nltk.corpus import stopwords\n",
    "nltk.download('punkt')\n",
    "from nltk.tokenize import word_tokenize\n",
    "\n",
    "from gensim.test.utils import get_tmpfile\n",
    "from gensim.models import Word2Vec, FastText\n",
    "\n",
    "from tqdm import tqdm\n",
    "tqdm.pandas(desc=\"progress-bar\")\n",
    "from gensim.models import Doc2Vec\n",
    "from gensim.models.doc2vec import LabeledSentence\n",
    "import multiprocessing\n",
    "\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from sklearn.svm import SVC\n",
    "from xgboost import XGBClassifier\n",
    "from sklearn.decomposition import PCA\n",
    "from sklearn.manifold import TSNE\n",
    "from sklearn.pipeline import Pipeline\n",
    "\n",
    "\n",
    "import tensorflow as tf\n",
    "import tensorflow_hub as hub\n",
    "\n",
    "from keras.preprocessing.text import Tokenizer\n",
    "from keras.preprocessing.sequence import pad_sequences\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Embedding,Dense,Flatten,GlobalMaxPooling1D,LSTM,Dropout, Activation,Bidirectional\n",
    "from keras.layers.convolutional import Conv1D,MaxPooling1D \n",
    "from keras.optimizers import Adam\n",
    "\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.metrics import confusion_matrix, classification_report, accuracy_score, roc_curve, precision_recall_curve \n",
    "from sklearn import utils\n",
    "\n",
    "\"\"\"\n",
    "We will ignore FutureWarning and DeprecationWarning\n",
    "\"\"\"\n",
    "import warnings\n",
    "warnings.filterwarnings(\"ignore\", category=FutureWarning)\n",
    "warnings.filterwarnings(\"ignore\", category=DeprecationWarning)\n",
    "\n",
    "\"\"\"\n",
    "We will ignore warnings\n",
    "\"\"\"\n",
    "warnings.filterwarnings(\"ignore\")\n",
    "\n",
    "if not sys.warnoptions:\n",
    "    warnings.simplefilter(\"ignore\")\n",
    "program_start_time=time.time()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Perform some housekeeping tasks"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "data_source_url = 'http://sifaka.cs.uiuc.edu/~wang296/Data/LARA/Amazon/AmazonReviews.zip'\n",
    "datasets = '/datasets/'\n",
    "data_file_path = os.getcwd()+datasets+'AmazonReviews.zip'\n",
    "data = 'data_'\n",
    "AMAZON = 'amazon'\n",
    "unzip_folder = os.getcwd()+datasets+data+AMAZON\n",
    "AMAZON_LAPTOPS = 'amazon_laptops'\n",
    "data_folder = os.getcwd()+datasets+data+AMAZON_LAPTOPS\n",
    "data_json_file = data+AMAZON_LAPTOPS+'.json'\n",
    "\n",
    "def check_if_file_exists(file):\n",
    "    '''\n",
    "    Checks if 'file' exists\n",
    "    '''\n",
    "    try:\n",
    "        fh = open(file, 'r')\n",
    "        return True\n",
    "    except FileNotFoundError:\n",
    "        print('Please make sure file: ' + file + ' is present before continuing')\n",
    "        return False\n",
    "\n",
    "def check_if_dir_exists(directory):\n",
    "    '''\n",
    "    Checks if 'directory' exists\n",
    "    '''\n",
    "    return(os.path.isdir(directory))\n",
    "\n",
    "def store_json(write_this_data):\n",
    "    '''\n",
    "    Store json if we are processing the first time\n",
    "    '''\n",
    "    open(data_json_file, 'w').write(json.dumps(write_this_data))\n",
    "\n",
    "def return_data_json(json_file):\n",
    "    '''\n",
    "    Return data json from data json file\n",
    "    '''\n",
    "    with open(json_file, encoding='utf-8') as data_file:\n",
    "        return json.loads(data_file.read())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Download data source**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Data file already exists. Not downloading again!\n"
     ]
    }
   ],
   "source": [
    "if not check_if_file_exists(data_file_path):\n",
    "    print('Start of data download')\n",
    "    wget.download(data_source_url, os.getcwd()+datasets)\n",
    "    print('Download complete')\n",
    "else:\n",
    "    print('Data file already exists. Not downloading again!')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Data foler exists. Won't copy again!\n"
     ]
    }
   ],
   "source": [
    "if not check_if_dir_exists(data_folder):\n",
    "    zip_ref = zipfile.ZipFile(data_file_path, 'r')\n",
    "    zip_ref.extractall(unzip_folder)\n",
    "    zip_ref.close()\n",
    "    try:\n",
    "        shutil.copytree(unzip_folder+'/laptops/', data_folder)\n",
    "    except FileExistsError as exc: \n",
    "        print('Already copied')\n",
    "else:\n",
    "    print('Data foler exists. Won\\'t copy again!')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "40762"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data = {}\n",
    "all_reviews = []\n",
    "if not check_if_file_exists(data_json_file):\n",
    "    for root, dirs, files in os.walk(data_folder):\n",
    "        for name in files:\n",
    "            full = os.path.join(root, name)\n",
    "            json_string = return_data_json(full)\n",
    "            for review in json_string['Reviews']:\n",
    "                curr_review = {}\n",
    "                if float(review['Overall']) <= 2.0:\n",
    "                    curr_review[\"rating\"] = 0\n",
    "                else:\n",
    "                    curr_review[\"rating\"] = 1\n",
    "                curr_review[\"review\"] = review['Content']\n",
    "                all_reviews.append(curr_review)\n",
    "    data[AMAZON_LAPTOPS] = all_reviews\n",
    "    store_json(data)\n",
    "pre_loaded_data = return_data_json(data_json_file)\n",
    "len(pre_loaded_data[AMAZON_LAPTOPS])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Creating review dataframe and Data clean up**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Before Cleanup : Shape of the Data Frame : (40762, 2)\n",
      "Remove missing values.\n",
      "Drop columns with duplicate data.\n",
      "After Cleanup : Shape of the Data Frame : (40744, 2)\n",
      "Counting null data per column.\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "rating    0\n",
       "review    0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "amazon_df = pd.DataFrame(pre_loaded_data[AMAZON_LAPTOPS])\n",
    "print('Before Cleanup : Shape of the Data Frame : {}'.format(amazon_df.shape))\n",
    "print('Remove missing values.')\n",
    "amazon_df.dropna(inplace=True)\n",
    "amazon_df.reset_index(drop=True,inplace=True)\n",
    "print('Drop columns with duplicate data.')\n",
    "amazon_df.drop_duplicates()\n",
    "print('After Cleanup : Shape of the Data Frame : {}'.format(amazon_df.shape))\n",
    "print('Counting null data per column.')\n",
    "amazon_df.isnull().sum()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* **The dataframe contains 40744 rows and 2 columns**"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Let us look at the data types of columns**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "rating     int64\n",
       "review    object\n",
       "dtype: object"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "amazon_df.dtypes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([1, 0])"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\"\"\"\n",
    "Ratings\n",
    "\"\"\"\n",
    "amazon_df.rating.unique()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Let us explore the data a bit using head(), tail(), info(), describe()**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>rating</th>\n",
       "      <th>review</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>I've had the S7-391 with 4Gb RAM and a 256 GB ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>This would be a 5-star review if it were not f...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>Exactly as described! Fast laptop. One of the ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1</td>\n",
       "      <td>I bought this laptop after doing my research a...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>After living with the Aspire S7 for 5 months, ...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   rating                                             review\n",
       "0       1  I've had the S7-391 with 4Gb RAM and a 256 GB ...\n",
       "1       1  This would be a 5-star review if it were not f...\n",
       "2       1  Exactly as described! Fast laptop. One of the ...\n",
       "3       1  I bought this laptop after doing my research a...\n",
       "4       0  After living with the Aspire S7 for 5 months, ..."
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "amazon_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>rating</th>\n",
       "      <th>review</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>40739</th>\n",
       "      <td>1</td>\n",
       "      <td>Great to use once you get used to Windows 8 -w...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>40740</th>\n",
       "      <td>1</td>\n",
       "      <td>I love this product. It is amazing and it has ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>40741</th>\n",
       "      <td>1</td>\n",
       "      <td>It fast and easy to use win 8 battery very goo...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>40742</th>\n",
       "      <td>1</td>\n",
       "      <td>The computer is a bit light,the outside of the...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>40743</th>\n",
       "      <td>0</td>\n",
       "      <td>Screen failed after 3 months and only limited ...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       rating                                             review\n",
       "40739       1  Great to use once you get used to Windows 8 -w...\n",
       "40740       1  I love this product. It is amazing and it has ...\n",
       "40741       1  It fast and easy to use win 8 battery very goo...\n",
       "40742       1  The computer is a bit light,the outside of the...\n",
       "40743       0  Screen failed after 3 months and only limited ..."
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "amazon_df.tail()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 40744 entries, 0 to 40743\n",
      "Data columns (total 2 columns):\n",
      "rating    40744 non-null int64\n",
      "review    40744 non-null object\n",
      "dtypes: int64(1), object(1)\n",
      "memory usage: 636.7+ KB\n"
     ]
    }
   ],
   "source": [
    "amazon_df.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>rating</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>40744.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>0.807702</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>0.394111</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             rating\n",
       "count  40744.000000\n",
       "mean       0.807702\n",
       "std        0.394111\n",
       "min        0.000000\n",
       "25%        1.000000\n",
       "50%        1.000000\n",
       "75%        1.000000\n",
       "max        1.000000"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "amazon_df.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>review</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>40744</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>unique</th>\n",
       "      <td>36471</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>top</th>\n",
       "      <td>Customers service is a crapI bought this lapto...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>freq</th>\n",
       "      <td>11</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                   review\n",
       "count                                               40744\n",
       "unique                                              36471\n",
       "top     Customers service is a crapI bought this lapto...\n",
       "freq                                                   11"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "amazon_df.describe(include='object')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>rating</th>\n",
       "      <th>review</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>40744.000000</td>\n",
       "      <td>40744</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>unique</th>\n",
       "      <td>NaN</td>\n",
       "      <td>36471</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>top</th>\n",
       "      <td>NaN</td>\n",
       "      <td>Customers service is a crapI bought this lapto...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>freq</th>\n",
       "      <td>NaN</td>\n",
       "      <td>11</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>0.807702</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>0.394111</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              rating                                             review\n",
       "count   40744.000000                                              40744\n",
       "unique           NaN                                              36471\n",
       "top              NaN  Customers service is a crapI bought this lapto...\n",
       "freq             NaN                                                 11\n",
       "mean        0.807702                                                NaN\n",
       "std         0.394111                                                NaN\n",
       "min         0.000000                                                NaN\n",
       "25%         1.000000                                                NaN\n",
       "50%         1.000000                                                NaN\n",
       "75%         1.000000                                                NaN\n",
       "max         1.000000                                                NaN"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "amazon_df.describe(include='all')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Creating a new column called \"review_length\" which is the length of the review column.**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>rating</th>\n",
       "      <th>review</th>\n",
       "      <th>review_length</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>I've had the S7-391 with 4Gb RAM and a 256 GB ...</td>\n",
       "      <td>2341</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>This would be a 5-star review if it were not f...</td>\n",
       "      <td>1771</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>Exactly as described! Fast laptop. One of the ...</td>\n",
       "      <td>166</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1</td>\n",
       "      <td>I bought this laptop after doing my research a...</td>\n",
       "      <td>919</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>After living with the Aspire S7 for 5 months, ...</td>\n",
       "      <td>1920</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   rating                                             review  review_length\n",
       "0       1  I've had the S7-391 with 4Gb RAM and a 256 GB ...           2341\n",
       "1       1  This would be a 5-star review if it were not f...           1771\n",
       "2       1  Exactly as described! Fast laptop. One of the ...            166\n",
       "3       1  I bought this laptop after doing my research a...            919\n",
       "4       0  After living with the Aspire S7 for 5 months, ...           1920"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "amazon_df['review_length'] = amazon_df['review'].apply(len)\n",
    "amazon_df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### **Data Visualization and Exploratory Data Analysis**"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Using FacetGrid from the seaborn library to create a grid of two histograms of review_length based off of the ratings**\n",
    "\n",
    "[FacetGrid reference](https://seaborn.pydata.org/generated/seaborn.FacetGrid.html)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "sns.set_style('darkgrid')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x360 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "g = sns.FacetGrid(amazon_df,col='rating',size=5)\n",
    "g.map(plt.hist,'review_length',bins=50)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's try to explain why the x-axis goes all the way to 20000ish, this must mean that there is some really long message!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "count    40744.000000\n",
       "mean       828.389873\n",
       "std       1270.138594\n",
       "min          3.000000\n",
       "25%        185.000000\n",
       "50%        407.000000\n",
       "75%        932.000000\n",
       "max      21649.000000\n",
       "Name: review_length, dtype: float64"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "amazon_df.review_length.describe()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* Max review_length 21649.000000 means review is 21649 character long"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Creating a boxplot of review_length for each rating category.**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAngAAAH2CAYAAADjxfEPAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvhp/UCwAAIABJREFUeJzs3Xd4VGXe//HPJNRQMhAJTWD1kYBSA5JO0CgBNJSgrkpbsCIsCFaw4QKCiy6KBBFEdxdxlaIBDatPfoAQICRIE2mLqECkJECYINKSmfP7I09mnbSZyIQTh/frurzkvufOme+ZOTP55D7NYhiGIQAAAPgMP7MLAAAAgHcR8AAAAHwMAQ8AAMDHEPAAAAB8DAEPAADAxxDwAAAAfAwBD7/Zp59+qjZt2jj/u/HGG9W9e3c9/vjj+uGHH8wuT5LUpk0bzZ49u8I/949//EOpqaler+edd97RLbfcoptuukn9+/f3+vIraujQoRo6dKhpz//TTz+pTZs2+vTTT02rocjevXs1e/Zs2Wy2Eo+1adNGb7zxxm9edlV7370pLi5OTz31lGnPP3v2bG3atKlE/4QJExQXF2dCRd5T1roBnqhmdgH4/Zs1a5aaNGkiu92urKwsvf322xo+fLhWrlypevXqmV3eb7Jw4UJ16dJF8fHxXlvmzp079cYbb+jBBx/U7bffrjp16nht2b/VpEmTzC6hyti7d6+SkpLUr18/Wa1Wry23Kr7vviQpKUkjR45UZGSkS/+oUaM0bNgwk6ryjrLWDfAEAQ+X7cYbb1SrVq0kSV27dlVwcLBGjBihbdu2qUePHiZXV3V8//33kqT7779fLVq0qNDPGoah/Px81ahRw6s13XDDDV5dHkq6nPe9NJcuXfL6dlCVeGv9WrZs6YVq4I6vb4+/Z+yihdfVrVtXklRQUODSn5aWpnvvvVcdO3ZU165dNWrUKJddufv371fHjh01bdo0l5+bOXOm2rdvrz179kiSMjMz1aZNG/3v//6vJkyYoG7duqlLly568skndfr0abf1uasjLi5OR44c0eeff+7c/TxhwoRyl7lz504NHz5coaGh6ty5s/70pz9p586dzseHDh3qXMbtt9/udtdx0W6vZcuWqXfv3mrfvr3WrVsnSTp//rxee+01xcXFqX379oqLi9PcuXPlcDgkSTk5Obrpppv0wQcflFju/Pnz1a5dO+Xm5jrrKr6LNjc3V5MmTVL37t3Vvn179e7dW4sXL3Z5vG3btlqxYoWzb82aNWrTpo3Lrrrz58+rffv2+vDDD8t97UqzefNm/elPf3K+ng8++KD279/vMmbo0KG6//77lZ6ersTERHXq1EkJCQlatWpVieWlpKSod+/e6tChg/r27avVq1e7rPunn36qiRMnSpLi4+Od7/tPP/3kspyFCxcqLi5OoaGhGjJkiL777rty16O89/3s2bOaPHmyYmJi1L59e/Xq1Uv/+Mc/9OubCxVt66mpqXrhhRcUERGhqKiocp8zKytLTz75pCIiItS+fXv1799f/+///T+XMYcOHdLTTz+tuLg4dezYUbfddpsmTZqkvLy8EsvbvHmzRowYoa5du6pz587q16+fli5dWmLcypUr1adPH3Xu3FkDBw7Uli1byq1TKtwF2aZNG+3fv18PPvigQkND9fjjj0uSNmzYoIcfflgxMTHO9/b999+X3W53/nybNm0kFe4CL3rPil7f4rtoiw4H+PjjjzVr1izFxMTo5ptv1siRI3X8+HGXus6fP69JkyYpPDxcoaGhGj16tLZt2+bx4QT79u3T6NGjFR4ero4dO6pXr16aN2+e8/HLXTfJs8+I3W7XG2+84XyeYcOG6fvvvy/1+8fd96L038/cmjVrNGDAALVv317/+te/1LdvX40ePbrE61C0/a5fv97tawbvYwYPl81ut6ugoEAOh0NZWVmaOXOmgoKCFB4e7hyTlpamRx99VBEREXrjjTd07tw5vfXWWxo0aJBWrFihxo0bKyQkRM8++6ymTJmi6Oho9ejRQxkZGXr33Xf1zDPP6KabbnJ53mnTpikqKkp/+9vfdOjQIc2cOVM5OTmlBpuK1JGUlKRHHnlEbdq00ZgxYyRJDRs2LHOZ+/bt05AhQ3TDDTdo+vTpslgsmj9/voYMGaIlS5aobdu2mjRpkj777DPNmzdPSUlJatSokZo0aVLu65qZmal9+/bpz3/+s4KCgtS8eXMVFBTowQcf1Pfff6/HHntMbdq00Y4dO/T2228rLy9PEyZMUHBwsCIjI7VixYoS4e3zzz9X9+7dy1yfs2fP6v7779fFixc1ZswYXXvttVq/fr1efvllXbp0SUOHDlXDhg3VunVrZWRkOI8ny8jIUK1atZSZmelc1pYtW5Sfn6+IiIhy17O4tWvXatSoUerRo4dee+01SdKCBQs0ePBgffbZZ2ratKlzbFZWll555RU98sgjatCggf7+979r7Nix+uKLL5yzyhs3btRTTz2luLg4TZgwQadPn9a0adN08eJFXXfddZKkW265RY899pjmzp3rPORAkoKDg11eu+uuu07PP/+88vPzNWPGDI0aNUpffPGFqlUr/au0rPfd4XDokUce0Z49ezR27FiFhIRo7dq1mj59unJzc/XEE0+4LGfKlCmKjY3VjBkzdOnSpTJfu2PHjumPf/yjgoKCNHHiRDVs2FD//ve/NWbMGM2ZM0e33XabpMI/Apo0aaLnnntOgYGBysrK0rx58/TII4+4hPlVq1Zp7Nix6tKliyZPnqwGDRrou+++09GjR12ed+vWrfrxxx/1+OOPq2bNmpo1a5ZGjhypNWvWqH79+uW/4SrcnXr33XfroYcekp+fn/O9jYyM1JAhQ1SzZk3t2rVLs2fPVm5urvMPicWLF+vee+/VwIEDde+990qS28/V/PnzFRoaqldeeUW5ubl69dVX9dRTT2nRokXOMS+++KK+/PJLjRkzRu3bt9emTZs8Ps5w586dGjp0qFq2bKmJEyeqcePGOnTokP7zn/84x1zuunn6GXnrrbc0b948Pfjgg4qKitLu3bs1atSoEjV78r1Y5ODBg5o6dapGjRqlFi1aKDAwUNWrV9e0adOUnZ3tMnbx4sW69tprFRMT49FrBy8zgN/ok08+MUJCQkr8FxMTY3zzzTcuYxMTE42ePXsa+fn5zr7Dhw8bN910kzFt2jSXsY899pgRGRlp7N+/34iJiTEeeOABw+FwOB/PyMgwQkJCjAceeMDl51asWGGEhIQY6enpzr6QkBDjrbfeqnAdt956q/Hkk0969DqMGTPG6Nq1q5GXl+fs+/nnn41u3boZo0ePdvYtWbLECAkJMbKystwu89ZbbzU6duxo5OTkuPQnJycbISEhxubNm1363377baNdu3bGyZMnDcP472vx/fffO8fs2bPHCAkJMVauXOnsGzJkiDFkyBBnOykpyWjfvr3x448/uiz/+eefN8LCwpyv25QpU4xbb73V+Xi/fv2M6dOnuzzna6+9ZkRHR5e7nllZWUZISIjxySefOPtuv/12Y9iwYS7jfv75ZyMsLMyYOnWqS+033XSTS60nT5402rZta8ydO9fZd++99xp33nmnyza0a9cuIyQkxGXdi7bngwcPlqgzJCTE6Nmzp3Hp0iVn3xdffGGEhIQYW7duLXcdS3vf16xZU2K9DcMwnnvuOaNdu3bGqVOnDMP477Y+atSocp+jyMSJE43w8HAjNzfXpX/48OFGv379yvy5/Px84+uvvzZCQkKM3bt3G4ZhGA6Hw7j11luNxMREw263l/mzt956q3HzzTcbNpvN2bdz504jJCTE+Oyzz8qt96233jJCQkKMf/zjH+WOczgcRn5+vvH2228bN998s0s9ISEhxsyZM0v8zLPPPuuyjRZta4MHD3YZt2DBAiMkJMQ4fvy4YRiG8f333xtt2rQx5s+f7zJuypQppb5nxQ0aNMiIjY01zp07V+64y1k3Tz4jNpvN6Ny5szFp0iSXce+///5v/l4cMmSI0aZNG2PPnj0lnjs0NNRISkpy9p06dcpo166dMW/ePI9eB3gfu2hx2ebMmaNly5Zp6dKlmjNnjm644QY98sgjzmOPzp07pz179qhPnz4uMx0tWrRQly5d9PXXX7ss75VXXlG1atU0cOBA2e12/fWvf5XFYinxvH369HFp9+7dW35+ftq+fXupdVa0Dk99/fXXuuWWW1xmKurWrau4uLjfvExJ6tSpkxo1auTSt379ejVv3lyhoaEqKChw/hcdHa38/Hzt2LFDktSzZ08FBAS47EZdsWKF6tWr55zFKc369evVqVMnXXvttS7Lj4mJkc1m04EDByRJEREROnLkiLKysnT69Gn95z//Uf/+/fWHP/xBGRkZkgpn9X49i+uJgwcP6vDhw+rbt6/L89eqVUuhoaEldvu1atVKf/jDH5ztoKAgBQUFOWeY7Ha7du3apfj4eJdtqF27drr22msrVFtUVJSqV6/ubIeEhEgqnDWrqK+//lp+fn5KSEhw6e/Xr5/L+1ikZ8+eHi13/fr16tGjh+rVq1fi/du3b5/Onj0rqfC4qXfeeUe9e/dWx44d1a5dOw0ePFiS9OOPP0qSfvjhBx05ckT33HOPc1atLJ07d1ZgYKCzXbR70dPXprT1y8nJ0UsvvaRbb71V7du3V7t27fTmm2/qzJkzOnXqlEfLLU3x44KLv487d+6UYRjq3bu3y7hevXq5Xfb58+e1bds29e3bV7Vr1y5z3OWsm6efkf379+vcuXNu16Oi34vNmzfXjTfe6NJXt25d9evXT8uWLXMeKlK0K3vgwIHlrg8qD7tocdlat27t3B0mSTExMerRo4dmz57t/NIyDMNld1eRa665RkeOHHHpa9CggW655RYtXrxYd955p6655ppSn7d4f40aNVS/fn1lZ2eXOr6idXgqLy+vRBArWmZpxzR5qrRl5ubm6siRI2rXrl2pP1N0iY/atWurV69e+vzzzzVu3Dg5HA7ncWg1a9Ys8zlzc3N16NAht8sPCwuTn5+fMjMzVa9ePQUGBqpt27YKDw9XZmam+vXrpz179uiPf/xjhda56Jfb888/r+eff77E482aNXNp/zpUFKlRo4ZzN+bp06eVn5+voKCgEuPK2q7KUvy5ig4sv3jxYoWWIxVuM4GBgSUOTi+qqfh2U9q2UJrc3FwtX75cy5cvL/Xx06dPq27dupo5c6YWLVqkUaNGKTQ0VHXq1FF2drb+/Oc/O9en6L12t8tTuvzXpvj6ORwOPfbYY8rJydGYMWN0/fXXq2bNmlq1apXeeeed3/SaFyl+hnTxWnNyciSpxDZT2jZU3JkzZ+RwOMp9zS533Tz9jJw4caLUuotv9xX9XixrWxw0aJA++ugjrVu3TrfccouWLFmi22+/vcKfM3gPAQ9eV6tWLbVo0cJ5zEn9+vVlsVicXzi/dvLkyRJfuOnp6VqyZInat2+vjz76SP369VOHDh1K/dlfu3Tpks6cOeNyDMivVbQOTwUGBpaopWiZpQUQT5U2a2m1WnXttdfqzTffLPVnmjdv7vx3//79lZycrK1bt+rChQs6ceKE22uwWa1WNWzYsNRfHJKcx6zVr19fN954ozIyMlSvXj2FhYXJYrEoIiJCU6ZMUWZmpux2e4Vn8IregyeffLLUS0P8egbNEw0aNFD16tVLnRU5efJkicB4pQQGBiovL6/EGYhF21HxbbG0baE0VqtVXbt21cMPP1zq40WfjZUrV6p///4ux2MVzbwWadCggSSV+QeTNxVfv8OHD2vXrl2aMWOGyzb71VdfVXotRUHn1KlTCggIcPZ7MmtYv359+fn5lfuaXe66efoZKQpip06dUuvWrZ2PF/+uquj3YlnbYkhIiG6++WYtXrxYNWvW1KFDh/SXv/zFo3VC5WAXLbzu/PnzysrKch7IHxAQoHbt2unLL790OUvsyJEj2r59u8LCwpx9ubm5euaZZ9SjRw99/PHHuvHGG/XUU0/pl19+KfE8X3zxhUv7yy+/lMPhUGhoaKl1VaSO6tWrezxL0K1bN61bt865+0sqPFnhq6++clmmN3Tv3l3Hjx9XQECAOnToUOK/X588ER4eriZNmmjFihVasWKFmjdvrptvvtnt8n/88Uc1a9as1OUXnSFdtPzMzExlZmY6T6SIiIjQ6dOn9cEHH6hp06YuM7ueuP7669W8eXN99913pT5/27ZtK7Q8f39/tW/fXqmpqS5np+7atavEGbJFQevChQsVeo7fIiwsTA6HQ19++aVL/+eff67q1aurU6dOv2m53bt313/+8x+1bt261Nfv1+tY/MSQ4meHXnfddWrevLmWLl3q8tpdCUXvwa8DfX5+vj7//PMSYyvyWfVEx44dZbFYSrw3xdulqV27trp27arPPvuszO3octfN089ISEiIAgIC3K5HRb4X3Rk0aJDS0tI0e/Zs/eEPf+D6fSZjBg+Xbe/evTp9+rQMw9CJEye0aNEi2Ww2DRkyxDnm8ccf16OPPqpHH31UgwYN0rlz5zR79mzVrVtXI0aMcI577rnnJEnTp09X9erV9be//U0DBgzQ1KlTNX36dJfnPXDggCZOnKg77rhDBw8e1BtvvKGwsLByv1Q8reOGG27Qli1b9NVXX+maa65RgwYNyjxma9SoUVq7dq2GDx+uhx9+WBaLRe+++67Onz9f6qUDLkffvn316aefavjw4XrggQfUtm1bXbp0SVlZWVqzZo3mzJnjPPbHz89Pffv21eLFi1VQUKA//elPbmeChg8frn//+98aNGiQhg8fruuuu07nz5/XDz/8oC1btmju3LnOsREREXr//feVk5PjnKkrOsN206ZNGjBgQIXXz2KxaNKkSRo1apTy8/PVp08fNWjQQCdPntT27dvVrFkzl/fJE2PGjNEDDzyg0aNH695779Xp06c1e/ZsNWrUyOX1KLom4IcffqjExERVq1ZNbdq0qZRrfMXGxqpr166aNGmScnNz1bp1a61bt05Lly7Vo48+Wu5Z2+UZO3as7rnnHg0ePFhDhgxR8+bNdebMGe3fv19ZWVnOz1D37t21fPlyhYSEqFWrVkpNTS1x7KrFYtFzzz2nMWPGaNiwYbr//vvVoEED/fDDDzp16pTGjh172a9DWYpCzBtvvCE/Pz9Vq1ZN//znP0sde8MNN2jt2rXq3r276tevr+Dg4DJn8T197oSEBM2aNUuGYahdu3bKyMhwzrC5Ox7xmWee0dChQ3XvvfdqxIgRatKkibKysrRv3z69+OKLXlk3Tz4jgYGBGjZsmObNm6c6deo4z6L95JNPJLnOxHn6vehOfHy8pk2bpm3btpV6aakNGzbokUce0euvv6477rjD4+XityHg4bIVXbdK+u8v+AULFqh79+7O/tjYWM2bN09z5szRuHHjVL16dYWFhenpp592fhkvWrRIa9eu1Xvvvef8BdeyZUtNmjRJzzzzjLp37+7ypfD8889rzZo1Gj9+vOx2u+Li4srctViROiTpiSee0Isvvqhx48bpwoULSkxM1KuvvlrqMtu2basPPvhAb7zxhiZMmCDDMNSpUyctWrSowjNO7lSvXl3vvfee5s+fr8WLF+unn35SQECAWrRooVtuuaXELsz+/fvr3XfflVR4AL879erV08cff6w5c+bo3XffVU5OjurVq6frrruuxF09unbtqmrVqslqtbpcMDk8PFz79++v8O7ZIj169NCiRYv0zjvv6IUXXtCFCxfUqFEjderU6Tf9UoiOjtbrr7+upKQkjR49Wq1atdKECRM0Z84clzuttG3bVmPGjNHixYu1dOlSORwOrV69usInY3jCz89P8+fP18yZM7VgwQLZbDY1b95cEydO1J/+9KffvNxmzZrpk08+0ezZszVz5kydPn1aVqtVrVu3dgncL7zwggzDcO7qj42N1d/+9jfdc889Lsu7/fbb9f777+vtt992frZatGhxWTV6okaNGpozZ44mT56sZ599VoGBgbrrrrvUrFkzvfDCCy5jX3zxRb3yyisaOXKkLl26pD//+c/Oyxv9VlOmTFGdOnW0YMEC5efnKzw8XC+99JIeffRRl1ns0nTs2FEfffSR3nrrLU2dOlWXLl1Ss2bNnCcbeGPdPP2MFIXwZcuW6YMPPlCnTp00ffp03X///S7bvqffi+5Ur15dt912m5KTk0v9A8/hcMhut1/xGeGrlcXglcbvTGZmpoYNG6a///3vbi/6CpTm+PHj6tmzp0aOHOn1WVb4pgULFuj111/XmjVrTDt20xu++OILjRs3Th9++KHbQzYqqqCgQPHx8eratavz+nwwDzN4AHzahQsXNH36dEVFRalBgwbKysrSggULVLt27RIzVoBUeMLDd999p7Zt28rPz09btmzR+++/rz59+vyuwt0333yjtWvXqlOnTs4LKr/77rvq3Lmzunbt6rXnOXv2rPbv36+UlBQdO3ZMDzzwgNeWjd+OgAfAp/n5+enkyZOaMmWKbDab80D4WbNmlXppCKBOnTpatWqV5s+fr/Pnz6tx48YaOnRopR53WBkCAgK0ZcsW/etf/9LZs2fVsGFD9e7dW08++aTHZ2Z7Yvfu3Ro2bJiCgoL0/PPPl7hOHszBLloAAAAfw2VSAAAAfAwBDwAAwMdc9cfgFZ62zV5qAABQ9VWv7u/RuKs+4Nnthmy2c2aXAQAA4FajRvXcDxK7aAEAAHwOAQ8AAMDHEPAAAAB8DAEPAADAxxDwAAAAfAwBDwAAwMcQ8AAAAHwMAQ8AAMDHEPAAAAB8DAEPAADAxxDwAAAAfAwBDwAAwMcQ8AAAAHwMAQ8AAMDHEPAAADCJzWbTjBlTlZdnM7sU+BgCHgAAJklJSdaBA/uVkrLc7FLgYwh4AACYwGazKT19vQzD0MaNacziwasIeAAAmCAlJVkOhyFJcjgMZvHgVQQ8AABMkJm5SXZ7gSTJbi9QRka6yRXBlxDwAAAwQXh4pCTL/7UsioiIMrMc+BgCHgAAJoiNjZNk/F/L+L824B0EPAAATJCWtkYWS+EMnsViUVraGpMrgi8h4AEAYILMzE0yjMIZPMMwOAYPXkXAAwDABOHhkfL3ryZJ8vevxjF48CoCHgAAJkhISJSfX+EuWj8/ixISBphcEXwJAQ8AABNYrVZFRXWXxWJRdHSsAgOtZpcEH1LN7AIAALhaJSQk6ujRI8zewessRtERnlep/Hy7bLZzZpcBAADgVqNG9Twaxy5aAAAAH0PAAwAA8DEEPAAAAB9DwAMAAPAxBDwAAAAfQ8ADAADwMQQ8AAAAH0PAAwAA8DEEPAAAAB9DwAMAAPAxBDwAAAAfQ8ADAADwMQQ8wAM2m00zZkxVXp7N7FIAAHCLgAd4ICUlWQcO7FdKynKzSwEAwC0CHuCGzWZTevp6GYahjRvTmMUDAFR5BDzAjZSUZDkchiTJ4TCYxQMAVHlXLOAdO3ZMQ4cOVZ8+fXTnnXfqn//8p6TC2ZERI0YoPj5eI0aMUF5eniTJMAxNnTpVPXv2VN++fbV7927nspKTkxUfH6/4+HglJyc7+3ft2qW+ffuqZ8+emjp1qgzDuFKrBx+WmblJdnuBJMluL1BGRrrJFQEAUL4rFvD8/f01YcIEffHFF1q8eLH+9a9/6cCBA5o/f74iIyOVmpqqyMhIzZ8/X5KUlpamgwcPKjU1VVOmTNHLL78sqTAQJiUlacmSJVq6dKmSkpKcofDll1/W5MmTlZqaqoMHDyotLe1KrR58WHh4pPz9q0mS/P2rKSIiyuSKAAAo3xULeMHBwWrXrp0kqW7durr++uuVnZ2t1atXa8CAAZKkAQMGaNWqVZLk7LdYLOrcubPOnDmjnJwcbdiwQdHR0bJarQoMDFR0dLTWr1+vnJwcnT17VqGhobJYLBowYIBWr159pVYPPiwhIVF+fhZJkp+fRQkJA0yuCACA8plyDN5PP/2kvXv3qlOnTjp16pSCg4MlFYbA3NxcSVJ2draaNGni/JkmTZooOzu7RH/jxo1L7S8aD1wuq9WqqKjuslgsio6OVWCg1eySAAAoV7Ur/YS//PKLxo4dq+eee05169Ytc1xpx89ZLJYK97vj72+R1RrgdhyuboMHD1JOznENGnQ/2wsAoMq7ogEvPz9fY8eOVd++fRUfHy9JCgoKUk5OjoKDg5WTk6OGDRtKKpyBO378uPNnjx8/ruDgYDVp0kSbN2929mdnZyssLKzM8e7Y7YZstnPeWkX4KIulpp54YqIksb0AAEzTqFE9j8ZdsV20hmHo+eef1/XXX68RI0Y4++Pi4rR8eeFlJ5YvX67bbrvNpd8wDO3YsUP16tVTcHCwYmJitGHDBuXl5SkvL08bNmxQTEyMgoODVadOHe3YsUOGYbgsCwAA4GpiMa7QtUS2bNmiwYMHKyQkRH5+hbnyiSeeUMeOHTVu3DgdO3ZMTZs21axZs2S1WmUYhiZPnqz169erdu3amjZtmjp06CBJWrZsmebNmydJGjlypO666y5J0rfffquJEyfqwoULio2N1Ysvvuh2N21+vp0ZGQAA8Lvg6QzeFQt4VRUBDwAA/F5UuV20AAAAuDIIeAAAAD6GgAcAAOBjCHgAAAA+hoAHAIBJbDabZsyYqrw8m9mlwMcQ8AAAMElKSrIOHNivlJTlZpcCH0PAAwDABDabTenp62UYhjZuTGMWD15FwAMAwAQpKclyOAovRetwGMziwasIeAAAmCAzc5Ps9gJJkt1eoIyMdJMrgi8h4AEAYILw8Ej5+1eTJPn7V1NERJTJFcGXEPAAADBBQkKi/PwK75fu52dRQsIAkyuCLyHgAQBgAqvVqqio7rJYLIqOjlVgoNXskuBDCHgAAJgkNjZONWvWUmxsnNmlwMcQ8AAAMEla2hpdvHhBaWlrzC4FPoaABwCACbgOHioTAQ8AABNwHTxUJgIeAAAm4Dp4qEwEPAAATMB18FCZCHgAAJiA6+ChMhHwAAAwAdfBQ2WqZnYBAABcrRISEnX06BFm7+B1FsMwDLOLMFN+vl022zmzywAAAHCrUaN6Ho1jFy0AAICPIeABAAD4GAIeAACAjyHgAQAA+BgCHgCBa1bRAAAgAElEQVQAgI8h4AEAAPgYAh7gAZvNphkzpiovz2Z2KQAAuEXAAzyQkpKsAwf2KyVludmlAADgFgEPcMNmsyk9fb0Mw9DGjWnM4gEAqjwCHuBGSkqyHI7CG744HAazeACAKo+AB7iRmblJdnuBJMluL1BGRrrJFQEAUD4CHuBGeHik/P2rSZL8/aspIiLK5IoAACgfAQ9wIyEhURZL4b8tFikhYYC5BQEA4AYBD3DDarWqUaNgSVJwcGMFBlpNrgiArzh8+JDGjHlEWVmHzS4FPoaAB7hhs9l04sQJSdKJEzmcRQvAaxYseFsXLpzXggVvm10KfAwBD3AjJSVZhsFZtAC86/DhQzp27Kgk6ejRI8ziwasIeIAbnEULoDIUn7VjFg/eRMAD3OAsWgCVoWj2rsjRo0dMqgS+iIAHuJGQkCg/v8LTaP38LJxFC8Arrrkm2KVddDIX4A0EPMANq9WqqKjuslgsio6O5SxaAF5RdGxvWW3gclQzuwDg9yAhIVFHjx5h9g6A15w6dcKlffLkiTJGAhVHwAM8YLVa9cwzL5hdBgAfEhAQoHPnzv2qXcfEauBr2EULAIAJCgrsxdoFJlUCX0TAAwDABJGR0S7tqKgYkyqBLyLgAQBggoSERPn7+0uS/P39OcYXXkXAAwDABFar6xn5nKEPbyLgAQBggt27d8luLzwOz263a+/e3SZXBF9CwAMAwATz5r3l0p47d5ZJlcAXEfAAADDB+fPny20Dl4OABwAA4GMIeAAAmKBDh84u7U6dQk2qBL6IgAcAgAkGDLjbpd2//91ljAQqjoAHAIAJVq36slj7C5MqgS8i4AEAYILNmze5tDMzN5UxEqg4Ah4AAICPIeABAGCCbt0iXNphYZEmVQJfRMADAMAEERExLu3IyJgyRgIVR8ADAMAEixd/4NL++OMPyhgJVBwBDwAAExw7dtSlffToEZMqgS8i4AEAYIKAgIBi7TomVQJfRMADAMAEBQX2Yu0CkyqBLyLgAQBggsjIaJd2VBQnWcB7CHgAAJggISGxWHuASZXAFxHwAAAAfAwBDwAAE3zwwXvltoHLQcADAMAEO3fucGl/882OMkYCFUfAAwAA8DEEPMADNptNM2ZMVV6ezexSAPiImjVrldsGLgcBD/BASkqyDhzYr5SU5WaXAsBHtGnT1qXdtm3bMkYCFUfAA9yw2WxKT18vwzC0cWMas3gAvGLfvr0u7b1795YxEqg4Ah7gRkpKshwOQ5LkcBjM4gHwimrV/Iu1q5lUCXwRAQ9wIzNzk+z2wlsI2e0FyshIN7kiAL7g3Llzxdq/mFQJfBEBD3AjPDxS/v6Ff1n7+1dTRESUyRUB8AX+/v7F2szgwXsIeIAbCQmJ8vOzSJL8/CzcTgiAV9jt9mLtApMqgS8i4AFuWK1WRUV1l8ViUXR0rAIDrWaXBMAHNG3azKXdrFlzkyqBLyLgAR5ISEjUDTeEMHsHwGseemhUuW3gchDwAA9YrVY988wLzN4B8JqWLVs5Z/GaNWuuFi1amlwRfAkBDwAAkzz00CjVqlWb2Tt4ncUwDMPsIsyUn2+XzXbO/UBc1Ww2m+bPT9Kjj/6ZWTwAgGkaNarn0Thm8AAPcKsyAMDvyRULeBMnTlRkZKQSEhKcfbNnz1b37t3Vv39/9e/fX+vWrXM+Nm/ePPXs2VO9evXS+vXrnf1paWnq1auXevbsqfnz5zv7s7KydM899yg+Pl7jxo3TpUuXrsyKwedxqzIAlcVms2nGjKl8r8DrrljAGzhwoBYsWFCif/jw4VqxYoVWrFihHj16SJIOHDiglStXauXKlVqwYIH+8pe/yG63y263a/LkyVqwYIFWrlyplJQUHThwQJL0+uuva/jw4UpNTVX9+vW1bNmyK7Vq8HHcqgxAZfn44w/03Xf/0ccfLzK7FPiYKxbwunXrpsDAQI/Grl69Wnfeeadq1KihFi1aqFWrVtq5c6d27typVq1aqUWLFqpRo4buvPNOrV69WoZhKCMjQ7169ZIkJSYmavXq1ZW5OriKcKsyAJXBZrNp69bNkqQtWzKZxYNXmX4M3ocffqi+fftq4sSJysvLkyRlZ2erSZMmzjGNGzdWdnZ2mf2nT59W/fr1nTdqbtKkibKzs6/sisBncasyAJXh448/KNZmFg/eY+qN7+6//36NGjVKFotFs2bN0quvvqrp06ertBN7LRaLHA5Hqf2lKau/OH9/i6zWgIoVjqvK4MGDlJ6+QXa75O/vp0GD7mebAXDZtm372qW9detmvlvgNaYGvGuuucb573vuuUcjR46UVDgDd/z4cedj2dnZCg4OlqRS+xs0aKAzZ86ooKBA1apV0/Hjx53j3bHbDS6TgnJZLDUVFRWjtLSv/u+WZTXZZgBctuKTGYbB7yO497u4TEpOTo7z36tWrVLr1q0lSXFxcVq5cqUuXbqkrKwsHTx4UB07dlSHDh108OBBZWVl6dKlS1q5cqXi4uJksVgUHh6u//3f/5UkJScnKy4uzpR1gm/iVmUAvM3Pz69Y29+kSuCLrtgM3hNPPKHNmzfr9OnTio2N1ZgxY7R582bt27dPktS8eXNNnjxZktS6dWv16dNHd9xxh/z9/fXSSy/J379ww3/ppZf00EMPyW6366677nKGwqefflrjx4/Xm2++qRtvvFH33HPPlVo1XAWKblUGAN5S/LAjh8NuUiXwRdzJgjtZAABM8PzzTysn57+HHTVu3ERTp75mYkX4Pfhd7KIFAOBq1aJFy3LbwOUg4AEAYIJdu3a6tL/9dmcZI4GKI+ABAGCChg0burSDgoJMqgS+iIAHAIAJTp065dI+efKkSZXAFxHwAAAwQfEZu19fGxa4XAQ8AABMkJub69IuPqMHXA4CHgAAJih+X+vIyGiTKoEvIuABAGCC2Ni4ctvA5SDgAQBggn//+7Ni7RUmVQJfRMADAMAEW7dudmlv2bK5jJFAxRHwAAAAfAwBDwAAE1gslnLbwOUg4AEAYIIuXbq5tLt2DTOpEvgiAh4AACa4776hxdpDTKoEvoiABwCACc6cySvWPmNSJfBFBDzAAzabTTNmTFVens3sUgD4iAUL3i63DVwOAh7ggZSUZB04sF8pKcvNLgWAjzh27KhL++jRIyZVAl9EwAPcsNlsSk9fL8MwtHFjGrN4ALyiadNmLu1mzZqbVAl8EQEPcCMlJVkOhyFJcjgMZvEAeMVDD40qtw1cDgIe4EZm5ibZ7QWSJLu9QBkZ6SZXBMAXtGzZSv7+/pIkf/9qatGipckVwZcQ8AA3wsMj5e9fTVLhl3BERJTJFQHwBYcPH5LdbpdU+MdjVtZhkyuCLyHgAW4kJCTKz6/wCvN+fhYlJAwwuSIAvoCzaFGZqnk68JtvvtGmTZt06tQpGYbh8tgLL7zg9cKAqsJqtSoqqrvS0r5SdHSsAgOtZpcEwAdwFi0qk0cB77333tNrr72mVq1aKTg42OUx7p2Hq0FCQqKOHj3C7B0ArwkICNC5c+d+1a5jYjXwNR4FvIULF+qFF17QkCHcRgVXJ6vVqmeeYaYagPcUFBQUa+ebVAl8kUfH4J09e1Y9evSo7FoAALhqWK0NXNoNGjQ0qRL4Io8C3p133qm0tLTKrgWosrhVGQBvO3nypEv7xIkTJlUCX1TmLtq///3vzn83bdpUs2fP1rZt29SmTRtVr17dZeyIESMqr0KgCvj1rcoGDx5udjkAfEDxQ9g5pB3eVGbA++CDD1zaAQEB2r59u7Zv3+7Sb7FYCHjwacVvVZaQMIAzaQFctrCwSG3atMHZDg+PNLEa+JoyA96aNWuuZB1AlVXarcqYxQNwuW6/vbdLwLv99j4mVgNf49ExeElJSTp//nyJ/gsXLigpKcnrRQFVCbcqA1AZ0tLWlNsGLodHAW/OnDku1+opcv78ec2ZM8frRQFVCbcqA1AZMjM3ubT54xHe5FHAMwyj1Asa7927V4GBgV4vCqhKuFUZgMrQuXMXl3ZoaFeTKoEvKvdCx6GhobJYLLJYLLrttttcQp7D4dDFixd13333VXqRgJm4VRmAypCfn19uG7gc5Qa8l156SYZh6LnnntP48eNVr14952PVq1dX8+bNFRoaWulFAmbjVmUAvG3Hjq0u7e3bt5hUCXxRuQEvMTFRknTttdcqNDS0xPXvgKsFtyoD4G1FZ+eX1QYuh0f3or322mtLvcK2xWJRzZo11bAht1cBAKAiatasqQsX/nuFilq1appYDXyNRwEvLi6u1JMsitStW1cDBw7U008/rWrVPFokAABXtV+HO0mlXo4M+K08SmMzZ87Ua6+9pvvuu08dO3aUJO3cuVOLFy/WmDFjdObMGc2dO1d16tTR2LFjK7VgAAB8QXBwE+XkHHe2GzduYmI18DUeBbyPPvpIEydOVHx8vLMvMjJS1113nRYuXKhFixYpKChIs2fPJuABAOCBFi1augS8Fi1amlgNfI1H18H75ptvFBISUqK/devW+vbbbyVJnTt31vHjx0uMAQAAJe3atdOl/e23O8sYCVScRwGvWbNmWrx4cYn+JUuWqGnTppKk3NxcLnoMAICHrFZruW3gcni0i3bChAkaM2aM0tLS1KFDB1ksFn377bc6fPiwZs+eLUn69ttv1acPN0oGAMATJ0+eKLcNXA6PAt4tt9yi1NRU/etf/9KPP/4owzAUFxen++67T82aNZMkDR48uFILBQAAgGc8vqZJ06ZN9eSTT1ZmLQAAXDW6dYtQRsZGZzssLNLEauBrPA5458+f1969e5WbmyuHw+Hy2K/PrgUAAO7dddd9LgHvrrvuNbEa+BqPAl56erqeeOIJ2Wy2Eo9ZLBbt3bvX64UBAADgt7EYhuH25nd33nmnOnTooPHjx6tx48ZXoq4rJj/fLpvtnNllAACuMnPnvqVt2752trt2DdPIkWNMrAi/B40a1fNonEczeEeOHNHcuXN9LtwBAGCWX4c7Sdq6dbNJlcAXeXQdvC5duujHH3+s7FoAAADgBR7N4N13333661//qpycHIWEhKhaNdcfa9euXaUUBwAAgIrzKOAV3V/2xRdfLPEYJ1kAAABULR4FvNWrV1d2HQAAAPASjwJe8+bNK7sOAAAAeIlHJ1lI0rp16/Too4/qjjvu0LFjxyRJS5cu1aZNmyqtOAAAAFScRwHvs88+07hx49SqVSv99NNPKigokCTZ7XYtWLCgUgsEqgKbzaYZM6YqL6/kxb4BAKhqPAp4CxYs0NSpU/Xcc8/J39/f2d+5c2dOsMBVISUlWQcO7FdKynKzSwEAwC2PAt6hQ4fUuXPnEv0BAQE6e/as14sCqhKbzab09PUyDEMbN6YxiwcAqPI8CnjBwcE6ePBgif6vv/5aLVu29HZNQJWSkpIsu90hSbLbHcziAQCqPI8C3h//+EdNnTpVW7dulSQdO3ZMycnJeu2113T//fdXaoGA2TIzN8nhsEuSHA67MjLSTa4IAIDyeXSZlIcfflhnz57VAw88oIsXL2rYsGGqUaOGHnjgAQ0ePLiyawRM1blzF2VkbHS2Q0O7mlgNAADueRTwJGn8+PEaOXKkDhw4IMMw9D//8z+qU6dOZdYGVAkWi6VY26RCAADwkMfXwZOk2rVrq0OHDurYsaPq1Kmj7777TpGRkZVVG1AlbN++1aW9bdvWMkYCgOdK/vHIX4/wngoFvOLsdrtsNs4ohG8LD490Xh7I399fERFRJlcEwBcQ8FCZLivgAVeDhIRE+fkVflT8/PyUkDDA5IoA+IJrrmnk0m7UKNikSuCLCHiAG1arVVFR3WWxWBQdHavAQKvZJQHwAcX3gJ0+fdqkSuCLCHiABxISEnXDDSHM3gHwmqCgIJf2NddcY1Il8EXlnkUbGhpa7jEBDofD6wUBVZHVatUzz7xgdhkAfEhubq5L+9SpUyZVAl9UbsB76aWXrlQdQJVms9k0f36SHn30z+yiBeAVERFRWrdujbMdGRltYjXwNeUGvMTExCtVB1ClpaQk68CB/UpJWa7Bg4ebXQ4AHxAa2s0l4HXp0s3EauBrOAYPcMNmsyk9fb0Mw9DGjWnKy+PSQAAu36JF77u0Fy58v4yRQMUR8AA3UlKS5XAYkiSHw1BKynKTKwLgC06ePFGsnWNSJfBFBDzAjczMTbLbCyRJdnuBMjLSTa4IAIDyEfAANwrvZFF4uKq/fzXuZAEAqPI8Cnh2u72y6wCqrMI7WRReLsjPz8K18AAAVZ5HAe/mm2/Wgw8+qHnz5mn79u0EPlxVuJMFAOD3ptzLpBRJSkrS5s2btXbtWs2ePVvVq1dXly5dFBYWprCwMIWGhlZ2nYCpEhISdfToEWbvAHhNhw6d9e23O5ztTp34XQrv8SjgRUdHKzq68AKMFy5c0LZt2/T555/rrbfeksPh0N69eyu1SMBs3MkCgPcZri3DKGMcUHEeBTxJOnnypDZv3qzMzExlZGTo2LFj6tKli8LDwyuzPgAAfNK3337j0t65c0cZI4GK8yjg3XnnnTpy5Ig6dOig8PBwTZ48WaGhoapRo0Zl1wcAAIAK8ugki59//ll+fn6qXbu2ateurTp16qh69eqVXRsAAAB+A49m8NLS0nTo0CHnLtqFCxfq3Llz6tq1qyIiIjR8+PBKLhMAAN9So0ZNXbp00dmuWbOmidXA11iMCh7VabfbtXPnTi1ZskSfffbZ7/4ki/x8u2y2c2aXAQC4ysyc+Vft3bvL2b7ppvYaP/5ZEyvC70GjRvU8GufRLtqdO3fq3Xff1UMPPaRu3bpp6NCh+v777zVixAjNnz/foyeaOHGiIiMjlZCQ4Oyz2WwaMWKE4uPjNWLECOXl5UkqPJNo6tSp6tmzp/r27avdu3c7fyY5OVnx8fGKj49XcnKys3/Xrl3q27evevbsqalTp3I2ErzKZrNpxoypysuzmV0KAB+xb99ul/bevbvLGAlUnEcBb9CgQVq1apXatm2rN998U5s3b9aSJUv01FNPqXv37h490cCBA7VgwQKXvvnz5ysyMlKpqamKjIx0hsW0tDQdPHhQqampmjJlil5++WVJhb9kk5KStGTJEi1dulRJSUnOUPjyyy9r8uTJSk1N1cGDB5WWlubpawC4lZKSrAMH9islZbnZpQDwEcUnIpiYgDd5FPA2b96sxYsX66mnnlJsbKwCAgIq/ETdunVTYGCgS9/q1as1YEDhhWMHDBigVatWufRbLBZ17txZZ86cUU5OjjZs2KDo6GhZrVYFBgYqOjpa69evV05Ojs6ePavQ0FBZLBYNGDBAq1evrnCNQGlsNps2blwvwzC0cWMas3gAgCrPo5MsAgICdPHiRX311Vc6fPiw7rvvPtWvX1+HDx9W/fr1ZbX+tls3nTp1SsHBwZKk4OBg5ebmSpKys7PVpEkT57gmTZooOzu7RH/jxo1L7S8a7wl/f4us1ooHVlw9li5dJLu9QJJUUFCg//f/UvTQQ4+YXBWA3zs/P385HP+99ae/vz+/j+A1HgW8Q4cOacSIEfrll1/0888/q0+fPqpfv74++ugjnTlzRq+88opXiyptmtpisVS43xN2u8FJFijXunXrnNuYYRhau3ad7r57iMlVAfi9+3W4kwpPYuT3Edzx6kkW06ZNU3R0tNLT011O446Li1NmZuZvq1BSUFCQcnJyJEk5OTlq2LChpMIZuOPHjzvHHT9+XMHBwSX6s7OzS+0vGg94Q9F2WSQoKMikSgAA8IxHAW/79u164IEH5O/v79LftGlTZ0D7LeLi4rR8eeFB68uXL9dtt93m0m8Yhnbs2KF69eopODhYMTEx2rBhg/Ly8pSXl6cNGzYoJiZGwcHBqlOnjnbs2CHDMFyWBVyuokMHipw6dcqkSgAA8IzH96ItKCgo0Xfs2DHVq+fZVOETTzyhzZs36/Tp04qNjdWYMWP0yCOPaNy4cVq2bJmaNm2qWbNmSZJ69OihdevWqWfPnqpdu7amTZsmqfCG76NGjdLdd98tSRo9erTz+L+XX35ZEydO1IULFxQbG6vY2FhPVw0oV0RElNat+0qFNwa3KDIy2uySAAAol0cXOh4/frwzaIWGhuqzzz5TgwYN9Nhjj6lFixbOAPZ7xIWO4Y7NZtPEieNVUFCgatWq6dVX31Bg4G87sQgAijz88NASfe+++4EJleD3xKvH4E2YMEFbt25Vr169dOnSJY0fP15xcXE6efKknnrqqcsqFKjqrFaroqNjZbFYFBPTg3AHAKjyPNpF27hxY61YsUIpKSnas2ePHA6H7r33XvXt21e1atWq7BoB0yUkJOro0SNKSBhgdikAALjl8TF4tWrVch77BlxtrFarnnnmBbPLAADAI2UGvNTUVN16662qXr26UlNTy11IfHy81wsDAADAb1NmwBs7dqw2btyooKAgjR07tswFWCwW7d27t1KKAwAAQMWVGfD27dtX6r8BAABQtXl0Fm3xC70CAIDLU7t27XLbwOXwKODFxsZq5MiR+ve//62LFy9Wdk1AlWOz2TRjxlTl5dnMLgWAj3A4HOW2gcvhUcB75513ZLVa9eKLLyoqKkoTJ07Upk2b5ME1kgGfkJKSrAMH9islZbnZpQDwEcUnTJhAgTd5FPBiYmL06quvKj09XVOmTJHNZtPDDz+sHj16aMaMGZVdI2Aqm82m9PT1MgxDGzemMYsHAKjyPAp4RWrWrKk77rhDc+fO1YoVK9SwYUP9/e9/r6zagCohJSVZDkfhbLXDYTCLBwCo8ioU8M6dO6cVK1booYceUv/+/fXLL7/oscceq6zagCohM3OT7PYCSZLdXqCMjHSTKwIAoHwe3cli7dq1+vzzz7VmzRrVrFlTvXv31sKFC9WlS5fKrg8wXXh4pNatW+NsR0REmVgNAADueRTwHn/8cd1yyy16/fXX1aNHD1Wr5vEdzoDfvdDQbi4Br0uXbiZWAwCAex4ltY0bN6pu3bqVXQtQJS1e/IFL++OPP9Bf/vKqSdUAAOCeR8fg1a1bVydPntR7772nSZMmOS98vHXrVmVlZVVqgYDZjh076tI+evSISZUAAOAZjwLerl271Lt3b33++ef65JNP9Msvv0iS0tPT9eabb1ZqgYDZmjZt5tJu1qy5SZUAAOAZjwLeX//6Vw0bNkzLly9X9erVnf0xMTHatm1bpRUHVAUPPTSq3DYAAFWNRwFv9+7dSkxMLNHfqFEjnTx50utFAVVJy5atFBAQIEkKCKijFi1amlwRAADl8yjg1apVS3l5eSX6f/jhBwUFBXm9KKAqsdlszlsIXbx4gTtZAACqPI8C3m233aakpCRdunTJ2ffTTz/p9ddfV3x8fKUVB1QFKSnJkiz/17JwJwsAQJXnUcB79tlnlZeXp4iICF24cEGDBg1SfHy86tevr3HjxlV2jYCpuJMFAOD3xqPr4NWtW1cfffSRNm3apD179sjhcKhdu3aKiuKK/vB94eGRWr9+nRwOu/z8/LmTBQCgyqvQLSkiIyMVGRnp0nfs2DE1bdrUq0UBVUlCQqLS0r6SJBmGQwkJA0yuCACA8nm0i7Y0J06c0F/+8hf16tXLm/UAVZLFYnH5PwAAVVm5Ae/MmTN68sknFRERoZiYGC1cuFCGYSgpKUm33367du7cqWnTpl2pWgFTpKQkuwQ8TrIAAFR15e6inTlzprZs2aLExEStX79e06dPV3p6us6dO6d3331XYWFhV6pOwDSFJ1nYJUl2u10ZGekaPHi4uUUBAFCOcmfw1q1bp+nTp+vZZ5/V3LlzZRiGWrRooYULFxLucNUID3c97pSTLAAAVV25AS8nJ0f/8z//I0lq0aKFatasqT/+8Y9XpDCgqggN7ebS7tKlWxkjAQCoGsoNeA6Hw+Xes35+fqpVq1alFwVUJR999E+X9ocf/sOMMgAA8Fi5x+AZhqGnn37aGfIuXbqkF198sUTIe+eddyqvQsBk2dnHy20DAFDVlBvwEhMTXdr9+vWr1GIAAABw+coNeNOnT79SdQBVmEWS8d8W18IDAFRxv/lCx8DVws/P9WNisfCxAQBUbfymAtwIC4twaRe/bAoAAFUNAQ9wo2fPPuW2AQCoagh4gBvLly9zaa9YsayMkQAAVA0EPMCNb7/d4dL+5pvtJlUCAIBnCHgAAAA+hoAHuFG7du1y2wAAVDUEPMCNgQPvdWnfffd9JlUCAIBnCHiAG2vWpLq0V69OLWMkAABVAwEPcOPYsaMu7aNHj5hUCQAAniHgAW4EBAQUa9cxqRIAADxDwAPcKCiwF2sXmFQJAACeIeABbkRGRru0o6JiTKoEAADPEPAANxISEou1B5hUCQAAniHgAW6cOZNXrH3GpEoAAPAMAQ9wY8GCt8ttAwBQ1RDwADe4TAoA4PeGgAdUmMXsAgAAKBcBD6gww+wCAAAoFwEPcIMLHQMAfm8IeIAbgwePcGkPHTqijJEAAFQNBDzAjf3797q0//OfvWWMBACgaiDgAW5kZm5yaWdkpJtUCQAAniHgAW5cc00jl3ZwcKMyRgIAUDUQ8AA3fvrpsEv78OHDZYwEAKBqIOABAAD4GAIeAACAjyHgAQAA+BgCHgAAgI8h4AEAAPgYAh4AAICPIeABbgQFuV73rvh18QAAqGoIeIAbP/98xqV95syZMkYCAFA1EPAAN+rVq+fSrl+/vkmVAADgGQIe4MapUydd2idPnjCpEgAAPEPAAwAA8DEEPMCNa64Jdmk3ahRcxkgAAKoGAh7gRsuWrYq1/2BOIQAAeIiAB7ixZ88ul/bu3d+aVAkAAJ4h4AFutGvXwaXdvn2HMkYCAFA1EPAANw4d+tGlffDgQXMKAQDAQwQ8wI3il0U5eTLHpEoAAPAMAQ8AAMDHEPAAAAB8DAEPAADAxxDwAAAAfAwBDwAAwMcQ8AAAAHwMAQ8AAMDHVDO7AEmKi4tTnTp15OfnJ39/f3366WH+lF8AAA3tSURBVKey2WwaP368jhw5oubNm+vNN99UYGCgDMPQK6+8onXr1qlWrVp69dVX1a5dO0lScnKy5s6dK0l67LHHlJiYaOZqAQAAmKLKzOD985//1IoVK/Tpp59KkubPn6/IyEilpqYqMjJS8+fPlySlpaXp4MGDSk1N1ZQpU/Tyyy9Lkmw2m5KSkrRkyRItXbpUSUlJysvLM2t1AAAATFNlAl5xq1ev1oABAyRJAwYM0KpVq1z6LRaLOnfurDNnzignJ0cbNmxQdHS0rFarAgMDFR0drfXr15u5CgAAAKaoMgHvwQcf1MCBA7V48WJJ0qlTpxQcHCxJCg4OVm5uriQpOztbTZo0cf5ckyZNlJ2dXaK/cePGys7OvoJrAAAAUDVUiWPwPvroIzVu3FinTp3SiBEjdP3115c51jCMEn0Wi6XMfnf8/S2yWgMqVjCuemwzACoD3y3wlioR8Bo3bixJCgoKUs+ePbVz504FBQUpJydHwcHBysnJUcOGDSUVztgdP37c+bPHjx9XcHCwmjRpos2bNzv7s7OzFRYW5va57XZDNts5L68RfB3bDIDKwHcL3GnUqJ5H40zfRXvu3DmdPXvW+e+NGzeqdevWiouL0/LlyyVJy5cv12233SZJzn7DMLRjxw7Vq1dPwcHBiomJ0YYNG5SXl6e8vDxt2LBBMTExpq0XfMcNN4S4tENC2phUCQAAnjF9Bu/UqVMaPXq0JMlutyshIUGxsbHq0KGDxo0bp2XLlqlp06aaNWuWJKlHjx5at26devbsqdq1a2vatGmSJKvVqlGjRunuu++WJI0ePVpWq9WclYJPyclxPZYzO/t4GSPx/9u7u9ioy3yB47+hpS7y6rp9ydFuQ0BKoj1aokt4UZImLgloTAQSzMKJGvEteEOCUjBcQCAoamI0kACBsB6TjSjcwJXhgpdK9EbTkBPOHhRwlpWpG0DYgrSZ9lx47JEFoZS2f/r087lipsPMbyBtv/M88/8PALeGXOfV3rw2iLS3Fy2Jc02LFi284rrNmz/IYBIgJX620BMDZosWAIDeJfAAABIj8AAAEiPwAAASI/AAABIj8AAAEiPwAAASI/AAABIj8AAAEiPwAAASI/AAABIj8AAAEiPwAAASI/AAABIj8AAAEiPwAAASI/AAABIj8AAAEiPwAAASI/AAABIj8AAAEiPwAAASI/AAABIj8AAAEiPwAAASI/AAABIj8AAAEiPwAAASI/AAABIj8AAAEiPwAAASI/AAABIj8AAAEiPwAAASI/AAABIj8AAAEiPwAAASI/AAABIj8AAAEiPwAAASI/AAABIj8AAAEiPwAAASI/AAABIj8AAAEiPwAAASU5r1AKThs88ORlPTvqzH6Dfr16/JeoReN23ajJg6dXrWYwDQC6zgAQAkJtfZ2dmZ9RBZam8vxtmzF7Ieg1vYokULr7hu8+YPMpgEBp+Udwf++tcjV1w3YcLEDCbpW3YHeld5+chu3c4KHlxHScnl72QoLR2a0SQA0D1W8Kzg0Q2/XMWzegf0BrsD9IQVPOgDVu8AGAgEHnTDhAkTY8KEibFx49asRwES8a+rdVbv6E0CDwAgMc6DBwAZ+fmo2aVLV2Q8CamxggcAkBiBBwCQGIEHAJAYgQcAkBiBBwCQGIEHAJAYgQcAkBiBBwCQGIEHAJAYgQcAkBiBBwCQGIEHAJCY0qwHGEz+8pf/jHz+RNZj0AM//7+tX78m40noqerqmpg/f0HWYwD0C4HXj/L5E/HfJ45FsaIi61G4QblhwyIi4r8utmY8CT1R0tKS9Qh9yovHgcuLx4HvVn3xKPD6WbGiIi786U9ZjwGDyu0ffpj1CH0qnz8RJ779n6isGp71KNygYcM7IyLix7a/ZzwJPVE4deu+6Bd4AAmorBoe//Hsv2c9Bgwqf97anPUIv8pBFgAAiRF4AACJEXgAAIkReAAAiXGQRT86d+6HKDlzOvkj+uBWU9LSEufu+G3WYwD0Gyt4AACJsYLXj0aNGh1/G1rqPHjQz27/8MMYNcw54oDBQ+ABDHDnzv0QZ8603tLn5IIUFb5rjTvu+CHrMa7KFi0AQGKs4PWzkpYWB1kMQLnWnz6OpnO4bb6BqKSlJaJmbNZj9JlRo0ZH2W9afZIF9LM/b22O35SNznqMqxJ4/ai6uibrEeih/D/+ERER1b+ryHgSeqRmrO8/YFBJLvD2798fa9asiY6Ojpg3b148//zzWY/UZf78BVmPQA+tX78mIiKWLl2R8SRwdYVT3oM3EP3zn20RETFiRFnGk9AThVOtUfP7rKe4uqQCr1gsxqpVq2Lbtm1RWVkZc+fOjYaGhhg/fnzWowH0GauTA9f3rSciIuJ3v/23jCehJ2p+f+t+/yUVeM3NzVFTUxPV1dURETF79uzYu3evwOsHn312MJqa9mU9Rp/J53/6IfzzSl6Kpk2bEVOnTs96DHrA7sDAZXeAvpJU4BUKhaiqquq6XFlZGc3Ntiy4eaNHj8l6BBiUvHgc+Lx4zEZSgdfZ2XnFdblc7pp/p6QkF2PG3N5XIw0as2b9MWbN+mPWYwCJGT68LEpLS7Ieo8/c8X8foZfycxw+vMzv2QwkFXhVVVVx6tSprsuFQiEqKq591GOx2Blnz17o69EA6IH77/9D3H//H7Ieg5vk92zvKS8f2a3bJXWi47q6ujh+/Hjk8/loa2uLPXv2RENDQ9ZjAQD0q6RW8EpLS2PlypXx3HPPRbFYjDlz5sQ999yT9VgAAP0q13m1N64NIu3tRUvHAMCAMCi3aAEAEHgAAMkReAAAiRF4AACJEXgAAIkReAAAiRF4AACJEXgAAIkReAAAiRF4AACJEXgAAIkReAAAiRF4AACJEXgAAIkReAAAicl1dnZ2Zj0EAAC9xwoeAEBiBB4AQGIEHgBAYgQeAEBiBB4AQGIEHgBAYgQedMP+/ftj5syZ8eijj8amTZuyHgdIQGNjY0yZMiUee+yxrEchQQIPrqNYLMaqVatiy5YtsWfPnti9e3ccPXo067GAAe7JJ5+MLVu2ZD0GiRJ4cB3Nzc1RU1MT1dXVUVZWFrNnz469e/dmPRYwwD300EMxevTorMcgUQIPrqNQKERVVVXX5crKyigUChlOBADXJvDgOq72aX65XC6DSQCgewQeXEdVVVWcOnWq63KhUIiKiooMJwKAaxN4cB11dXVx/PjxyOfz0dbWFnv27ImGhoasxwKAX5XrvNr+E3CZffv2xdq1a6NYLMacOXPipZdeynokYIBbsmRJfPHFF3HmzJm4884745VXXol58+ZlPRaJEHgAAImxRQsAkBiBBwCQGIEHAJAYgQcAkBiBBwCQGIEH0EcWLlwYq1atynoMYBBymhRgUFu2bFns2rUrIiJKSkqioqIiZsyYEUuWLOn2B8Hv3LkzVq9eHV9++eVl1589ezZKS0tjxIgRvT43wLWUZj0AQNamTp0ab775ZhSLxTh69GgsX748zp8/H++8885N3e+YMWN6aUKAG2OLFhj0ysrKory8PKqqqmL69Okxa9asaGpq6vr6tm3b4vHHH48HHnggHn744VixYkWcO3cuIiI+//zzaGxsjAsXLkRtbW3U1tbGe++9FxFXbtE2NDTEhg0bYuXKlTFp0qR45JFHYsuWLZfNcuzYsViwYEHU1dXFzJkzY9++fVFfXx87d+7sh38JIBUCD+AX8vl8HDhwIEpL/3+DI5fLxfLly2P37t3x9ttvR3Nzc6xevToiIurr62P58uUxbNiwOHjwYBw8eDCeffbZX73/7du3x4QJE2LXrl2xaNGiWL9+fdfWbkdHRyxevDhKSkrio48+inXr1sX7778fbW1tffukgeTYogUGvQMHDkR9fX0Ui8W4dOlSREQ0NjZ2ff3pp5/u+vPdd98dS5cujZdffjneeOONKCsri5EjR0Yul4vy8vLrPta0adNiwYIFEfHTCt8HH3wQhw4divr6+mhqaopjx47F1q1bo7KysmuOp556qhefLTAYCDxg0HvwwQdj9erV8eOPP8aOHTvi22+/jYULF3Z9/dChQ7Fp06b4+uuv4/z589HR0RHt7e3x/fffd4VYd9XW1l52uaKiIk6fPh0REd98801UVFRcdp91dXUxZIjNFuDG+KkBDHrDhg2LmpqaqK2tjddffz0uXrwYGzZsiIiIkydPxgsvvBDjxo2Ld999N3bu3Blr166NiIj29vYbfqxfbv1G/LT929HRERERnZ2dkcvlbvLZAAg8gCssXrw4Nm/eHIVCIQ4fPhzt7e3R2NgY9fX1MXbs2Ghpabns9kOHDo1isXjTjztu3LgoFApRKBS6rjt8+HBXAAJ0l8AD+BeTJ0+O8ePHx8aNG6OmpiY6Ojpi+/btkc/nY/fu3bF9+/bLbn/XXXfFpUuXoqmpKU6fPh0XL17s0eNOmzYtxo4dG8uWLYsjR47EV199FevWrYvS0lIre8ANEXgAV/HMM8/Exx9/HCNHjowVK1bEtm3bYvbs2bFjx4549dVXL7vtpEmTYv78+bFkyZKYMmXKFac+6a4hQ4Z0HTU7d+7ceO211+LFF1+MXC4Xt912W288LWCQ8EkWALewI0eOxBNPPBGffPJJ3HfffVmPAwwQjqIFuIV8+umnXQd9nDx5MtatWxcTJ06Me++9N+vRgAFE4AHcQlpbW+Ott96K7777LkaNGhWTJ0+OxsZG78EDbogtWgCAxDjIAgAgMQIPACAxAg8AIDECDwAgMQIPACAxAg8AIDH/C0Q5oSS6nW/qAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 720x576 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(10,8))\n",
    "sns.boxplot(x='rating',y='review_length',data=amazon_df,palette='rainbow')\n",
    "plt.title(\"Boxplot of review length for each rating category.\",fontsize=16)\n",
    "plt.xlabel(\"Rating\",fontsize=14)\n",
    "plt.ylabel(\"Review Length\",fontsize=14)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* Looking at the box and whisker plots for the review_length in words, we can see an exponential distribution. We can observe that the mass of the distribution can possibly be covered with 900 to 1000 words."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Creating a countplot of the number of occurrences for each type of rating.**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x576 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(10,8))\n",
    "sns.countplot(x='rating',data=amazon_df,palette='winter')\n",
    "plt.title(\"Number of occurrences for each type of rating\",fontsize=16)\n",
    "plt.xlabel(\"Rating\",fontsize=14)\n",
    "plt.ylabel(\"Number of occurrences\",fontsize=14)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Pre-processing of review text**\n",
    "\n",
    "[Regex reference](https://stackoverflow.com/questions/9662346/python-code-to-remove-html-tags-from-a-string/9662410)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "def review_preprocess(review):\n",
    "    \"\"\"\n",
    "    Takes in a string of review, then performs the following:\n",
    "    1. Remove HTML tag from review\n",
    "    2. Remove URLs from review\n",
    "    3. Make entire review lowercase\n",
    "    4. Split the review in words\n",
    "    5. Remove all punctuation\n",
    "    6. Remove empty strings from review\n",
    "    7. Remove all stopwords\n",
    "    8. Returns a list of the cleaned review after jioning them back to a sentence\n",
    "    \"\"\"\n",
    "    en_stops = set(stopwords.words('english'))\n",
    "    \n",
    "    \"\"\"\n",
    "    Removing HTML tag from review\n",
    "    \"\"\"\n",
    "    clean = re.compile('<.*?>')\n",
    "    review_without_tag = re.sub(clean, '', review) \n",
    "    \n",
    "    \n",
    "    \"\"\"\n",
    "    Removing URLs\n",
    "    \"\"\"\n",
    "    review_without_tag_and_url = re.sub(r\"http\\S+\", \"\", review_without_tag)\n",
    "    \n",
    "    review_without_tag_and_url = re.sub(r\"www\\S+\", \"\", review_without_tag)\n",
    "    \n",
    "    \"\"\"\n",
    "    Make entire string lowercase\n",
    "    \"\"\"\n",
    "    review_lowercase = review_without_tag_and_url.lower()\n",
    "    \n",
    "    \"\"\"\n",
    "    Split string into words\n",
    "    \"\"\"\n",
    "    list_of_words = word_tokenize(review_lowercase)\n",
    "    \n",
    "    \n",
    "    \"\"\"\n",
    "    Remove punctuation\n",
    "    Checking characters to see if they are in punctuation\n",
    "    \"\"\"\n",
    "\n",
    "    list_of_words_without_punctuation=[''.join(this_char for this_char in this_string if (this_char in string.ascii_lowercase))for this_string in list_of_words]\n",
    "     \n",
    "    \n",
    "    \"\"\"\n",
    "    Remove empty strings\n",
    "    \"\"\"\n",
    "    list_of_words_without_punctuation = list(filter(None, list_of_words_without_punctuation))\n",
    "    \n",
    "    \n",
    "    \"\"\"\n",
    "    Remove any stopwords\n",
    "    \"\"\"\n",
    "  \n",
    "    filtered_word_list = [w for w in list_of_words_without_punctuation if w not in en_stops] \n",
    "    \n",
    "    \"\"\"\n",
    "    Returns a list of the cleaned review after jioning them back to a sentence\n",
    "    \"\"\"\n",
    "    return ' '.join(filtered_word_list)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "40739    Great to use once you get used to Windows 8 -w...\n",
       "40740    I love this product. It is amazing and it has ...\n",
       "40741    It fast and easy to use win 8 battery very goo...\n",
       "40742    The computer is a bit light,the outside of the...\n",
       "40743    Screen failed after 3 months and only limited ...\n",
       "Name: review, dtype: object"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\"\"\"\n",
    "Here is the original reviews:\n",
    "\"\"\"\n",
    "amazon_df['review'].tail()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Applying pre-processing to reviews**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Elapsed time for review preprocessing :  1.522184415658315  in minutes\n"
     ]
    }
   ],
   "source": [
    "start_time=time.time()\n",
    "amazon_df['review']=amazon_df['review'].apply(review_preprocess)\n",
    "print('Elapsed time for review preprocessing : ',((time.time()-start_time)/60),' in minutes')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "40739    great use get used windows lovevery slickcompu...\n",
       "40740    love product amazing cool feature like multifu...\n",
       "40741    fast easy use win battery good feel comf brand...\n",
       "40742    computer bit light outside computer elegant dr...\n",
       "40743    screen failed months limited use contacted son...\n",
       "Name: review, dtype: object"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\"\"\"\n",
    "Here is the reviews after preprocessing :\n",
    "\"\"\"\n",
    "amazon_df['review'].tail()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### The term frequency distribution of words in the review is obtained using nltk.FreqDist(). This provides us a rough idea of the main topic in the review dataset.\n",
    "\n",
    "[FreqDist reference](http://www.nltk.org/api/nltk.html?highlight=freqdist)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of vocabulary : 84381\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "['laptop',\n",
       " 'nt',\n",
       " 'computer',\n",
       " 'windows',\n",
       " 'use',\n",
       " 'one',\n",
       " 'screen',\n",
       " 'like',\n",
       " 'would',\n",
       " 'great',\n",
       " 'get',\n",
       " 'good',\n",
       " 'keyboard',\n",
       " 'battery',\n",
       " 'really',\n",
       " 'time',\n",
       " 'also',\n",
       " 'work',\n",
       " 'well',\n",
       " 'even',\n",
       " 'drive',\n",
       " 'much',\n",
       " 'new',\n",
       " 'price',\n",
       " 'fast',\n",
       " 'machine',\n",
       " 'used',\n",
       " 'gb',\n",
       " 'bought',\n",
       " 'need',\n",
       " 'still',\n",
       " 'buy',\n",
       " 'could',\n",
       " 'love',\n",
       " 'using',\n",
       " 'first',\n",
       " 'macbook',\n",
       " 'back',\n",
       " 'little',\n",
       " 'hard',\n",
       " 'life',\n",
       " 'got',\n",
       " 'better',\n",
       " 'chromebook',\n",
       " 'nice',\n",
       " 'problem',\n",
       " 'apple',\n",
       " 'works',\n",
       " 'thing',\n",
       " 'hp']"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "reviews = amazon_df['review'].str.cat(sep=' ')\n",
    "\n",
    "\"\"\"\n",
    "function to split review into word\n",
    "\"\"\"\n",
    "tokens = word_tokenize(reviews)\n",
    "\n",
    "vocabulary = set(tokens)\n",
    "print('Number of vocabulary : {}'.format(len(vocabulary)))\n",
    "\n",
    "frequency_distribution = nltk.FreqDist(tokens)\n",
    "sorted(frequency_distribution,key=frequency_distribution.__getitem__, reverse=True)[0:50]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Wordcloud visualization of frequent words\n",
    "\n",
    "[Wordcloud reference](https://www.geeksforgeeks.org/generating-word-cloud-python/)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "wordcloud = WordCloud().generate_from_frequencies(frequency_distribution)\n",
    "plt.imshow(wordcloud)\n",
    "plt.axis(\"off\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Train, test split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "def split_train_test(x, y):\n",
    "    SEED = 2000\n",
    "    x_train, x_validation_and_test, y_train, y_validation_and_test = train_test_split(x, y, test_size=.2, random_state=SEED)\n",
    "    x_validation, x_test, y_validation, y_test = train_test_split(x_validation_and_test, y_validation_and_test, test_size=.5, random_state=SEED)\n",
    "    return x_train, y_train, x_test, y_test, x_validation, y_validation, pd.concat([x_train,x_validation,x_test])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "X = amazon_df.review\n",
    "y = amazon_df.rating\n",
    "x_train, y_train, x_test, y_test, x_validation, y_validation, all_reviews = split_train_test(X, y)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### We are done pre-processing! Now on to the data analysis and machine learning tasks...\n",
    "\n",
    "* In the next code cell we reload data if needed or for first time load we simply read data into lists of tokenized items\n",
    "* Then we train the Word2Vec model"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<span style=\"display:block;text-align:center\">![hobbits](hobbits.jpg)\n",
    "\n",
    "### Concerning embeddings:\n",
    "* Developed by [Tomas Mikolov in 2013 at Google](https://arxiv.org/pdf/1310.4546.pdf), **Word2Vec** is one of the most popular algorithms to train \"word embeddings\" using a shallow two layer neural networks having one input layer, one hidden layer and one output layer. There are two models for generating word embeddings, i.e. CBOW and Skip-gram.\n",
    "* **Word Embedding** is a language modeling technique that uses vectors with several dimensions to represent words from large amounts of unstructured text data. Word embeddings can be generated using various methods like neural networks, co-occurrence matrix, probabilistic models, etc.\n",
    "* **CBOW (Continuous Bag of Words) model** CBOW model predicts the current word given a context of words. The input layer contains context words and the output layer contains current predicted word. The hidden layer contains the number of dimensions in which we want to represent current word at output layer. The CBOW architecture is shown below ([image credit: Google](https://www.google.com/search?q=cbow&newwindow=1&source=lnms&tbm=isch&sa=X&ved=0ahUKEwibmL_l0YLiAhWoT98KHX8QAisQ_AUIDigB&biw=1855&bih=906))\n",
    "\n",
    "<span style=\"display:block;text-align:center\">![CBOW](CBOW_ORIGINAL.png)\n",
    "\n",
    "* **Skip-gram model** flips CBOW's network architecture and aims to predict context given a word. Given current word, Skip gram predicts the surrounding context words. Input layer for it contains the current word and output layer provides the context words. The hidden layer consists of number of dimensions in which we want to represent current input word.The skip-gram architecture is shown below ([image credit: Google](https://www.google.com/search?q=skip+gram&newwindow=1&source=lnms&tbm=isch&sa=X&ved=0ahUKEwjSirSW0oLiAhUOd98KHbbgA5EQ_AUIDygC&biw=1855&bih=952))\n",
    "\n",
    "<span style=\"display:block;text-align:center\">![SG](skip.png)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* For our Word2Vec modeling we have used CBOW as it is faster and have better representation for more frequent words.\n",
    "\n",
    "[Word2Vec reference](https://radimrehurek.com/gensim/models/word2vec.html)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Reload previously processed model if exists**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "list_of_tokenized_reviews = []\n",
    "skip_modeling = False\n",
    "\n",
    "filename_to_save_model = AMAZON_LAPTOPS + \".model\"\n",
    "if check_if_file_exists(filename_to_save_model):\n",
    "    skip_modeling = True"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "W0506 18:16:04.071289 139801131669312 smart_open_lib.py:385] this function is deprecated, use smart_open.open instead\n"
     ]
    }
   ],
   "source": [
    "if not skip_modeling:\n",
    "    for one_sentence in all_reviews:\n",
    "        list_of_tokenized_reviews.append(gensim.utils.simple_preprocess(one_sentence))\n",
    "    model = Word2Vec(list_of_tokenized_reviews, size=150, window=10, min_count=2, workers=10)\n",
    "    model.save(filename_to_save_model)\n",
    "    model = Word2Vec.load(filename_to_save_model)\n",
    "else:\n",
    "    model = Word2Vec.load(filename_to_save_model)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Let's look at some output"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[('horrible', 0.863848865032196),\n",
       " ('awful', 0.8366063833236694),\n",
       " ('poor', 0.6874389052391052),\n",
       " ('crappy', 0.6861168146133423),\n",
       " ('bad', 0.680998682975769),\n",
       " ('suck', 0.6512565016746521),\n",
       " ('sucks', 0.6458801031112671),\n",
       " ('subpar', 0.6320646405220032),\n",
       " ('worst', 0.6144826412200928),\n",
       " ('stupid', 0.5905051231384277)]"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\"\"\"\n",
    "look up top 10 words similar to the word 'terrible'.\n",
    "\"\"\"\n",
    "w1 = \"terrible\"\n",
    "model.wv.most_similar(positive=w1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[('fantastic', 0.7522412538528442),\n",
       " ('outstanding', 0.7373641133308411),\n",
       " ('superb', 0.7291960716247559),\n",
       " ('terrific', 0.7222232818603516),\n",
       " ('wonderful', 0.6978937983512878),\n",
       " ('great', 0.695469856262207),\n",
       " ('exceptional', 0.646528422832489),\n",
       " ('incredible', 0.6273177862167358),\n",
       " ('fabulous', 0.6217853426933289),\n",
       " ('good', 0.6088477373123169)]"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\"\"\"\n",
    "look up top 10 words similar to 'excellent'\n",
    "\"\"\"\n",
    "w1 = [\"excellent\"]\n",
    "model.wv.most_similar (positive=w1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[('macs', 0.6919459700584412),\n",
       " ('imac', 0.5886913537979126),\n",
       " ('apple', 0.5822550654411316),\n",
       " ('osx', 0.5632225275039673),\n",
       " ('snow', 0.5617016553878784),\n",
       " ('apples', 0.5597377419471741),\n",
       " ('pcs', 0.5570902228355408),\n",
       " ('macintosh', 0.5568829774856567)]"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\"\"\"\n",
    "look up top 8 words similar to 'mac'\n",
    "\"\"\"\n",
    "w1 = [\"mac\"]\n",
    "model.wv.most_similar (positive=w1,topn=8)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[('problem', 0.9046769738197327),\n",
       " ('issues', 0.750947117805481),\n",
       " ('problems', 0.6805667877197266),\n",
       " ('bug', 0.678601086139679),\n",
       " ('glitch', 0.6227734684944153)]"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\"\"\"\n",
    "look up top 5 words similar to 'issue'\n",
    "\"\"\"\n",
    "w1 = [\"issue\"]\n",
    "model.wv.most_similar (positive=w1,topn=5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "-0.13857192"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\"\"\"\n",
    "similarity between two different words\n",
    "\"\"\"\n",
    "model.wv.similarity(w1=\"great\",w2=\"worse\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1.0"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\"\"\"\n",
    "similarity between two identical words\n",
    "\"\"\"\n",
    "model.wv.similarity(w1=\"outstanding\",w2=\"outstanding\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.73736417"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\"\"\"\n",
    "similarity between two related words\n",
    "\"\"\"\n",
    "model.wv.similarity(w1=\"excellent\",w2=\"outstanding\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'disapointed'"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\"\"\"\n",
    "Which one is the odd one out in this list?\n",
    "\"\"\"\n",
    "model.wv.doesnt_match([\"best\",\"great\",\"good\",\"disapointed\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'wire'"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\"\"\"\n",
    "Which one is the odd one out in this list?\n",
    "\"\"\"\n",
    "model.wv.doesnt_match([\"mac\",\"ipad\",\"mackbook\",\"wire\"])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* **Using dimensionality reduction algorithms like PCA and t-SNE to convert multi-dimensional word vectors to two dimensional plots**\n",
    "* The goal is to plot our 150 dimensions vectors into 2 dimensional graphs, and check if we can spot interesting patterns.\n",
    "* Using PCA and t-SNE implementation from scikit-learn for dimension reductions.\n",
    "* In the visualizations we will look at query word (in **blue**), and most similar words (in **green**), and list of words passed in the function (in **red**)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [],
   "source": [
    "def word_vectors_plot(model, input_word, word_list):\n",
    "    \"\"\" \n",
    "    Seaborn plot results of query word and most similar words, alongwith other words in corpus\n",
    "    \"\"\"\n",
    "    word_arrays = np.empty((0, 150), dtype='f')\n",
    "    word_tags = [input_word]\n",
    "    color_list  = ['blue']\n",
    "\n",
    "    \"\"\"\n",
    "    Creating Vector of query word\n",
    "    \"\"\"\n",
    "    word_arrays = np.append(word_arrays, model.wv.__getitem__([input_word]), axis=0)\n",
    "    \n",
    "    \"\"\"\n",
    "    Find similar words\n",
    "    \"\"\"\n",
    "    similar_words = model.wv.most_similar([input_word],topn=8)\n",
    "    \n",
    "    \"\"\"\n",
    "    Insert word vector for similar words into array\n",
    "    \"\"\"\n",
    "    for word_score in similar_words:\n",
    "        word_vector = model.wv.__getitem__([word_score[0]])\n",
    "        word_tags.append(word_score[0])\n",
    "        color_list.append('green')\n",
    "        word_arrays = np.append(word_arrays, word_vector, axis=0)\n",
    "    \n",
    "    \"\"\"\n",
    "    Insert word vectors for other words into array\n",
    "    \"\"\"\n",
    "    for word in word_list:\n",
    "        word_vector = model.wv.__getitem__([word])\n",
    "        word_tags.append(word)\n",
    "        color_list.append('red')\n",
    "        word_arrays = np.append(word_arrays, word_vector, axis=0)\n",
    "        \n",
    "    \"\"\"\n",
    "    Dimensionality from 150 to 17 dimensions with PCA\n",
    "    \"\"\"\n",
    "    reduce = PCA(n_components=17).fit_transform(word_arrays)\n",
    "    \n",
    "    \n",
    "    \"\"\"\n",
    "    Finds t-SNE coordinates for 2 dimensions\n",
    "    \"\"\"\n",
    "    np.set_printoptions(suppress=True)\n",
    "    \n",
    "    Y = TSNE(n_components=2, random_state=0, perplexity=15).fit_transform(reduce)\n",
    "    \n",
    "    \n",
    "    \"\"\"\n",
    "    Sets everything up to plot\n",
    "    \"\"\"\n",
    "    df = pd.DataFrame({'x': [x for x in Y[:, 0]],\n",
    "                       'y': [y for y in Y[:, 1]],\n",
    "                       'words': word_tags,\n",
    "                       'color': color_list})\n",
    "    \n",
    "    fig, _ = plt.subplots()\n",
    "    fig.set_size_inches(9, 9)\n",
    "    \n",
    "    \n",
    "    \"\"\"\n",
    "    Original plot\n",
    "    \"\"\"\n",
    "    p1 = sns.regplot(data=df,\n",
    "                     x=\"x\",\n",
    "                     y=\"y\",\n",
    "                     fit_reg=False,\n",
    "                     marker=\"o\",\n",
    "                     scatter_kws={'s': 40,\n",
    "                                  'facecolors': df['color']\n",
    "                                 }\n",
    "                    )\n",
    "    \n",
    "    \n",
    "    \"\"\"\n",
    "    Annotating word in plots\n",
    "    \"\"\"\n",
    "    for line in range(0, df.shape[0]):\n",
    "         p1.text(df[\"x\"][line],\n",
    "                 df['y'][line],\n",
    "                 '  ' + df[\"words\"][line].title(),\n",
    "                 horizontalalignment='left',\n",
    "                 verticalalignment='bottom', size='medium',\n",
    "                 color=df['color'][line],\n",
    "                 weight='normal'\n",
    "                ).set_size(15)\n",
    "\n",
    "    \n",
    "    plt.xlim(Y[:, 0].min()-50, Y[:, 0].max()+50)\n",
    "    plt.ylim(Y[:, 1].min()-50, Y[:, 1].max()+50)\n",
    "            \n",
    "    plt.title('t-SNE viz for input: {}'.format(input_word.title()),fontsize=16)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Eight Most Similar Words Vs. Eight Random Words"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 648x648 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "word_vectors_plot(model, 'mac', ['cheap', 'perform', 'risk', 'life', 'issue', 'fix', 'buy', 'love'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Eight Most Similar Words Vs. Ninth To Sixteenth Most Similar Words"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAjgAAAIoCAYAAABko5FtAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvhp/UCwAAIABJREFUeJzs3Xd4U9Ufx/F3kiZp071boAMEyip7ytKy9xJZIiAge8geP1BAEBFlT1kiSxAZMhRBBREEVMDFkj0LlJbSmXV/f1QCoaUUKBTC9/U8fSQn5557ToL0k3POvVEpiqIghBBCCOFA1DndASGEEEKI7CYBRwghhBAORwKOEEIIIRyOBBwhhBBCOBwJOEIIIYRwOBJwhBBCCOFwJOCIF9727dtZvHjxQx/Trl07KlWqRPHixXn11Vfp2bMnu3btstXZt28fERERFClShNOnT6dro1q1agwbNsz2+MKFC0RERNz358iRI48+yP/MmDGDiIiIx24H4Msvv6R27doUK1aMsmXLZkubGbn9Ou7bt++JnSOr4uPjmTFjBn///fcjt/E03udHFRERwYwZM2yPs/Pvy/3MmDGDvXv3PtFziBeTU053QIictn37dvbs2UOnTp2yVH/p0qWMHz+eFi1a0LlzZ1xcXDh//jw//vgjv/zyC9WqVbOrb7FYmD59OlOmTMlS+926dSMqKipdeXh4eJaOz0zLli2pWrXqY7cTHR3N6NGjadSoERMmTECv1z92m/dTtGhRvvjiC/Lnz//EzpFV8fHxzJw5k6CgIIoWLfpYbT3J9/l5MnPmTLp3706lSpVyuivCwUjAEeIhLVq0iJo1azJhwgRbWaVKlXj99dexWq3p6lepUoWtW7fSrVs3ChUq9MD2Q0JCKFmyZLb2+bagoCCCgoIeu52zZ89isVho2rRptszemEwmnJycUKlU6Z5zc3N7Yq9HTnqS77MQQpaoxAtu2LBhrFu3jujoaNsSQUafqu928+ZN/Pz8MnxOrU7/v1S7du3w9/dn6tSp2dLnjDRo0IA+ffqkKz98+DARERFs374dSL/k0L59+/sulVy4cCHDcw0bNoz27dsD0LFjRyIiImxLbSaTiSlTphAVFUWxYsWIiopiypQpmEwm2/G3l2iWL1/OpEmTqFKlCpGRkcTHx2d4voyWqNq3b0+bNm3Ys2cPzZo1o0SJEjRs2NA2zttuj/fYsWO0b9+eEiVKUKVKFaZNm2YXRr/66qsMx3z363XhwgVq1KgBwP/+9z/b6/TVV19l2O/HNWXKFIoVK8Yff/xhK0tKSqJOnTq0atUKs9lsK9+/fz+dOnWiTJkylCxZksaNG7NmzRq79lavXk3jxo2JjIykQoUKjBgxgri4uIful9lsZt68edStW5dixYpRpUoVJk6cSGpqqq3O7fd41apVTJs2jSpVqlC2bFm6d+/OlStXbPVuv7Zz5861vZ53L5EJ8ThkBke80Hr27MmNGzf4888/mTNnDgA6nS7TYyIjI1m/fj0hISHUqFGDvHnzZlrf2dmZHj16MGbMGA4dOvTAT+1Wq9XulxeASqVCo9Hc95gmTZowffp0bt68iaenp61848aNeHl5pVs2u+3dd98lISHB9thisTBixAgSEhLw8vLK8JiePXtStGhR3n//fUaPHk3RokXx8fEB0sLP7dmqMmXKcOjQIebMmcOFCxf4+OOP7dqZO3cukZGRjBs3DovF8tDLXOfPn2f8+PG8/fbbeHt7s3jxYvr27cvWrVsJCwuzq9urVy9atGhBt27d2L17N7Nnz0atVmcYCu8nICCAmTNn0rt3b7vlpdDQUCAtJA0fPpylS5dSoUKFB7b3oPe5T58+7Nu3j0GDBrFu3TpcXV0ZO3Ys169fZ8GCBTg5pf3zvX37dvr27Uvp0qUZO3Ys3t7enDhxgkuXLtnanTx5MosXL6Z9+/YMGTKE6Ohopk6dyokTJ1i1alWmf7fuNXjwYH744Qe6dOlC6dKlOXnyJNOmTePixYvpwsn8+fMpVaoU48eP58aNG0ycOJFBgwaxbNkyAL744gtatWpF8+bNadWqFUC2zDAKARJwxAsuNDQUHx8ftFptlpcLxowZQ9++ffnoo4/46KOP8PLyonLlyjRv3pwqVapkeEzLli1ZtGgRn3zyCUuXLs20/dGjRzN69Gi7MoPBwMGDB+97TKNGjZgyZQpbt26ldevWQNpsyubNm6lXr959Q9u9+1rGjh3L5cuX+fzzz3Fzc8vwmNDQUF566SXb8bdft+PHj7Np0yZ69+5tCw5VqlRBrVYzbdo0unbtardE5+fnx6xZszJclsqK2NhYli1bZtuzUrRoUdtyYPfu3e3qvv7667z99tu2PiUkJLBo0SI6dOiAh4dHls6n0+koXLgwkPHyklqtRqPRZHk8D3qfnZycmDx5Mk2bNmXs2LFUrVqVdevWMXnyZEJCQgBQFIUJEyZQqFAhli5daptBfPnll21tXrhwgYULF9KrVy969+5tKw8PD6dt27b88MMP1KxZM0t9/vXXX9myZQsffvghTZs2tZ3L09OTwYMHc+TIEdtrBJArVy67YHvjxg0mTZpEdHQ0gYGBttcwICBAlutEtpOAI8R93Pvp+vYvr7x587J+/Xp+//13du/ezeHDh/nuu+/YvHkz/fr1o2fPnuna0mq19O7dm6FDh7Jnzx67X0D36tGjh20p5O5zZyY4OJhy5cqxYcMGW8D56aefiI2Ntf0iepDly5ezYsUKpkyZQokSJbJ0zN0OHDgAQOPGje3KGzduzLRp0zhw4IBdwKlRo8YjhxuAsLAwuw25vr6++Pr62s1c3FavXj27xw0aNGDNmjUcP348264Aa9q0aZZfa8ja+5wnTx7GjBnDgAED2LJlC02bNqVRo0a250+dOsXFixfp2rVrhsujAHv27MFqtdK4cWO7v9MlSpTAzc2NAwcOZDng/PTTT2i1WmrXrm3X1u1gf+DAAbuAU716dbvjCxYsCMDly5cJDAzM0jmFeFQScIS4j3uvkrl76UGj0VCuXDnKlSsHpF1V1KVLF2bNmkW7du3slolua9y4MQsWLGDKlCmZBpzcuXMTGRn50P1t2rQpw4cP5/z584SEhLBhwwbCwsKy9Ml49+7djB8/nv79+6cLA1l18+ZNAPz9/e3Kbz++/fxtAQEBj3Se2zJ6jXU6HUajMV25r69vho+vXr36WH14HFl9n6tXr46XlxdxcXF07NjR7rnbe2gyW9aJiYkBoFatWhk+/zD7cGJiYjCZTJQqVSpLbd27zHl7JvHu/TpCPCkScIS4jy+//NLucWZ7bQIDA2nZsiXjx4/n7NmzFC9ePF0dtVpNv3796N27d7rNsNmhdu3ajB07lo0bN9KhQwd++OEH27JMZk6ePEn//v1p3LhxuqWdh3E7cFy/ft22LwXg2rVrQPpfdo8ze/OwYmJiMBgMdo/hTsi6vf/n7s3Q8HC//J+UsWPHYrVaCQ0NZdSoUaxcuRKtVguAt7c3kBaw7+f2675o0aIMl+Put9fqfm3p9XqWL1+e4fOPG1qFyE4ScMQLT6fTZfiJ8n6fri9fvkxwcHC68lOnTgHc9worSPsUHRkZme4qnuzg5uZGjRo12LhxIwEBAaSmptKkSZNMj4mNjbVdvj527NjHOn/58uUB2Lx5Mz169LCVf/311wBP9GaAD7J161a7sLd582YMBoNtySRXrlwAnDhxwhZkzWYzu3fvtmvn9gxESkrK0+g2X3/9NRs2bGDq1KmEhITQunVrpk+fzsCBA4G00J07d27WrFlDq1atMgyNlStXRq1Wc+nSJSpXrvxY/alatSqffvopCQkJ2XbfGq1WKzM64omQgCNeeC+99BJxcXGsWLGCYsWKodfrM717a8OGDalQoQK1atUiT548JCQksHPnTlatWkW9evVsvyzv55133uGtt9667/Pnz5/n0KFD6crDw8Mf+Gm7SZMmbNq0iRkzZlCmTBnbZtT7GTRoELGxsYwaNYp//vnH7rkiRYo88IqyuxUoUICGDRsyc+ZMLBYLpUqV4uDBg8yZM4eGDRtm6R5AT8rq1auxWq1ERkaye/du1qxZQ58+fWwzGpGRkYSGhjJp0iSsVis6nY4VK1akm9Hx8/PDy8uLzZs3ExERgYuLC3ny5MHb25v169czYsQIlixZYgt7mXnQ+3z+/Hnee+89XnvtNduyYf/+/fn444+pXLkyFStWRKVSMWLECPr06cObb75JmzZt8Pb25tSpU8TExNC3b19CQ0Pp2rUr48aN4/Tp05QvXx69Xs/ly5f5+eefadmyJRUrVszS61ihQgUaNmxI37596dixI8WLF0etVnPx4kV27tzJoEGDHnhV4b3y58/Pjz/+SNWqVfHw8CAgIED254hsIQFHvPBatmzJ4cOHmTJlCvHx8eTOnZvvv//+vvUHDRrEzp07mT59OtevX0ej0RAeHs7AgQPp0KHDA89XuXJlypcvz/79+zN8ft68ecybNy9d+bRp06hbt+4D2/b39yc6OppevXo9sC+nT58mISEhw6WsHTt2kCdPnge2cbeJEyeSJ08e1q5dy5w5cwgICKBLly52V+/khNmzZzNu3Dhmz56Nu7s7PXr0sNsM7uTkxOzZsxk7dizDhw/H09OTDh06UKJECWbOnGmrp1arGT9+PJ988gmdOnXCbDbzwQcf0Lx5c6xWKxaLBUVRstSnzN7nmjVrMmjQIPz8/Bg5cqTtuc6dO7Nnzx6GDBnChg0b8Pb2pmbNmixatIjZs2fb6oaEhNj9XRwwYAD58uVjxYoVrFixApVKRVBQEJUqVXroOyd/9NFHfP7556xdu5a5c+ei0+nInTs3VapUyXT28n5GjRrF+PHj6d69O0aj0e4qPCEeh0rJ6v+NQgjxnJkxYwYzZ87k77//tt03RgjxYpA7GQshhBDC4UjAEUIIIYTDkSUqIYQQQjgcmcERQgghhMORgCOEEEIIh+PQlxWkXbbpmCtwGo3KYceWFTJ+Gb+MX8b/onrRx6/VZv7dfLc5dMCxWBTi4pJyuhtPhJeXwWHHlhUyfhm/jF/G/6J60cfv7++epXqyRCWEEEIIhyMBRwghhBAORwKOEEIIIRyOBBwhhBBCOBwJOEIIIYRwOBJwhBBCCOFwJOAIIYQQwuFIwBFCCCGEw5GAI4QQQgiHIwFHCCGEEA5HAo4QQgghHI4EHCGEEEI4HAk4QgghhHA4EnCEEEII4XAk4AghhBDC4UjAEUIIIYTDkYAjhBBCCIcjAUcIIYQQDkcCjhBCCCEcjgQcIYQQQjgcCThCCCGEcDgScIQQQgjhcCTgCCGEEMLhSMARQgghhMORgCOEEEIIhyMBRwghhBAORwKOEEIIIRyOBBwhhBBCOBwJOEIIIYRwOBJwhBBCCOFwJOAIIYQQwuFIwBFCCCGEw5GAI4QQQgiHIwFHCCGEEA5HAo4QQgghHI4EHCGEEEI4HAk4QgghhHA4EnCEEEII4XByNOAMHz6cSpUq0bBhQ1vZjBkzqFq1Kk2aNKFJkybs3LnT9ty8efOoVasWderU4aeffsqJLgshhBDiOeCUkydv3rw5b7zxBkOHDrUr79ixI507d7Yr+/fff9m8eTObN28mOjqaTp068e2336LRaJ5ml4UQQgjxHMjRGZxy5crh6emZpbo7duygQYMG6HQ6QkJCCAsL448//njCPRRCCCHE8yhHZ3DuZ/ny5axfv55ixYoxbNgwPD09iY6OpkSJErY6gYGBREdHZ9qORqPCy8vwpLubIzQatcOOLStk/DJ+Gb+M/0X1oo8/q565gNOmTRt69uyJSqVi2rRpTJw4kQ8++ABFUdLVValUmbZlsSjExSU9qa7mKC8vg8OOLStk/DJ+Gb+M/0X1oo/f3989S/Weuauo/Pz80Gg0qNVqWrZsyZ9//glAUFAQV65csdWLjo4mICAgp7ophBBCiGfYMxdwrl69avvz9u3bKVCgAABRUVFs3rwZo9HI+fPnOXPmDMWLF8+pbgohhBDiGZajS1QDBgxg//79xMbGUq1aNfr06cP+/fs5evQoALlz52bs2LEAFChQgHr16lG/fn00Gg2jR4+WK6iEEEIIkSGVktHmFgdhMlkcdp3yRV+DlfHL+GX8Mv4X1Ys+/ud2D44QQgghxOOSgCOEEEIIhyMBRwghhBAORwKOEEIIIRyOBBwhhBBCOBwJOEIIIYRwOBJwhBBCCOFwJOAIIYQQwuFIwBFCCCGEw5GAI4QQQgiHIwFHCCGEEA5HAo4QQgghHI4EHCGEEEI4HAk4QgghhHA4EnCEEEII4XAk4AghHtmFW+fp/30vin8WQZ65fpT5vBgjfxpCTHLMI7W39O/FbDm16aGOMVqMTNo/gT+v//FI53wctdZUp8+O7rbHfXZ0p9aa6k+9H0KI9JxyugPixWSYNAHXyRMzrWN8uQo31295Sj3KnPOiT7HmyYOxdj27ct+i+Ul+402Sho/OoZ7lnKM3jtBsfX38XPwZUWE0oe5hnIg7ztTfJrPt7Dd83exbglyDH6rNz/9ZQiGfwtTP1zDLxxitRib/OpFQjzAi/Yo/7DCy1YCyQ0gxp+RoH4QQaSTgiByR8kYHjFE1bY9dFsxDu3sX8UuW28oUd4+c6FqGXJYswFS+UrqA86JSFIWe27viqfdiS4vtuOvS3quXc1ehdlhdXvmiEkN2DWBpvZU53NOnK69nvpzuwjMvIMA9w/LwcCv79yc+1b78/LOGZs0M7NyZSOHCViCtfx98kELnzqZsP9+MGTpKl7ZQubIl29sW6ckSlcgR1ly5MZctb/uxBgaBXm9XZokolNPdFPex99LP/HX9D94pM9gWbm4LdstFl+Ld+fb0Fs7FnwVg1dHlBMz2IMGUYFe3zOfFePfnkQA0XV+fw9cO8sWxFQTM9iBgtgerjqYF3m9Ob6HmmmqEzw+iwMJQKi+pxJ6LuwHI92kuAPp+38N23O3zjtv7LtVXVSR8fjAlPitE9+86E50UnWEf5h6eSYnPClFgYShvb+vIzdQ4u3pHYv6hwVe1CJnnT+UVZfnmdPrZxXuXqG6P+5+Yv3ltYxPC5wfx8ooybDq50e44RVGYuG8cRRbnI9+nuen3fU/WnfjSbiyOpEcPI1u2JNr9LFyYnNPdeuJmzdLy88+anO7GC0MCjnguOC9egHflsvjl8cOnbCTq6dPsntfu/RmPdi3xKVYAv/BgvGpURbdxnX0bSxbiH+CB5u+/8GpUB7+wQLxqVkPz91+oEm7h3qMLvvly41OuOLqvN9iO86oXhdPRI7gsXYR/gAf+AR7o131p17bLjKn4RBbEt2Ao7j27okq4Zfe86vp13Pr3wrdIPvxCA/BqVAenwwez+VV6evZe/hmAenkbZPh8vbwNUVDYd3lvltv8sNonFPAqSM3Q2mxpvp0tzbdTM6wOp2+eovO37amSuxqf1/+COTU/pX7+BsSmxgLwVZO0PTsDygy2HRfoGgTA9eRr9CszkBUN1vB+lYmcjT9Diw0NsVjtP0FvPLmOny7s5ONXpjGq4hi2nfmW8b+MsT2fbE6m1aZmJJoSmVNzIe+UHcz/fh7GxYTzWRpbj+86Uze8HovrLief50t0+64TlxIu2p6f98cspv7+MW8WfYuFdZbi7OTM2L2Ou+wZGmqlbFn7n8hIa053SzgYWaISzzzDxx9imPIRSX3ewVTxZbS/HcAwfBjOah0pb3QAQH3+HKaXq5LyVlcUrS4t8HR7i3iNE8YGjeza8+j1NsldumHt+w6uY0fj0bUDlojCmAsXIX7R57gsXYxHzy7EVKiEEhDArSmz8GzfCnPRSJJ69wPAku8lW3vOq1dhLl6ShKkzUZ87h9t7I7H6+pE47oO0CsnJeDVvAKmpJIyZgOLjg/OCeXi2aMyNfYdQfH2fzguZjS4nXMZT74WH3jPD50PcQ9LqJV7OcpsRPoUwaF3xdfGjbFB5W/neS7tx07rx3svv28pe82pGXFwSACUDSgMQ7pnP7jiAaVGzbX+2WC2UDSxPiaWF2H/lFyrlqmx7zkntxGf1VuKkTvsn8XjsUdadWMuk6lMAWHl0GdeTr/FNi+/J5Zb7vzGG0Whd7SyNrVuJXrQt3B6AEgElKbo4P9vOfEPHYp2xWC3MPDiNDkXfYlj5/wHwamgNzsWf5WLChSy170hu3oRq1VypUsXCrFl39jO1b+/CiRNqvv8+ES+vtLLNm52YPl3HkSNqXFygdGkLkyalEBKiAHDkiJpx4/Ts3Zs2axIVZWbChFQCA5WH6tPWrU588omOo0fVeHgovP66mREjUtFq056fNEnHokVa1qxJZsgQZ/75R81LL1mZMCGVihXTwnSZMq7cuKFm8mQ9kyfrAVi3LkmWq54gCTjimaaKvYFh2sckDhlJct93ADBVfxVnYzKGjz+0BZzU19vcOchqxfRyFTTnz+G8bEm6gJPUfyCpTVukPUg14vnWG5iqVCNpyAgAzMVL4Fs4H7od20ht8waWQoVRXFyw+vljLmv/CxRAMRiIX7wM1GkTok5//4V+y9e2gOO84nM0Z05zY/cBrKFhABirVMenQklcPp1N0rBR2feCPSNUqLKtrcI+RYk3xtN7RzdaFHid8sEV8cKQpWN3nN3Gx79O4ljsUW4Z423lJ+P+tQs4VXJXs4UbgILehbiefA2jxYhOo+Ng9G+U8C9pCzcAFYIr4ufin6V+vBISZfuzj7Mvfi7+XE5Mm8G5mHCBq0nR1Amvb3dMnfD67Dj3XZbaf95YrWA225ep1Wk/np4wZUoKrVsbaNDATP36ZlaudOK77zRs3JiE4b+3fvVqJ3r3dqFZMxMDBqSiKLB7txMxMSpCQhROnVLRsKGBkiXTgpLFAh9+qKN9exe+/TYJVRb/im7Y4ES3bs68+aaJESNSOXNGzfjxeqxWGDMm1VYvOVlF797OdO9uJCBAYfJkPR07OvP774kYDLBkSTLNmhlo1MhEu3Zp+3siImTW6kmSgCOeadpf9qJKSSG1cVO7fxGVqCicPvkY1dWrKAEBqG7E4PrheHTbvkF95TIqS9qnIkt43nRtGqu+YvuzJW++/8ru7JtQfHxRPD3RXL6UpT6aqla3hRsAS0QE6hVLQVFApUK360dMpctizZX7zhg0GkwVK6E99HwuUwW7BXMzNY5bxvh0e3AAzt06l1bvIa+iykh+7wIsrbeS6Qen0Hbza2jVWppENOXdChPwc/G773EHo3+j/dbW1M/biL6lB+Dn4odKpaLe2hqkWuyvdPLQ2c9E6TQ6FBSM1rSAczUpOsMw45/FgHPvTJdOoyPFnPbL8ep/e4LuHYtvJmN73o0c6czIkc52Za1amZgxI+19iYqy0L69kUGD9ISEWBk1ypmePY2UL58WCKxWeP99PfXrm5g37857WbfundmQyZP1BAQorFyZjE6XVla0qIWXX3Zl+3YNtWo9eOZEUWDMGD2vv25m0qTbYcaCXq8wbJgz/fql4uOTVpqcrOL991OpWjWt3YCAFGrUcOWXXzRERVmIjLTi5KQQHKxQtqwEm6dBAo54pqlvpN1Pxbd8iQyf11y6gDkgAI/undEcPULSO4OxFCiI4uaGy/w5aH/+Kd0xiuddv2z++5dP8bhnqUWrg9SsXe5r9bQ/VtHqUJnNYLGAkxPqGzFo9+3FP5dPumPNhQpn6RzPmkrBabMf35zeQsuI1ume//bMFlSoqJjrZQD0mrQpeZPFCNo79eLu2ch7P7XC61IrvC7xqTf57uy3jN4znFTjIObXXnLfY7ac3oSvsx+f1l6C6r+P6+f/C14PK8AQyL9xx9OVX0u+9kjt3ds2wPXk63blMfc8diS9ehlp0sT+KiUfH/tlozFjUtm504n69Q3kzWtl6FCj7bljx+DKFTVt2tz//9FduzS0amVCrb7zuSI0VCEkROHQoawFnJMnVVy4oKZJkxS7GacqVSykpKg4elTDyy+ntaPVKnbLTbdnZy5dUgOyDJUTJOCIZ5rVyxuAuNXrUW4vvAPu7s7cupWCuUAEqlvx6H78nvjpc0ht3e7OwZZ75sBziNXbG1P5iiS8n/6+P4qzSw706PFVylWZYn7F+eS3SdTL2wA33Z1Lf6MTr/DpH3Oom7cBIe6hAAT/t7RzPPY4FYIrAvBb9AG7ZSMArVqbbnblbh56T1oUfJ3fru/j53N7ANCp00LqvfefSTEno1VrbeEGYO3x1Y803pIBpVl7YjWXEi7alqn2Xf6F69kQcHK75SHAEMg3pzcTFXrn1gnfnnk27gH1JOTJY6VkycxnMdzcoHZtMwsW6Gjb1oRef+e5GzfS/pvZXpobN1TMmKFnxgx9uucuXcra+lRMTNrMbJs2GS+JXrx4px13d+XuiVzbrFFqKiKHSMARzzRThUooOh3qq9GkvnJnH4PiZcD83yZTrl5N++/tf1EAVVwsuu3fobi5ZUs/FK0OVRZndO5lqlodwyeTsISFo3inn8V5ViQazSSbrPgYtKgfsEFBpVIxq8Z8mm9oQP2vatKrZD9CPcI4EZt2oz93nQcfVvvYVr90QBmCXXMxcvcQhpUfSWxKLDMPTku3vFXAuyA/nN/B9+e24+PsQ6hHGF+f3MCvV/YTFVqTINdgTsWdZO3RL2lZMG3mSKfREeoRzsaT6yjkWwRnjZ4ivsWoHvIq8/6Yzf92D6V2eD0OXNnHl8e/eKTXpk2hN5jy20e03dySweWGk2JOZuL+9/F1fvwN4hq1hl4l+zFm7//wdfGjfFBFvj2zhSMx/wCgVr2YF7seOqRmyRItkZEWpkzR06yZ2RZobi8LRUff/++pl5dC/fom3ngj/f1s7p0tuh9v77R6H3+cQmRk+lmY0FBZanqWScARzzTFz4+k/oNwHzIAzelTmCpUQmUxo754Fvddu7m14DOUgADMhYti+HA8it4ZrFYM0z5OCxMm44NPkgWWAgXQ/vwT2h+/R/HywhKeF+W/2aUHSW7XAedln+HVtAFJPXpjDQ1DFROD9tf9WELDSOn8drb08VHFp5iY/fNR9py9iIIRHxdPulUsxst5M//lXdi3CN+13MXkAxN5/5f3uJESQ6AhiHp5GzCg7FB8Xe4cr9PoWFJ3OUPqbTUqAAAgAElEQVR3DaDzt2/yklcBJlWfQs/tXezafKfMYC4kXKDrto7cMsYzPWoORXyL8u3pLYz+eQRxqbEEGoJ4q2Rn+pcYajvuo2pTeG/P/2i5sTGpllR+feNPaobVYVSlsSz8Yx7L/vmMMkHlWV5/NRVXlH7o18igNfBFw3UM3tmfbts6EeIRypiXJzDlt48euq2MdC/Ri5upsSz5awFzD8+iTng9+pUZyNBdA3DXZXxjPEeWkgK9ezvz6qsWPv00mVdfdWXQIGc+/zztXjkRERAcbOWLL7TUqZPx8k+1ahaOHlVTooQ1yxuK75U/v5XgYCvnz6to3/7xw4xWKzM6T5MEHJEtriWksmH/af46cQmdk4YKJfPSsEQe9E6P/+kzadAwLLnz4PLpXAwzp6K4uEBEBKYmLWx14hd8htugfnj07ILV14/kt3ugvnYN/ZeP9ok9XR8GD8dtUH883mqPOuEW8fMWkdrstawdbDAQt34LrhPfx3XCWNQx17H6B2AuU47UJs2ypX+PSlEU3v3mIIdjv0btuheV2sR5YyATfmjLROcoigVnfjfpPO4hTI2alaVzlQosw7aWO+3Kfmv/l93jcM+8rG1sfxM8gBUN7e875OVlsF0mDmmXVb8aWiPdcX1K9adPqf52ZVd72i+L3dsHgNaF2tG6UDu7sqJ+xdjSYrtd2b1fKTGjxtwHtpPROVUqFcMqjGJYhTtX1L3zQ2/yuIXgqfe69/BnwoW4ZNYdvMqxCymE+utoWtqfggFZmzE9d07Nr7/a/9ugUkGZMmkh4oMP9Fy9qmbt2rQrkGbMSKZxYwOrVjnRurUZtRpGj06lRw8XuneH5s1NqFTw009ONG9uomRJK4MHp1Knjitt27rQtq0JHx+Fy5dV7NzpROvWpixdnq1Wp+0F6tXLmVu3VNSoYUarhbNn1Wzd6sTChcm2q7qyokABK9u3OxEVZcHVVSF/fivZNMksMqBSFOXhbgjwHDGZLHb/CDqSe/+Bz0mxSUbGLN9DnX1biLryD0kaHWvyV+ZG5VcY0qz0A5c7HsWzNP6ckF3jPxJ9i8FbNpHiOtfuU64pOT+V/Xowvn7Zxz7Hk+Bo7/+RmH/Y8O9aygVVQKVS8/2571jw5zxGVRxLr1J909XP6fEfi05g+OrTxB7MjSrWE6tbIh4lzzOkaTBVXsp8GfZ+X9Wg0ShcvpzAvn0amjRxYdasFFq0uLOP7r339CxbpmXXrkSKFHEhLi6JTZucmDpVx7FjagwGKFPGwocf3rkPzokTaj74QMfu3U6kpEBQkELVqmYGDjSSK5eS5a9q2LFDw9SpOv78U4NaDWFhVmrXNjN4sBEnpzv3wTl61P6rJu5t6/BhNcOGOXPkiJqkJNUj3wcnp9//nObvn7VZTQk4z6ln6S/4ql9O47R0CV1O/Ggrs6JiQPWuNOvUkBK5M74Z3ON4lsafE7Jr/DuOX2Pinvngus2u3GpxI9A8kBVvRN3nyJzlaO//2fgz9P++F3/F/EmSKZE87iG8WeQtepbsY7dJ+racHn+/Fcc4vDEEbcydS9ktrgkE1v2bz7sWxUnzZPcN5fT4c9qLPv6sBhxZohKP7d+Tl+gUfcyuTI1CxVO/c/zqK08k4IjsEeSuR6+Ek5J2yx4biymAEG+ZO39awjzCWdd0c053I0sSjWZORqfiFGO/R0uT6EZCrI5zccnk83XNod4JcceLuT1fZCs3Nxeu6dL/Mrzm4Ye7szaDI8SzonCQO+Ge+bAkVUBR0m5nbzH54GZqSOuSETncO/Es0qhUqNUKqO033Soo4GRB94Rnb4TIKvmbKB5b1dL5WBNZk5tOd+5MesItgL0vlaFSeNauNBI5Q61SMa5uGaoEtsbl1kBcEvoQaOnDwKpVKZVHZt5Ees5aDZUKuGMJvWhXbg64RlguNbk9ne9zpBBPlyxRicdWOo8nZ2tVoqfBg9KXjpGgc+FYrgJ0bVwWTxeZwXnWeRt0jKtXmtgkI4lGC0Hu+ie+h0I837q/kocz1//lfHA8Cec9MQQmEPjSLYbUfynDPUNC5ATZZPycehY3md1IMvLX5VvoNWpK5PbAWat5Yud6Fsf/NMn4Zfw5PX6zxcpv529yLjaZAHc9FcK8nuj/83d7Fsafk1708csmY/HU+Rh0VHvp8e/sKoR49jlp1FQI96aCLEOLZ5TMQwshhBDC4UjAEUIIIYTDkYAjhBBCCIcjAUcIIYQQDkcCjhBCCCEcjgQcIYQQQjgcCThCCCGEcDgScIQQQgjhcCTgCCGEEMLhSMARQgghhMORgCOEEEIIhyMBRwghhBAORwKOEEIIIRyOBBwhhBBCOBwJOEIIIYRwOBJwhBBCCOFwJOAIIYQQwuFIwBFCCCGEw5GAI4QQQgiHIwFHCCGEEA5HAo4QQgghHI4EHCGEEEI4HAk4QgghhHA4EnCEEEII4XAk4AghhBDC4UjAEUIIIYTDkYAjhBBCCIcjAUcIIYQQDkcCjhBCCCEcjgQcIYQQQjgcCThCCCGEcDgScIQQQgjhcCTgCCGEEMLhSMARQgghhMPJ0YAzfPhwKlWqRMOGDW1lcXFxdOrUidq1a9OpUydu3rwJgKIovP/++9SqVYtGjRrx999/51S3hRBCCPGMy9GA07x5cxYsWGBXNn/+fCpVqsS2bduoVKkS8+fPB2DXrl2cOXOGbdu2MW7cON57770c6LEQQgghngc5GnDKlSuHp6enXdmOHTto2rQpAE2bNmX79u125SqVipIlSxIfH8/Vq1efep+FEEII8exzyukO3CsmJoaAgAAAAgICuHHjBgDR0dEEBQXZ6gUFBREdHW2rmxGNRoWXl+HJdjiHaDRqhx1bVsj4Zfwyfhn/i+pFH39WPXMB534URUlXplKpMj3GYlGIi0t6Ul3KUV5eBocdW1bI+GX8Mn4Z/4vqRR+/v797luo9c1dR+fr62paerl69io+PD5A2Y3PlyhVbvStXrmQ6eyOEEEKIF9czF3CioqJYv349AOvXr6dGjRp25YqicOjQIdzd3SXgCCGEECJDObpENWDAAPbv309sbCzVqlWjT58+vP322/Tv358vv/yS4OBgpk2bBkD16tXZuXMntWrVwsXFhQkTJuRk14UQQgjxDFMpGW1ucRAmk8Vh1ylf9DVYGb+MX8Yv439Rvejjf2734AghhBBCPC4JOEIIIYRwOBJwhBBCCOFwJOAIIYQQwuFIwBFCCCGEw5GAI4QQQgiHIwFHCCEeoORnhXnnh952ZQmmBILneFN+WYl09Rt8VYuWG5s8re5RaFE4k/Zn373B9l3+hRqrqxIyz5+A2R5ZPm7V0eUEzPYgwZSQbX0R4lE9N99FJYQQOaVcUAUOXNlnV/Z79K/oNXrOxJ/mWtI1/A3+ABgtRv64doi+pQfkRFezxeCd/fBz8eeLhuvQaXQ53R0hHonM4AghxAOUCyrPidjjxKXE2sp+vbKfSrkqE+Ieahd+Dl87SKollfJBFXOiq49MURRSzCkAnIg9Tr28DXg5dxXKBpXP4Z4J8Wgk4AghxAOUD66IgsKv0fttZQeu7KNsUHnKBpazCzgHruxHo9JQJrCsrexs/Bne3NqGfJ/mJu+nuXhj8+ucunnS7hwBsz2Yf3g2438ZQ+FFeSmyOB9Ddw0g1ZJqV2/vpZ8ps6A0IfP8qbmmGvsv288s3bb19GZqralOyDx/ii7Oz5g9ozBZTLbnJ+2fQKFF4fxyeS+1/6u38eQ6AmZ7YFEsjNw9lIDZHvTZ0d3Wv4V/zrM7x+02hHgWScARIhv4B3hk+OP0y94st6Hd+QP+AR5oThx/7P6oT53EP8AD7fffPXZbAor5FcfgZLAFGUVR+C36AGUDy1M2qPw9AWcfhX2L4qZLu518qiWVFhsbcyL2GJ+8Mp3pUXM4d+ssTdfXJzblht155hyeyZXEy8yq+Sk9S/Zj6d+Lmf/HHNvzVxIv02ZTC3ycvVlYZylvFulEz+1dSDYn27Wz4d+v6PRNO0oFlGZpvVUMKjeMz/9Zwvu/vGdXL9mcTJ8d3WhXpAOrGn1FEd9ibGm+HYAeJfqwpfl2BpQdkm2voxBPk+zBESKbJPXoQ2oj+42llkKFc6g3Ijs5qZ0oGVCaA1fSZnBOxB4n3hhP6cAyeOg8GLf3XYwWIzqNjl+v7KdBvka2Y1ceWcbFW+fZ2/Z3wj3zAlAmsCzllhVn6d+L6VdmoK1uiHsoM2rMBSAqtCb7r/zCllMb6VOqPwDzDs9Gr9GzodXXGBPTjjFoDfTc3tXWhqIojNkzitcj2jCp+hRbuV6jZ9iugfQrMwAfZ18gLeCMrfwB9fI2SDfmUI9QWZ4SzzWZwREim1hCQzGXLW/3o7hl7UvhxLOvfFBFfo/+DYvVwoEr+4jwLoy7zoNifsUB+OPaIc7GnyE66QrlgirYjjt49TeK+5ewhRuAXG65KR9UkX2X7Wf4XgmJsnsc4V2ISwmX7NqqHvIqBq3BVtYgX2O7Y07G/cuFhPM0eakZZqvZ9lMldzVSLCkcjTliq6tCRY3QWo/xqgjx7JKAI8RToDl2FPeuHfEpWRi/sEC8q1XAecFcUJR0ddWXLuLRpgV+YYH4lC6K8+dL7J73algbTbu2dmVZWt6yWDBMHJfWhzx+eFergH7dl+nadn+7Y+ZtKwqGKR/hU644fiH++BZ5Cc/WzVFdu5bl1+N5VC6oPEnmRP6O+ZNfo/fbQoxWo6W4f0kOXNlvW6oqH3xng3F00hX8XQLStedv8Cc2NdauzFPvafdYq9aSakmxPb6aFI2fi79dHRcnF1y1brbHMSkxALTZ/Bq55vrYfsouiwTgYsIFW10vvZdcJSUclixRCZFNVFYrmM13FahAowHSQoulQEFSX2uF4uaG0x+HcZ0wDlVKKsm9+9m1496vJymt25LctQf6TRtwH9gXS+7cmKIe75O26/gxuMybRdLg4ZiKl0S/cR0e3d7ipkaDsXGzLLfjvHIZLjOmkjhqDJaIQqhu3ED304+oUpJJH9ccR7mgCqhQceDKPg5c2Ufv/5aNANs+HH+DP8GuuQhxD7U9F2gI4tiNI+nau5Z0DW+990P1IcAQyPVk+yCZbE4m8a77ztxu8+NXphP53+zS3UI9wmx/VqlUWT63XqPHeNcmZYC4ewKaEM8SCThCZBO3kUNxGznU9thUviJxm7al/fnVGpherZH2hKJgqlAJVWICzsuWpAs4xtp1SRo2Ku24qJpozpzG9ZOPiHuMgKOKicFlwVySBg0jqf+gO21fvIDrRx88VMBx+v03jDVqkdKpy50+N2ycyRGOwcvZmwLeBdl+dhsnYo9T7q79KWUDy7P2+Gr8DQF2y1MApQPLsvrYSs7GnyHMIxyAywmXOHBlH4PLDX+oPpQMKM3KI5+TZEqylW0+tdGuTn7vAgS75uJ8/DnaF+n4cIPMRLBrLk7EHrM9tipWdl/clW3tC5HdJOAIkU2SevUjtcmdoGC3/yY5GcO0yTivXYP60kVUprRPwopKBVYrqO+sFqfWv7NBNe1xQ9zGjEpbznqIT9x3c/rnL1QpKaQ2bmrfdtMWuL/TG1XsDRRvnyy1ZS4Widu7I7BMmoCxVh3MJUrZ9f95YDRbiUky4umsxaDTZPm4ckEVWHHkc7ydvXnJq8Bd5eWJTrrC1aRoWkfYLx+2LtSOGQen0mZTC4aWH4lapWHygQ/wcfHlzaKdHqrf3Yr3ZPFfn9J0dWO6FO3BlcQrTP/9E1ycXGx11Co1Y14eT68db3PLFE+N0Fpo1TrOxp9h6+lNLKzzud0enqyqn68Ri//6lGL+xQn3CGfZP0u5Zbz10O0I8bRIwBEim1jy5MFcsnSGz7m9NxL96lUkDRyKObI4iqcn+k0bMUz/BIxGcHa21bX62e+xUPz8UaWmpoUQH99H6ps6+kpa2/72e0Gs/mnnUsXFZTngpLTviCoxEefln+E6eSJWPz+SO3QmafDwZz7oWBWFNYfOseaPExitN1HjRu0CYXSpWBCd04P7Xj6oIsuPLKVMYDm78kDXIELcQzl/61y6GRy9Rs+XjTYwes8I+v/QG0VRqJy7CovrLsPbOWuv+W3BbrlY0eBLRu0ZxlvftKeAdwSzan5Kh62t7eo1LdACd507U3//mJVHlqFWaQjzCKd2eJ1H3nMzqNwwridfY+K+cWjVOjpHvk0hn8Is+mv+I7UnxJMmAUeIp0D/9XqS3+5utxyl+2ZzhnXV169hueux6vo1FL3eFkAUvXNaKLr7mLjM90JYA4PutO1xZyOr+r+NwYqXl61tldF+n0W6tjUaknv1JblXX9QXzuO8ZhWGie9jzZ2HlDc6ZNqPnLbxr4ssPrgDk2Etak0CilXPuhN1SbVYead6kQce36bwG7Qp/EaGz/3W/q/7HhfumZel9VZm2vbVnvHpyoaUH8GQ8iPsyirnrsrvXQ8SF3dnmeroW2fSHVsjrDY1wmrf93wZtZ1ZX9y0bsysMS9d+d1ttC7UjtaF2t33nEI8Tc/2xy0hHEVyCuj0dx5bLOjXf5VhVf2Wr+95vCltGei/5Slrrlyojh2zq6Pd+UOmpzcXKYbi7Ix+43r7tjd8hTmikC08WXPlQvOv/ZVYmbVtzRNC0juDsYaGoTl2NNM+5DSrorDq0HGMLhtQa9I25arUqSiGLfxw8gxxSaYHtCCEeJ7IDI4Q94hPMbH39A1uJqaSL8CDUnk80agfbe/Lbabqr+KyYC6W0DAUT09cFs5HZbZkWFe37RsMPj6YyldCv2kDut27uLlije351PqNcP5iBa7vjsQYVRPdTzvR7fox0/Mrvr4kd+mOYfJEFLUac2QJ9F+vR/fj99xc8NlDte32Tm8UP39MpcuieHig/elH1OfOYqpa7VFfnqciyWgh0ZiMxjnOrlylNoHmBtcSU/EyaHOod0KI7CYzOELc5e/L8Qxf+COX5y/Be+YUti1ez7jVB0g0mh98cCZuffgJpnIVcBsyAPd3+mAuVpyke66estWdOgvt77/h2akduu+3c+ujqRhr1rE9b6xbH8uYsejXr8WjYzvUVy6TMGbCA/uQOGI0Sb374bJwPp5vvI721/3Ez11odwWVsW59EoePyrRtc7kKaH/+Cfd+PfBs1xL9N1u5NW02xtr1HvHVeToMOg2uOhcsZvtLsxWrFpXFB383/X2OFEI8j1SKksGdxhyEyWSxW6d2JF5eBocdW1Y8ifGbLFYGLtjJ4G3ziYxPu3usAswuXBvr661oXyV/tp7vccj7/2jj3/DnReYc2IbJZS1qp1tYrc6oE+vQMH89+lV7fr5WQ95/Gf+LPH5//6zdIV6WqIT4z5HoBIKjz9nCDYAKeO3UXgb+WemZCjji0TQqlgujpQarD4eQmhKPGjfqRITRpWKBBx8shHiuSMAR4j8mixWDKSVducFsxGix5kCPRHZTq1S0LBlKk2J5uJFkxNNFi4s26/fBEUI8P2QPjhD/iQhw41hAXq7p3OzKvw8qRPGX0n+XkHh+6ZzUBHk4S7gRwoHJDI4Q/3HTO9H4lWKMTHqL1/74jtwpcRwIKMCOEq8y/OWCOd09IYQQD0ECjhB3qRuZixC/Wvzwe35u3kwkb94g3i2eBz9X+cZlIYR4nkjAEeIeRYM9KNqgRE53QwghxGOQPThCCCGEcDgScIQQQgjhcCTgCCGEEMLhSMARQgghhMORgCOEEEIIhyMBRwghhBAORwKOEEIIIRyOBBwhhBBCOBwJOEIIIYRwOBJwhBBCCOFwJOAIIYQQwuFIwBFCCCGEw5GAI4QQQgiHIwFHCCGEEA5HAo4QQgghHI4EHCGEEEI4HAk4QgghhHA4EnCEEEII4XAk4AgA/AM8bD9+oQH4lCqCR8d26L7d+kjt6Vctxz/AAxISAFCfO4t/gAe6bQ/Xnvbnn/AP8EBz5J/Mz7fhK/Srlj9SX4UQQjgep5zugHh2JPXoQ2qjJqjMZtQXL6D7dgseb7YmpXU7EqbNfqy2rYFBxG7ZjqVAwWzqrT39hnWobsSQ2rrdE2lfCCHE80UCjrCxhIZiLlve9ji1xeuYXqmBe/9emCpVfrzwoNfbtS2EEEI8SbJEJTKV0rY9pjJlcflsoV250y978WxSD7+wQHwjwnAb0AdVwq37tpPhElVqKm6D38E3fwi+EWG4vvc/XObNSlvauvf4GzF4dH4Tv/BgfMoWRz13ju059z7d0W/agG7Pbtsym2HShMcfvBBCiOeWzOCIBzJWfxXD9ClgMoFWi9O+X/B6rRGp9RoSv3Apqhs3cH3/PdRxccQv+jzL7bqOHYXzF8tJHDEac8EInFcuR79+bYZ13Qb0IbVVW5Lbd0S/7ktc+vbBqWBRzKXLkjhgCOqLF1DdvEnChx8DYM2VOzuGLoQQ4jklAUc8kDU4NyqzGVVsLEpAAG7vv4upXAVufbrkrjq58GrRCM2Rf7AULvLANlU3YnD5fAmJQ0aS3L03AKZXa+JdrUKG9VObvUbSgCFp9SpXxfm7b9Bv/hpz6bJY8+ZD8fIGq1WWwYQQQgCyRCWyQlHu/DkpCadf95PauBmYzbYfU4VKKFotTocPZqlJpyP/oEpJwVi3/p1ClQpj7XoZ1je+UuPOA60WJX9+1JcuPspohBBCvABkBkc8kPrKJRStFsXbG/X1a6gsFtyHDsB96IB0dTVZDB3qq9EAWH197cqtvn4Z1lc8Pe0LdDpUqalZOpcQQogXjwQc8UC6H7/HXLwkaLVYPTxRVCqSBg/HWLN2urrWoOAstWkNCARAHRODxdvHVq6OuZ4tfRZCCPFik4AjMuW84nO0v/9G/Iy5aQWurpjLlEPz7wnMg4Y9crvmwkVQnJ3Rbd1Mcp/+aYWK8tA3ArxN0WlRp6Y8cn+EEEI4Fgk4DkhRFI5fS+RKfCq5PPXk93NFpVI98DjNuXM4/bo/7UZ/ly6i+2Yz+g3rSG7bntRWbW31EkaPw+u1RtBTTWqjpihubqgvXkD/3bckjhiF5aUCD+6jjy/Jb3TA9aMJoHWyXUWlunULJQt9vZclf0H032xBt2UT1ly5sAYFZ3k2SQghhOORgONgbiabmLbxIKmnTlMw5jzf+IXilj8v/RqVwk2f+dttmDMDw5wZKHo9Vl8/zCVLE790FcY69ht/zRUrEbdhK66TJuDe621UVguWPCEYX62J1T8gy31NHD0OlcmM4aOJoFaT0rIVKW3b4zJ/zoMPvkdyp644/fkH7v17oo6LI3HQMJKGjHjodoQQ2Ue/chke/XqS8lorbs3+NKe7Y0d97iy+ZSO5ueyL+17cIJ5vKkW5+xIZx2IyWYiLS8rpbjwRXl6GDMc29etD5N28hjdP7kYFWFGxqEB1ohu9Rq96kU+/ow/Js0VjMJu4uSHzpar7jf9FIeOX8T8P4/ds2QTdzh+wuroR889JcHHJlnazZfypqTj9eRhLgYIonl7Z0q+n5Xl5/58Uf3/3LNWTy8QdyM1kEyf+vUTr079we5FHjUK7kz/z57GLJKSac7R/99Lu3oXL9E/Q/rAD3Xff4N7rbXQ//Uhyl2453TUhxGNSXbuGdvcujFVfQZ2YgO67b3K6S/b++/qY5y3ciKyTgONAEoxm3E3J6K32QcbFasLVnEKyyZJDPcuY4uqKfusmPLp0wKPTGzj9eZj46XMwNmqa010TQjwm/cavUFksJEycjCU4F85ffZmujk+ZYri+OxKX6Z/gU6wAvi/lwXX0iLQLDrZ/i3fV8vjmzYXHm21QxcXeOTAxEbdhA/GuVBq/sEB8ykbiNnQAqlvxtiqGSRNsX91y949PmWJAxl8fY+vP3Jn4lCiEb4FQ3N/uiOpmnF2/NX//hVf9mviF+ONdtTy67d/iVas67n26Z/OrKB6H7MFxIIFuepI8vDlr8CEs6Yat/ISrP1YPT3wMuhzsXXrmUmWI2/p9TndDPIK7vy9McXbGEhJKSqt2JPfsA07Pxz8r+lXL8ejbg2unLoGb2xNvx7NpfRQf34f6OpPnmfNXX2KKLIGlQEFSmzTHZfGnqOJvonjY39NKv34t5lJluDV9NtrDh3D9YBwoVrR795A49H+oUpJxGz4Y1/fHkDB5atpBSUlgsZI4YjSKrx/qixcwTJ2MR+c3ubl6PQApb3TAGFXTdh5VUhIe3TphyfdSpv3Wb1yHuUhREj6ehvrSJdxGj0DxGkPCpCm2c3u2aoY1IJD4uYtQpabgOmo46ptxWAoVzr4XUDy25+NfIpElTho1TV8pyvi4tnT5/WsKJlzlqHsQC0s3onlUJBr1w1+dJMT9JPXoQ2qjJqhSUtBt+wa3999FZTbZvlJDvLjU59OuyEz83xgAUpu1wDB3Jrotm0ht3c6+sl5P/MKloNFgiqqF7pvNuCyYx41fDmINCwfA6e+/cP5ixZ2A4+9PwkdT7rRhNmMJDce7UW3UF85jzROCNVduu++kc+/aEcVJy62Z8zLvvJMT8Z+ttAV1zfGjOK9baws4ziuXoY69Qdx3O7EG5wLAEp4X77pRj/hqiSdFAo6DiSoUiHvbmnwRFkJ0XCLBPm60qRxB6TyyziyylyU01PbdX6Yq1XA6dgT96pXPfsCxWNJ+xBOjX5f2pbmpTZsDabO15rz5cP5qTbqAY6xcFTQa22NL3nyoY2Nt4eZ2mSrmOhiNoEubidavXonL3Fk4nTqJKinRVldz8l+seULszuEyfQr6LV8Tt24L1sCgTPturFLNbhbSUrAQquvXbOfWHvodc/GStnADpH0n3kNcQSqeDtmD44DKhXozqk1FZvaowchWFSTciKfCXDQyw6/q0G3djFet6viF+ONbND+uY0alfTP9fwyTJuBbKByn3w6k1QsNwKthbdRnz6C6dg2PN9vgFx6Md+WyaH/aade2/osVeDWsjW/BUHwLhOLZrAFOh363q+Pepztetaqj27IJ76rl8QvxR/v7rxmOwWXmNPxC/NF9swUApwP78BrmuSQAACAASURBVGjfCp/IgviFB+H9amX0X36R4bFOJ47h1agOfqEBeFcshW7z1w98zTRH/sGj7Wv45s2Vttek85uooqMfeNyzznndl5iLl0Tx8EB1Mw7VzTiMdeqj/WknqmvX7Oreu2SFVof1no2/ik6HSlHSQgagWr8ej97dMJctR/zCz4jduoObS1akPXfPDT+1P+zAdcIYEsZ+gLl8xl/mm1l/7j23+mo0Vr/0XymTUZnIWTKDI4TIFuqL57GEhtmV6Td8hXu3t0h5sxOJI0ajOXMa1/FjwGolccx4Wz1VcjLuA/uR1LsfisEVt5FD8Oj1NopejzGqFsmdumCYOQ2PLm8Sc/AIeBkA0Jw/R8rrbbCE50VlMqJfuwavJvW4sfMXrOF5be1rzp/FdewokgYOxeofgCU0DPWZ03Z9NXz8IYbpn3Dzs5WY/tu7oblwHlO5iqR0eAtF74x2/y+49+sJajWpzVvaHe/RtRPJnbqQ1H8gzsuW4tHlTWK/24WlWMa3Z1CfOolXw9qYS5bi1qz5YLHg+uH7/J+9+46v+fofOP66++Zm70GGEcTem9hbzVqlVapVWq0OHUq/tFSLLr9qabVmh9rEKrVrllJF7S1BIjs3uev3R+pyJSRiJK738/HweORzPp/POe9zQ/J2Pudzjme/HiSu2QgFWPCyKFAdO4r6n78B8IsMy3Fet2wxxoHP31MbyoULMNWoeWNeDKD5Y2vO606fwmPwADK797znNq+zBgSiOn4sZ1tXZZuZokYSHCFEgSisVjCbURgz0K5djS5mOSnXt/QAsNlwHTOKzB697b+ITAA6HW5vv076K69h88nebFWRkUHq+E8w1W8IQHrsJdzffp20Ee+SMXQYAKkhxfBpVBvt9q3QLftNu/SbtwuxWsmKbob6r73oF/zicE6ZkEDir8uwVKqca19cx41B/900kn5cgKlBI3t5ZpfuDv0x1WuA8uJF9HNn5UhwjH2ftsea1bQF3g1rYfhyMinTZ+be5qQJWAMCSPppof2xi6VCBbzr10S7bg1ZLdvkel9Rp1s4H5tKRfKcn7G5GBzOuY18C/3iBfeebGRkYNPqcrTrIC0Nz/5PYSkeRsrEz++tvZuYqlZHt3A+yksX7Y+p1Hv3oLxy+b61Ie4PSXCEEAXiNvIt3Ea+ZT9Of2GoQ0KgOnEc1flzpHbqAuYbSxdkNWyMwmhEfeSwPaGxabWY6ta3X2MpUfK/a6NzlCkvXbrRxtF/cR03Bs3unSiv3nj0YT5x3CFWS3DIbZMbt/ffRbd0MUm/LM7xCEOReA3DJ+PRrV6J8tJFFP/N3bHcNP/iusx2HW8cKJVktWmPbtmSXNsE0GzemL0FilJp/3wsYRFYQ8NQ/7WvyCQ4KUYz208nkJxhIjLAjUohHijvMLqkW7IQU3RTslq0znHO2LMPrv8bifLcWayhOUd38svWogXaYS9j+Gwipuo10a5bi3bzRodr3Ea/g+roEVK+mm4fUcoOUIe5UpUCt23s3RfDZxPxeKoH6W+8jcKYgeGT8Vj9/LApZdZHUSIJjhCiQNKHvkJmpy4okpNxmT4Vw7SvMEU3sf9iU8THA+DZu3uu9ysvnLd/bXNzy/5Ff91/Ixo2T88cZfw3x0KRmoJnj85Y/QNIHTseS/Ew0OtwH/4yisxMh7buNAFUt2Ip5ipVMVevkeOc+7AX0ezZTdrrI7CUKYfN3R39zBnoVsfkuPbWORhWP3+Ul2Nv264yIR7DlM8wTPks57lc5jIVhr8vJvPN4p3UOrGPoGtxLC1ZjRVR5Xi9UzX0GlWO69V/7UV98gTJb76Ta32ZXbvjOnYUusULyRg2vMBxWQc9T+bho7h8+zUuxkxM0U1J/mYG3m2b269RnTiOwmzG44UBDvdaQsNI+PNggdvGYCDp50W4jxiOx/P9sYSGkTb6A1zHjsLmnr8VdsXDUWQTnGbNmuHq6opSqUSlUrFo0SISExMZPnw4Fy5coFixYnz++ed4enrmXZkQ4r6zFC+OuWp1AEz1GuAdXRfXMaPIat4KFAps3t4ApEz+EnMuoye3zte5W+rdu1BdvEDSr0uxRJaxl9+82NuNwtuPOCTNnY9n3x64v/Q8KVO/u5FoGY1of1tD6keTMPYfeKMqa+57KimvXsXy3yO37OMrWANu/8aOzcsbY7uOGPs+neOc9aZ6CkuW2cr0ZXt4d/0MKqRkj5r1OP8nnye0Y2moLz3rlMhxj7lqda5czuXz/481MIirl24s2JdbouHwmPM/mb2e4srNb1+pVKSNGecwjwtwaDtpycrbdw6whoXniDW3eHK0DVgqViJx5Tr7sfLMaVRnz2CuUPS3w3mcFOnxtFmzZrF06VIWLVoEwPTp06lXrx5r166lXr16TJ8+vZAjFEIAoNGQ9vZ7qP89gnZN9sqwltKRWIJDUJ47i7lq9Rx/bPf4S1xhzB7JsWlvLGCp3rUT1dkzd1WPOaoCST8tRLt2DW5vvnqj/qxMFBYLNt2NuR6K1BS0a3L/xalbedNbU1Yr2tUxmKpXv227WY2jUR85hLlKtRyfjfUek7/74e9LyURcOmlPbiB765fuJ/5g5/7ThRdYEeDyxWR0P89Ds20Lul9+xLNPd6y+fmR1eKKwQxM3KbIjOLlZv349c+ZkrwLauXNn+vXrx5tvvlnIUQnx6LLZbPx7OZUTV9Px0KupFeaV66OH/Mjq0AlzZBkMX31BVpt2oFSSNmYc7kOfR5mSTFbzltg0WlRnTqNdtYLkGXPAYMi74tsw1aiF1dUN99eGkf7SK6guXcQw8aNc58fkxVy9Jsnz5uPZqys2N3fSxozD5uGJqVp1XCd/jM3NHZRKDFM+xebuiSI15yiFfu5sbBot5qgoXObMQnXqJCnTvr9tm+lvvoNX62Z49OmOsU8/bD6+KC9dRLtpA8ZeTzlMdi4MWRYrrlnGHOWu5kwyzdZCiKgIUShwnfQxyrhL2LQ6THXrkfa/cdjcPfK+Vzw0RTrBGThwIAqFgp49e9KzZ0/i4+MJCMh+lh4QEEBCQsId71epFHh5FfwHaFGmUimdtm/5If2/9/5nmix89OtuEv4+Qo0zf/OPdxALIyrwdr9GlAvKey6BwUWL/tYY3n4HzcBn8T6yH1vdetC/H5YgP3QfT0D/09zsBd1KlMTarh1eAV6gVqPUa1AoHP+tKtyyR03c3fX2V8JvblehUuJZJgLrzz+jeestPJ/pDaUjsUydinLyJDQalb0+lVYNuXxeiv+2LvHyMoCbAdq2xDL/V1y6dUXn54111GiY9yOKIYPxePkF8PXF+uIQSE9H8fVUe33X67H99COGN15HMeEDKF4cy7yfcGtUz96eSq2Em+KiRhUsW7eheX802jdegYwMKFYMa9NmqKpUyNHvmz2Mv/91y6iZFxJJst4Vb1OGvXxTSHlqViheqP/+Cv3f/+j3sI5+j+tpnhJ4mLNv7rr/cXEoJ32CMmYlnDubvZBhmbJYe/XCOmAgPKSpHsq33kS5aBHmYyceSnsKm81meygt3aW4uDgCAwOJj4/n2WefZdSoUbz44ovs2XNjga5atWqxe/fu29ZhMlmcdkt5Ly+D0/YtP6T/997/X3aeInn+Ql4/sAIl2T8G9niF8X+tn2PSgMaoi/DWHvL9fzj9X7b3HNvX/EHPA78RkJnCrsAy/F61Ke/2rk+Qh/6Bt3878v3Pf/9Vx47i2bUDuLiQMWgw5qgKkJWFZvdOXGbNIKtJc1Km5j6v7H5zfX8kuhVL722SN+Dvn790ssiO4AQGBgLg6+tLy5YtOXDgAL6+vly+fJmAgAAuX76Mj49PIUcpxKPrj79OMf7oZntyA1Az8SwBl85wODaFSiEy3P64e6J6KKEBbdj4ZyQpKRmUKhnM6CrF8XMtWhv3ittzHzwQm48viSvWODxCMzVrQcaQl9H+tqYQo3uwiuQk4/T0dFJTU+1fb9u2jcjISJo1a8aSJdnrSixZsoTmzZvfqRohxB0YTVY8TDnnWHgYU8kwyV5NIlu14p682qk6o/o2oE/9kpLcPEI0f2xF8/d+0kb9L9f5QTZ3jxwLVmq2bMKrTdPsrVXKl8JtxHD47/fxdcozp/F4uje+JYtlbzHStwfKk46PnRRJibgPHoBfRBA+FSMxfDbx/ncwD0VyBCc+Pp6hQ4cCYLFY6NChA40bN6ZSpUq8+uqrLFiwgODgYL744otCjlSIR1dUuB9b/CNpF3tjuDhR48KhgJL0C3ArxMiEEPeDZvs2bGq1w4KZd6L69wievbqSFd2U5O/noLxwAdcP/4fqzGmSflmcfVFmJl7dnsCmUZP66ZfYVGpcJ47Hq3M7rm3ajs07+8mK+7AhaP7YSuoHE7AGBOIy9UtUp085bGT6oBXJBCc0NJRly5blKPf29mbWrFmFEJEQzqdrw7JMPNuZtL16al49Razegx8rtaJF/Sg8XTSFHZ4Q4h4pY2Oz11TS3zJfymKB69NvFQr7bu6GyROwFA8lec4v9jKbtzceg/qj3r0Tc6066H+ai/LCORK277Xv92auUROfWpXRz/6BjFdeR3XkMLpVK0ie/gOZnbsB2bvG+1Yv/1DfNCuSCY4Q4sEL83bh3X6NWFk+lA2n4/ByN9ChZmlqh8vu80I4BZst10Uu/UoVR5GeBoDVx4f4I6cB0OzdS2bHTvbkBiCzQydsajWanTsw16qDet+fmCtXcdjM1hpSDFPtumh2bieD7BWtATJbt7vRqJsbWdHN0Oy98aLQgyYJjhCPsWAPPQOblgPKFXYoQoj7zBocjDL+KmRmwk0LViYuXw0WC/o5M9HF3Hhaorwcm3NbE5UKm7cPysTsFahVcblcA1j9/VGdO/dfPXFY3dzBxcXhGtst25k8aEVykrEQQggh7o2pXgMUZjOabZsdys2VqmSvmB3ouJWINSDIYdNaACwWFNcSsHplb71iCczlGkB55QrW/7ZnsQYEokxNyV7b6SaKq1fvtUt3RRIcIYQQwgmZ6jfEVKkKrh+OQZGakvf1NWqgXbk8e47Of7Qxy1CYzZjq1AWyV/1W7/8L5ZnT9muUly6i2b0TU53shS2v71Gnu3lbk9RUtJt+vw+9yj95RCWEEEI8Iu52bd6Ub2bg2bUDXs0bkTFoMJZy5cFiQXXyBLqli7C53nhjMn34CLybN8Tjmd4Y+w9EefEirh+8T1bT5phr1QHA2OspDFM+x7N3N9LfGolNqcJ10kdYfXwxPv0sAJZyUWS2aYfbiOEoUlKwBgbi8tWX2Fwe7urTRXYl4/tBVjJ2XtJ/6b/0X/r/ONl95hrLtxzibHw6AR46oquXom3lEJS5TCK+lSIuDsP/fY527SpUFy+ASo25VGmyWrUh47nB2HxvbHyr2bwR1/FjUP9zEJu7O5kdOpE6+gNwu5EIKU+fwm30u2i2bEJhs5HVoCFpH3yEpWTpG20mXsNtxHB0a1djM7iSMWAQipSUh7qSsSQ4j6jH8R/4zaT/0n/pv/T/cbHrTAK//LKRobsXUSXxPOfd/fimQmtC2zSlT4PSeVfgZPKb4MgcHCGEEKKIstlsLNl4iFd2LqB64jlU2CiZHs+7e35l657jJBtNhR1ikSUJjhBCCFFEmSw24q6lUin5gkO5uzmTiKRYzl3LuM2dQhIcIYQQoohSqxTodRridI6PZSwoiDV44W2QvcFuRxIcIYQQoohSKhQ0qVGK6RVbY1Rmv/hsRcHPJeriX6I4IZ76PGp4fMlr4kIIIUQR1qVGGD8kt2BgQEnKXjvPRc8A3MKL81LbKoUdWpEmCY4QQghRhKlVSga1iOJK3ZKcS8ygRJAn3uq8Xw9/3EmCI4QQQjwC/N10+LvpHrvX5AtK5uAIIYQQwulIgiOEEEIIpyMJjhBCCCGcjiQ4QgghhHA6kuAIIYQQwulIgiOEEEIIpyMJjhBCCCGcjiQ4QgghhHA6kuAIIYQQwulIgiOEEEIIpyMJjhBCCCGcjiQ4QgghhHA6kuAIIYQQwulIgiOEEEIIpyMJjhBCCCGcjrqwAxDiURQw1SPX8giPEuzqu/8hR5M/S48vIsOcQa9yTzmUd17SDh+9L9+3mVMocf18ZB7Dfn+Rk4Mu4qZxK5QYhBDORxIcIQroxSov07FUJ4cynVpfSNHkbenxxSQY43MkOB83/hSNUn4UCCGci/xUE6KAwjzCqBlUu7DDuGdlfcoVdghCCHHfyRwcIR6ApMxEqswqx9B1zzuU91vZk7rzqpFuSreXxZxcTusFTQibFkDZGeH0XtGNcyln7ecPxx+iz4rulPg2hBLfhjBwzdPEpsbaz2+7sIWAqR5sOLuep2KeJGJ6ENVml2fmwRn2a15eP5gVJ5fyx8WtBEz1IGCqB5/sGg9kP6IasLqfQ5xbzm+izYKmhE7zp/wPpRixaTipptQcbW67sIWBa54mYnowNedW5vuD3zrUszt2J/1W9qTSzDJETA+i6S8NWHD0l3v4ZIUQIn8kwRGigKw2K2ar2eGP1WYFwFPnxWdN/49fj/7MypMrAPjp8Fx+O7OGL5t9g0FjAGD+vz/x7OqniPAowbetZ/FFs68p5VWa+IyrAJxMOkGHxa3ItGTyVfPpfNnsa/5NOEyXXztjs9kc4hm+4SXK+1bk+zZzaRbWkhGbh7P29CoAXqs5gobFGlPJrworu65jZdd19C3/TK79+jfhCL1WdM2el9N6DiNqvcuiYwsYeEsSBPDaxpep4FuRmW3n0SCkIW9vfp29cXvs58+nnKNWUF0+azqFOe1+oUOpJ3jl9yEsOvbrPX76QghxZ/KISogCGrn1LUZufcuhrGfZPkxp/g0AzcJa0K/8s7yx6RVC3UMZte0dhlQdRu3gOkB2gvThjv/RrkRHprX6wV5HmxLt7F9P2j2BAEMAP3VYiFalBaCCbwXq/1STdWfW0DKijf3a5uEtGVn3fXvbZ5JP89mfE2kV0ZYSniXx0nljtVnzfKw2ec8EiruHMqfdL6iUKgC89d4MWtuf3bE7qRVUx35tl8juvFZzBAANQhqx9vQqYk4up3pgTfv562w2G/VCGnAx9SJzD82ia+ST+fmYhRCiQCTBEaKAhlZ9hU6luziU+eh9HY7HNBjHpvMbaLeoBSU8S/JW7ZH2c8evHSM27RK9o/reto3N5zfSs2wflAolZqsZgDCPCCI8I/jryj6HBKddiQ4O97Yv2ZGRW0dgsVrsiUp+7L28l44lOznc06FkJ9RKNTsv7XBIcJqENrd/rVFpKOFZioupF+xlicZrfLJ7PKtPreRS2kUsNgsAwa4h+Y5HCCEKIs8EZ+7cuXTs2BFPT8+HEY8Qj4zi7sWpGlD9jte4adxoFd6a7/6eRp+ofuhUOvu5hMwEAAINgbe9P8EYz5R9nzFl32c5zt2cSAD4ufjnODZbzcQb4wkwBOTZn+sup8Xif8v1KqUKb50PicZrDuWeWsefC1qVlkxLpv142O8vsiduN6/XHEEZ73K4a92ZeXAGq0/H5DseIYQoiDwTnCtXrtC9e3fKly9Pt27daNSoEQqF4mHEJsQj76/Le5n5zwwq+VXhsz0T6RL5pD2h8dH5ABCXHnfb+7103rQr0ZG+5Z92KHd316MxuTqUXc24kuNYrVTje8uoUl4CXINy1GWxWriWmYCX3jvf9RjNRn47s4aPGk2if8WB9nIr397hLiGEuD/ynGQ8fPhw1q5dS/fu3Vm8eDGtWrXi008/5ezZs3ndKsRjzWg28tL6F2ga2pzlXdbgpffmjY3D7OdLe0cS7BrCL0d+vG0djYtHcyThEFX8q1E1oLr9T43gmoR5hDtcu/LUCsfjk8up4l/V/qhJq9KQaTHmGXeNgBqsPLkci9ViL4s5uQyz1Uyd4Lr56jtAliUTi83iMGqVmpXCmlMr812HEEIUVL7m4CgUCvz9/fHz80OlUpGUlMSwYcOoX78+I0aMeNAxCvFA2Ww2Tsank2Q0EeFjwMegzdd9Z5PPsid2l0OZQqGgRmAtAD7a+QGX0+NY+MRyDBoDU5pN44klrfn5yDx6lXsKpULJ6HpjeXHdcwz+bSBdI7ujQMGWC5vpGtmdqgHVebPWO7Re0Iw+Md3pE9UPH70vl9Iusj1uC11L9qRBsUb2ttef+Y3xO8ZSL6QBMSeXs+n8Bma3/dl+vrRXGVafWsnKkysIcQshyDWYINfgHP0aXnMEzec35JlVvelfcSAXUy/ywY73aRra3GH+TV48dJ5UC6jO5D0f46Z1R6lQMmXvp7jrPEnNSs53PUIIURB5JjizZ89myZIleHt70717d0aMGIFGo8FqtdKqVStJcMQj7UpqJmPX7uNU0jlsigRU1lDalyvNoHqRKPN4FPv1/il8vX+KQ5lKoeLSi9fYeWkH0w58xVfNpxPoGgRA7eA6DK7yEu9tfZvGxZsQ4laMbmV6oFPp+XzvJAaueRqD2kCNwFr4uvgBUMorklXd1vPRzg94feMwjGYjQa7BtCiZPWn5Zp82ncL0A1OZduArvHTeTGg82eGNrGcrDuLvqwd4dcMQEjMTeaPm24yo/W6OfpXzieKnDgsZv2MMz67ui7vWnS6luzG6/gd3/fl+3XIGb2x8hZfXv4C33ocBlZ4nw5TO9wen33VdzuB2W3ws67KGusH17mtbi48tIMuSRc9yfe5rvXequ8OiVoS4hTC91cz73qYQd0thu3UxjVt88cUXdO/enWLFiuU4d+LECUqVKvXAgrtXJpOFxMT0vC98BHl5GZy2b/lxP/pvs9l4efEO/kn5BZXLPhQKsFm1qNKeZGitDjxRqei+6XNz/7dd2EKXpe3Z1HMHUb7lCzmyh+NR/fsfMNUj1y0+yvlE4aZ1z3c9+en/06t6k2ZKY+ETywoUa0Hq/jfhCBqVhpKeD/b3wqP6/b9fHvf++/vn799KniM4r7zyym3PFeXkRoi8HL+axumks6hcs5MbAIUyC5N+DQsPli3SCY54dDnLFh+5kW0/RFEiKxmLx1Zihgmb8iq3PolSqhJJTM8qnKDEY+3fhCMMWtOfqrOiCJ8eSOOf6/DdgW8cVq3edG4D2vFqdlz8gwGr+xExPYiacysz65/v7de8+NtzrD4Vw5bzG+1bc3y65xMA1pxeRbdlTxD1fQlKfluMtgubs+ncBoc4zqecY8DqfkR9X4KwaQHUnlvFvrXHnerusKgVz6/t71DXwat/02dFd0p9V5wS34bQdmEztpzf9CA+PiEcyEJ/4rEV4WNAaQnHZlWjUJrt5easMEr75z5XoihqUKwRl4fIpN1HxfUtPq5ToLC/6XYx9QKR3mXoXrYnbho3DlzZz/idH2C0ZPJSNcfR9Fc3DKVXuad4usKzLDj6C29uepVqAdWp7F+VEbXf5WLaBYzmDMY1zE4+irkVB+Bs8mnaRrTjpaqvoFAoWHdmDT1XdCGm62/2CfJD1g3CYrPwadP/w13rzpmk05xMOgFwx7pvdSThMB0WtaKsT1kmRX+Ol86b/Vf2cSH1/H38RIXInSQ44rHl76ajZelSrDjeDbN+LUp1EubMcDxMXXm2VlRhhyec1K1bfNQOqsuKrmsBaBrWnKZh2atD22w26gTXI82UytxDM3MkON3L9OTVGm8AUD+kIb+dXk3MyWVU9q9KCc+SeOq8UCs1OR6HDar8ov1rq81Kw2KNORx/iB8Pz7EnOPsu/8nMNvNoHt4KgIbFGtvvuVPdt5q4+yO89d4s7bwavVpv76MQD4MkOOKxNqRhOUI83Vl8MIqkZDNl/N3pX6s8FYMfnREc8Wi5dYsPN82NCZMZ5gy++HMSC4/9ysXUC5isJiB7lMdqs6JU3JhV0CS0mf1rrUpLhGcJLqZezLP9CynnGbdzDFsvbCYuLRYb2Y+/rm+jAVDRrxJjt7/PlYwrNCoWTTH33Edo8rL1/CaeKv+MPbkR4mGSBEc81tRKBd2rFKd7leLYbDZZpVs8cHfa4uN/f4xk/r8/83rNt6jkVxlPnScrTizjy32fkmXJckgUPHVeDvdqlNo8F3K0WC08tbIHmRYj79QeRbhHBAaNgXE7xpBqSrFfN6P1HMbtGMN7W98mOSuJSn5VGNNgnMNITn4kZibecSsSIR4kSXCE+I8kN6KwLT+xhOcrD3Z4HLX61P3bt+t44jEOxR/k145LiQ5tai833pIYhbgV46sW07FYLey9vIePd42n38pe/PX0oRyJ1Z146bzuuBWJEA+SvEUlhBBFRIbZiPamrS0sVgtLji8qUF1apZZMs2PiYjRnADhsn3E66RR/xu3OtQ6VUkWtoDq8UfMt0kypnE85f9u6c9OoeBOWHFvosAGrEA+LjOAIIUQBnElIZ93Ri8Snp1OtWBCNSvqi16juqc7o4k357sA3hLmH46nzZMbf0zHfNDfmbpT2jmTdmbWsOhVDsGswwa4hlPWJIsg1mFHb3uGt2u+SnJXMx7vGEeJ6YyHXBGM8fWN68mTZXpTyKo3RnMHUv6YQ5BpMae/I29Z9fcXum42o9S6tFzal85K2vFB5KN56Hw5c3Y+/iz+9yj1VsA9JiHySERwhhLhLvx+9zMtL1jP3yDesPD+FSX/8zCtLdpBsNN1TvR9Hf0qtoDqM2Pwawze8TEW/yrxU9faLrd7JwIov0Di0CcN+f5FWC5ow7/Bs9Go9P7SZC8CA1f34ZNd4Xq/5lsMeYy5qA2V9yjH9wFT6rezJK78PwV3rzvyOS+wjP7nVnZsyPmVZ1mU1HlpPXt3wEs+u7svKk8sp7h5aoD4JcTfy3KrhUSZbNTgv6b/0v7D6n5pppt+Pv5Osn4JSfWP9IUtqU3pH9WVQ3dIPPAb5/kv/H+f+53erBhnBEUKIu3DgYjJGxVGH5AZAod/LxuNnCykqIcStJMERQgghhNORBEcIIe5C5RAP9LYyWM2eDuU2Y3Walg4rpKiEELeSBEcIIe6Cm07NsIZVcc0YhDmtNqaMMihTu1HGcdAwCwAAIABJREFUrSU9qoYXdnhCiP/Ia+JCCHGXmkYGUNK3OeuPRhGfkUG1kEAa3ofXxIUQ948kOEIIUQDhPgYGPIQ3poQQBSOPqIQQQgjhdCTBEUIIIYTTkQRHCCGEEE5HEhwhhBBCOB1JcIQQQgjhdCTBEUIIIYTTkQRHCCGEEE5HEhwhhBBCOB1JcIQQQgjhdCTBEUIIIYTTeeQSnM2bN9O6dWtatmzJ9OnTCzscIYQQQhRBj1SCY7FYGDt2LN999x0xMTGsWLGC48ePF3ZYQgghhChiHqkE58CBA4SHhxMaGopWq6V9+/asX7++sMMSQgghRBHzSCU4cXFxBAUF2Y8DAwOJi4srxIiEEEIIURSpCzuAu2Gz2XKUKRSK216vUinw8jI8yJAKjUqldNq+5Yf0X/ov/Zf+P64e9/7n1yOV4AQFBREbG2s/jouLIyAg4LbXWyw2EhPTH0ZoD52Xl8Fp+5Yf0n/pv/Rf+v+4etz77+/vnq/rHqkEp1KlSpw+fZpz584RGBhITEwMkydPLuywnM4nu8Yzac+EO15TP6QhSzqvfCDtH7t2lAY/1eTXjkuJDm0KQHJmEq9ueImtFzaRmJnInE7z2HV2DwuO/sL+Z448kDiEEEI8uh6pBEetVjN69Giee+45LBYL3bp1IzIysrDDcjp9yz9Ds7AW9uPv/p7G1gubmdlmnr3MXevxwNoPdQ9jZdd1lPUp5xDDxnO/83/NpxFgCKBaWCUqelSjc+muDywOIYQQj65HKsEBiI6OJjo6urDDcGohbsUIcStmP15+Yik6lY6aQbXvS/0WqwWrzYpGpclxzmg2olfrc7R17NpRyvlE0a5kBwC8DAZUWS4Ucy9+X2ISQgjhXB6pt6hE0fTDwe9o8GNNin/jR805lZi2/yuH88+v7U+HRa1YenxR9nXT/Dh49QAzD84gYKoHB678RcfFrQmbFsB3f0/j2LWjBEz1YNO5DQBU+KE0C4/NZ0/cLgKmehA6zR+AD7a/T5VZ5RzaupJ+hVd/H0qFH0oTOs2fBj/W5PuD3z6cD0IIIUSR8ciN4IiiZfKej/lsz0Rerj6cusH1+TNuN2O3j8ZV40bf8s/YrzuZdJyPd43j9Zpv4aP3JcS9OPuv/AXAoLX9GVBxEG/VHomXzjtHGz91WMDY7e9zNeMKk6I/R6nIPS9PM6XRaUkbUrJSGFH7XUp6luJE4nEupl54MJ0XQghRZEmCIwrsmjGBL/6czIjaIxlWfTgA0aFNSclKYfKejx0SnPiMeJZ0WkUZn7I56hlSdRjPVBhgPz527ajD+cr+VfHRe2M0Z9zxMdm8Q7M4lXSSTT132NtpVFweZwohxONIHlGJAttxaTtGi5EnSnXGbDXb/zQqHs2F1PNcTr9svzbcIyLX5AagZXjr+xLP1gubqR5Y87btiMIRMNWDGX9PK+ww8qXc9xF8smu8/bjzknYMWN3PfvzJrvGU+z4iz3pqzKnI+9tGPogQhRD5JCM4osASMuIBqD2vSq7nL6aeJ8CQvU6Rv+H26xX5u9z+3F3FY0wg0BCU94VC5NPHjT9Fo5Qfk0I8iuRfrigwL332fJn5HZfgpfPKcT7S+8ZIioLbrzh9p9Wo74aP3oe49Ni8LxROKcOcgYva5b7WefNSBcK5XU6/zJR9n/Hb6dVcSD2PSqGmtHckXSOfpG/U03joPB9aLLP/+QE/F3/7W6OiYCTBEQVWJ7geWqWWy+lxNAltVtjh0Kh4NKO3vcuxa0eJ9C5T2OGI2+i8pB0+el/al+zIx7vGcTXjKrWD6/BpkykOyxNkmDP4ZNd4lh5fxOX0OIJcg+lcuhvv1fsfkP0YqH3JJ/DUeTL7nx+4knGZi4MTgOzHpx/tHMtfl/eiV+lpX/IJxjYYj5v2xgqo2y9u450tb3Ii8RhlfaIY33DibWP9vs0ch/Kdl3bw7pY3OXrtCKW9yvBR40nUDa53x37nJyZROI5dO0rXpR1wUbswqPJgonwrkGXJYnfsTv5v32ccvHqAqS0e3tuYcw7NdFgWQxSMJDiPCavNxtmEDExWKxE+BjSqe59+5efix6s13mDEptc4lXSSOsH1sFjNHE88xu7YXXzXetZ9iDz/+kQ9zax/vqf7sid4veZblPQqxemkU5xLOcM7dUY/1FjEne2N20Ns2iXGNBiP0ZzByK1v8frGYfzUYSGQve/c0yt7sSduN6/VHEEV/6pcSr3IzkvbHepZdOxXyvpE8XHjTzFbzUB28tF9aUfalujAjNazSTAm8OGO/5GYmWhPVGLTLtF7RTeqBdRgRuvZxKbFMmTdc2SYM/KMPcOcwdB1gxhW/TUCXYP4+q8p9F7RjR1P7SPQEJjrPfmJSRSewb8NxEfvy4quaxwWMW0W1oIhVV/mtzNrbnuvzWYj05KJXq1/GKGKuyAJzmPgxNU0Plq/j7j0S6Aw4aoKYljDKtQv4XvPdb9R622KuRXn27+/4f/2fY6L2oVSXpF0i3zyPkR+d1w1rizutJIPdozmo51jSTOlEeoexqDKLz70WMSdpZhS+LH9r/bHnJfT4xi17R37Y6YN59az6fwGZrf9mTYl2tnv61muT4665rWb7/DL5cMd71MrqA7ftp5pLwt2DaHbso4cjj9ElG95pu2fik6lY177XzFosjctNGgMDFk3KM/YM8wZvFNnFN3K9ACgQbFGVJ9dnun7pzKq3phc78lPTKJw/HFhK39f3c+P7X/NdYV2d60HXW/6efbJrvF8f3A6M9v+xOitb3Mo/h8+bTqFHmV7c82YwIc7xrDq1ApSspKp5FeFDxp+RI3AWvb7p/41hSXHFnAi6QQ6lY7qATUY2/AjSnqWArJHDfdf2cf+K/v45d8fAfiy2df0KvfUA/4knI8kOE4uNdPMyFU7iWU2atdTKBSQZvJjwob+fOrejNJ+rnnWMabBOMY0GHfb872j+tI7qu9tz09vNTPX8v4VB9K/4sAc5ZHeZbg8JDnPOkbVG5PjF4q/wZ8vm31921hE0VA1oLo9uQEo45091+VS2kVKepZi64XNeOu8HZKb3DQqFu2Q3KSb0tkTu4vxjSbaR3Qg+3GqRqlh/5V9RPmWZ9/lP4kObWpPbgDal3wCyDvBAWhXsqP9azeNG9HFm7Hv8p+5XpvfmETh2H5pG2qlmobF8r+kRIY5g5fXv8BL1V6llFdpggzBZFoy6b6sE0lZSbxf7wP8Df7MPDiD7ss6OYzuXUq9wIBKzxPqHkZKVgqz/plBh0Wt2NFnLx46Tz5u/CkDVvcl3COC12qOACDCs+QD6buzkwTHyf1xKoEE8140bqfsZSrNVZJNa1l+sBTDm8gPVvHweWodJ2xqVVoAMs2ZQPYaS4Gueb8Rd+vbeUmZiVhsFt7a/BpvbX4tx/XXF328nB5Hed8KDudc1C64atzybNNV45ZjMrOfwY9D8QdzvT6/MYnCEZsWi4/eN8cjJovVgg0bkP2ShEqpsp/LMGcwtsFHtC3R3l4279BsjiQcYkuvnZT0Kg1A4+JNqfdjdb7+awr/q/8hAB80nODQRnRoU8r/UIpVp2LoWa4PZX3KYdC44uvid9+2x3lcSYLj5C6lpJHJObS3lCtU8ZxPTiqUmITIi7feh7i0vN+Iu/XtPA+dJwoUvFnrHVqEt8pxfZBrMAABhkCuZlxxOJdhziDNlJpnm2mm1BxvbF1Nv3rbhCy/MYnCYbPZcn3Ls9R3xUk3pwHZb2geGXDafk6BguZhLR2u33x+A5X9qxLmEeEwUlc/pCH7L++zH++J3cWEXeP4+8pfXMu8Zi8/mXT8fnVJ/EcSHCdX2tcTA1GY+cvxhDmMqAD/wglKiDw0KhbN/+37nLWnV9Eqom2+73PVuFIjsBbHE4/xRq23b3td1YDq/HR4DummdPtjqpiTy/LdzsqTy+1zcFJNqWw6/zv9yj97TzGJwhHsFky88SqZlkx0Kp29fHmX1VhsFuYcmpnj74aXzss+6nhdvDGBP+N2E/KNT442IjxKAHA+5Rw9lnehWmANJjX5gkBDMFqVhj4xT2L8b/RS3D+S4Di52uFeFHetxPHUU6gMf4LCgtlYGn9FczqUl524xd05GZ9GzKGzXEhOpnxAAO3KF8PP9dbxwXvXJLQZTUObM/i353i95ltU9q9CXHos2y/+weQmX9zx3tH1P6D70o4MUSjpWKozbho3LqSe57cza3i3zihKeUXyQuUh/HDwW/qu7MHgKkOJTYvly72f5msdHRe1Cx/t/IA0UxpBrkFM/WsKWRYTz99hMnt+YhKFo15wA8xWM9subKbZTaMylfyzFzBde3p1jntyW7vLW+dNVf9qfBL9WY5z2v8Sp9/PriPDnM7stj/hqsme/2i2mkm8aSRH3D+S4Dg5jUrJxx1q8e0OD7adao7VZqV8oA+D61cgyENeaxT598epeCZs3E6KYh2or7L7SkliDkczqWN9Qr3v7wJ7CoWCmW1/ZMLOD5l+YCrxGVcJcg12eJvlduoG12Npl1V8sms8Q9c9j9Vmobh7KE3DWthXzQ52C+HH9gt4d8sIBqzuR6R3Wb5q8S3PrOqVZ/0uahf+r/k03tnyJseu/Utp7zL81GHBHecM5ScmUTjqF2tIJb8qfLhjDLWD6hZ4XaJGxaPZuON3irmF4m/IfXTcaM5AqVCivml17KXHFzk80gLQKDVkWowFikPcoLDZbLbCDuJBMZksJCamF3YYD4SXl+Gu+2ayWLFYbeg1qrwvLuIK0n9n8rD7b7JY6TtvA1fUX6HSxN8oTy9Po4Dn+KBtjYcWC8j3X/p/+/5brDYOXkomyWimlJ+BYp55J9/XF/ozaAwMqjSYcr7lsVgtnEw6wXcHvsFoMfJnv+xJ5NdfE795Tg6A0Wyk/aKWZFqMDKk6jHCPCBKMCey7/CcBhgAGV3mJQ/H/0Gx+AzqX7kqfqKf5N+EwU/+aQqoplT7l+tnfVn3l9yFsOLeez5t+hY/ehzCPcHz0N5b1eNy///7++UtCZQTnMaJRKXGC3EYUghNX00i3XnJIbgDULv+y70I8ZosV9X1YPFKIe3E+MYNRq/dwxXgKs+IKGktpokuU5tXo8qiVt98SJtK7DOt6bOH/9n3Ot39/w8XUC6gUakp5leaJ0l14rtLgPNvWq/Us7rSCj3eN45Nd47mScRk/F3+qBdSgdUT2cgflfSvwRdOpTNozgZUnV1DBryLftZ7F82sd528Nr/Em51PPM2htf1KykmUdnAKSEZxH1OOewUv/H27/j19JY/iK5RjdHNcYslm1GFJHsPjZFqju8AvkfpPvv/T/1v5bbTaen7+N45mz0LgcBcBmU6FM7coLNTrTrYrzzDl83L//+R3Bkf9yCSHyVNLPgK8+BHNmqEO5Jb06DUsEP9TkRojcHIlLJTb9jD25AVAoLJj161n6j7yC/TiSBEcIkSelQsHbzarhZ+kPqW3ISquMKrUHpQwdea6ubGwqCl9KphmrMufbSEpVKslGUyFEJAqbzMERQuRLuUB3vusRzZaTFbiUkkYZP2/qRnijU8v/k0ThK+3nispSgiyrFoUyy15uzoygQpD3He4UzkoSHCFEvnm6aOhQQVbeFUWPr6uWjlFlWHS4J5m631CqkjBnRuBj60T/WuUKOzxRCCTBEUII4RSeq1uacG8PFv5dlsT0TMoH+dK3Rpl8bSosnI8kOEIIIZyCUqGgdblAWpcLLOxQRBEgD8+FEEII4XQkwRFCCCGE05EERwghhBBORxIcIYQQQjgdSXCEEEII4XQkwRFCCCGE05EERwghhBBORxIcIYQQQjgdSXCEEEII4XQkwRFCCCGE05EERwghhBBORxIcIYQQQjgdSXCEEEII4XQkwRFCCCGE05EERwghhBBORxIcIYQQQjgdSXCEEEII4XQkwRFCCCGE05EERwghhBBORxIcIYQQQjgdSXCEEEII4XQkwRFCCCGE05EERwghhBBORxIcIYQQQjgdSXCEEEII4XQkwRFCCCGE05EERwghhBBORxIcIYQQQjgdSXCEEEII4XQkwRHiHpxLOcuQdYOoNrs8odP8qToriqdX9mL7xW2FHVquAqZ6MOPvaYUdhhBCPHDqwg5AiEdVovEabRc2J9AQxMi67xPkGsy55LOsPr2S3bG7qBfSoLBDFEKIx5YkOEIU0PKTS7mSfpkNPf7A3+CfXVgMekf1xWazFW5wt8gwZ+CidinsMIQQ4qGRR1RCFFBSZhJalRZvvXeOcwqFwv515yXtGLC6n8P5bRe2EDDVg8PxhwA4m3yGgKkeLDw6nyHrBlHi2xDK/1CSibs/ylH34fhDdPqlIyW+DaHEtyEMXPM0celxOer+/ew6+q3sScT0YN7Z/Ib9fJbFxMgtIygzI4zS34XyzpY3yLJk3fPnIYQQRYkkOEIUUGX/KmRaMhm6bhD7L+/DarPec51jto/CRW1gRuvZ9I3qz6TdE5jx93T7+ZNJJ+iwuBVGs5Gvmk/ny2Zf82/CYfrF9MgxajR8w0tU8K3InHY/0yfqaXv51/uncDHtIl+3+I7hNd5kzj8zGb9z7D3HLoQQRYk8ohKigBoXb8ILVYYyff9UFh9fiJvGnejQpvSvMJDo0KYFqrOsdzkmN/kCgGZhLbiacYUv9k7m2YrPoVQombR7AgGGAJb3iiE9xQxABd8K1P+pJuvOrKFlRBt7XU+U6szbdUblaMNN48aM1rNRKpQ0D29FliWTz/dO4pXqr+Gt9ylQ3EIIUdTICI4Q9+CDBh+x/am9vF/vQxoUa8iGs+vosbwzMw/OKFB97Up2dDhuX7IjsWmXuJh6AYDN5zfSrkRHlAolZqsZs9VMmEcEoe5h/HVln8O9LcJb59pGmxLtUSpu/NNvX/IJMswZHEk4XKCYhRCiKJIRHCHuUUnPUgytNoyh1YYRnxFPj+WdGb9zDM9UGOAwFyc//Fz8cz2OS4+luHsoCcZ4puz7jCn7Pstx7/Uk6Dp/Q0D+2jD4ZbeRFntXsQohRFEmCY4Q95Gviy+9yz3Fu1tHcCXjCgGGAHQqHSar4yTea8Zrud5/NeNKrseBhiAAvHTetCvRkRfrPE9KitHhWh+9r8OxgtyTqxxtpF/NbsM16E5dE0KIR4o8ohKigK5mXM21/GTSCXQqHR5aDwBC3Ipx7NpRh2s2nd+Q670rTy53OI45uZxAQxAhbsUAaFw8miMJh6geVIOqAdUd/oR5hOcr7tWnYhwmRMecXIaL2oVyPlH5ul8IIR4FMoIjxE2upGaSkmmmmKcLOvWd8/9fjvzIwmPz6VG2FxV8K2GymthyfhM/HPyO/hUHolfrAWhXogPzDs9m1Na3aRHemm0XtrDh3Ppc6/z32hFe3/gKHUo+wY5L25h3eDbjGn5snzPzZq13aL2gGZ3md+TJ0n3w0ftyKe0im85toFe5p2hQrFGefUw1pTJwzdP0K/8MRxKOMHnPxwyo9LxMMBZCOBVJcIQArqVnMXHj3/wdexGFMg2dwo9napanQ4WQ297TIrwVZ1NOM/fQLC6kXkClUBHhWYLxjSbSr3x/+3UtI9owss77/HDwO+Yenk2biHZ82GACT6/qlaPO0fXG8tvp1QxY0w+9SsdrNUcwsNIL9vOlvCJZ1W09k/aO5/WNwzCajQS5BtOoeBNKeJbMV19frPIyZ5JP8cJvA7HarDxV/mlG1n0//x+WEEI8AhS2orbk6n1kMllITEwv7DAeCC8vg9P2LT/uZ/9tNhvDFu/kYPIiVIbdKBRWrGYPdBl9GNWsNfUiHvzIxtnkM9ScW4m57X6hVUTbPK+X77/0X/ov/X9c+fu75+s6mYMjHnv/Xk7lVNIpVIadKBTZc1OU6mQytDH8tO9IIUcnhBCiICTBEY+9uJQszIoL3PpGt1J9ldjkx/d/SUII8SiTOTjisRfqrUdtLYHZpkChuPHE1mIKJsLH46HEEOYRzuUhyQ+lLSGEeBzICI547JXwMVAlKAJbWktsVh02G5izgnEzdeKp6mULOzwhhBAFUORGcKZMmcL8+fPx8cme2Pnaa68RHR0NwLRp01iwYAFKpZL33nuPRo3yfiVWiLwoFApGtqjKdzsM/H68FmZbFoFuHgxqUJEqxTwLOzwhhBAFUOQSHID+/fszcOBAh7Ljx48TExNDTEwMcXFxPPvss6xZswaVSlVIUQpnYtCqGNY4ihfql8VosuChV9/1NgtCCCGKjkfmEdX69etp3749Wq2W0NBQwsPDOXDgQGGHJZyMTq3E00UjyY0QQjziimSCM2/ePDp27Mg777xDUlISAHFxcQQF3dgrJzAwkLi4uMIKUQghhBBFWKE8ourfvz9Xr+bcx+fVV1+ld+/eDBkyBIVCwRdffMGECRP46KOPyG09wrz+l61SKfDyMty3uIsSlUrptH3LD+m/9F/6L/1/XD3u/c+vQklwZs6cma/rnnzySQYPHgxAUFAQsbGx9nNxcXEEBATc8X6Lxea0qz0+7itZSv+l/9L/nP33D7ixrIFNr8cSGoax51NkDHkZ1EVyymWB3K7/mm1b8OrSnoRNO7BElb/t/e4vD0Z15DCJv216kGE+MI/73/9HdiXjy5cv279et24dkZGRADRr1oyYmBiysrI4d+4cp0+fpnLlyoUVphBCFEnpL77MtZXrSPpxAVnNW+H24fsYvvy0sMMS4qErcin9xIkTOXIke3n8YsWKMXbsWAAiIyNp27Yt7dq1Q6VSMXr0aHmDSgghbmEJC8NcszYApoaNUf97GN38n0h/bUQhRybEw1XkRnAmTpzI8uXLWb58Od98843DY6gXX3yRdevWsWbNGvvaOEIIIW7PXKESqosXcpRrV8Xg1TIav1B/fCuUxnXMKDCZ7OdVx47i/nx/fKpG4RceiHej2rhM+wqs1huVmEy4vj8Sn2rl8Svuh0+lMng80weysm7U8/cBPLt1xC88EN/IMNwHD0Rx00i98uwZ/AM80C1dhNvrr+Bbqjg+Vcph+HicQ1s3x6P2dM89nut1xl7C46kn8YsIwqdaefQzZ+T5OSnPn8P9+f74lgnDLzwQzx6dUR0/lud9ougqciM4Qggh7h/lhXNYwsIdynRLF+H+wgCMTz9L2rujUZ0+heu4MWC1kjZmXPZ9ly5iKRVJZree2NzcUB88gOGTj8BoJOOV1wEwfDEZ3cL5pL33P6zhESgvx6FdtxYsFgAUV6/i1aU9lsgyJH89A0VaKq4f/g+vJztx7bdNoNXaY3IdO5rM9k+QPGM22i2bcJ38MZZyUWR26pojHtcgX4w7dueI5zr34S+R+WQvkga+gC5mOe4jhmMNCSGrVdtcPyPFtQS8OrbG6u1D6sTPsbm4YPjyMzy7P0HC9r3g4nJfvhfi4ZIERwghnIjCagWzGYUxA+3a1ehilpMy5ZsbF9hsuI4ZRWaP3qR+8hkAJgCdDre3Xyf9ldew+fhiatwEU+Mm9ntMdeqhyMjAZc4se0Kh3vcnmV2fJLPXU/bqryckAIavpwCQNH8xNvfsCdCWUqXxbtMM3YqlZHZ90n6tqW590saOz/66STO0v69DG7PMXt/N8Rg8XcioUD1HPNdlNW9J2sj3s+9r1gLVmdMYPpt42wTHZdpXKNLTSPp9Kzbv7FX0TbXr4lOjEvof52Ac+Hx+PnpRxEiCI4QQTsRt5Fu4jXzLfpz+wlAyu3S3H6tOHEd1/hypnbqA2Wwvz2rYGIXRiPrIYUz1G4LRiOGLyegXzkd54TyKmx5fYTaDWo25YiVcZs7A6h9AVrMWWMpXgJuW71Dv+5OsJs3syQ2AuXpNLGHhaHZud0hwspo0c+iHuWxZVOfP3yi4JR7/XOKx19Wug0Ndme074jZyRPbIUi5zN7WbNpIV3TQ7zv8+E5ubO+YqVdHs34cxxx3iUSAJjhBCOJH0oa+Q2akLiuRkXKZPxTDtK0zRTchq0RoARXw8AJ69u+d6v/JCdlLh+sFoXObNJu31tzFXroLN0xPt6hhcP50IRiO4uZE+fAQolLjM/A63D0ZjCQ4hY+gwMp4fkl1XXCyWsuVytGH190eReM2hzOZ5y75vGi1kZtoPb45HX782KSp9jnjs9fv5O7bn54/CbEYRH48tl+VFFAnx6P/cjX7Johznsho1yfVzEkWfJDhCCOFELMWLY65aHQBTvQZ4R9fFdcwospq3AoUCm7c3ACmTv8RcKedSG9fn6+iWLSFj4AtkvPyq/Zz2tzWOF+v1pL/9Hulvv4fq5HH0M7/H7b23MZeOxNSsJdbAIBS5LOqqvHIFc+Wqd9Wvm+PReRkwJ6bnjOd6/Vev5Di2qdXYfH1zvd7m7U1m2Xa5vmlmc8vfmiui6Clyb1EJIYS4TzQa0t5+D/W/R9CuWQWApXQkluAQlOfOYq5aPccfm092EqAwGrHdNAkYiwXdkoW3bcpSsjRpY8Zh0+lQ//svkP04SrthPYrUFPt16n1/ojp7BlOdenfVlbuJR7tyhcOxbuVyzFWq5vp4CsDUqAnqI4cxl43K8XlYSkfeVZyi6JARHCGEKGKsNhvbT19j9/5TWCxWKpcPIzrSH6367v9PmtWhE+bIMhi++oKsNu1AqSRtzDjchz6PMiWZrOYtsWm0qM6cRrtqBckz5oDBQFZ0U1x++BZLiZLYvL3Rf/8tiswsh7o9numDuUpVzJUqY9O7oFu+BMxmTPXqA5A++CX0M2fg2aML6S8Pt79FZY6qQGaHTnfXj5viURQPwmPK/+WI5zrt+t8wjB+LqV4DdDHL0W7aQNLsn29bd/rgl9At+AWvrh3IeO4FrMEhKK9cRvPHVkx16jnMFRKPDklwhBCiCLHZbHyz9hDXtu/iiX+3orWaWVOiFjtq1OGtLjXuPslRKkkf9hoeLw9GvXsn5lp1yOzcDZu7O4bPJ6P/aS42pQpreASZrVrbX91OHT8R9zdfxe3tN0Cvx9izD1ntOuL++jB71aZaddAtXYTLV1/E2dxAAAATW0lEQVSC1YqlbFmSv59rf0Rm8/MjafEKXN8ficfgAdg0WrJatCR17ASHV8Tz4+Z4FAYXsp7snSOe61I+nYLhv/lHVi9vUiZMzk7ubsPm60viqvW4jh+L26h3UCQnYQ0MwlS7LubyFe8qTlF0KGy57WLpJEwmi9Pu1/G470Ui/Zf+O2v//7mUzI8/rOSzjdPR2rLXk7EBH9TsQcX+PWhW1t+p+58f0v/Hu/+P7F5UQgjxOPvr9FWaHt1hT24AFEDLk7v569CZwgtMiEeMJDhCCFGEqFUqMtW6HOWZSjVqtey/J0R+SYIjhBBFSJ3S/qwrV49EzY3tATIVKpZFNaZOpYjCC0yIR4xMMhZCiCIkwsdA46bVeZUXafnvdrSWLH6PrEuJWpWoFe5V2OEJ8ciQBEcIIYqYLjXCqF7Cj10nq5FssfJMhC9lA9xQ3LQNghDiziTBEUKIIijcx0C4T1hhhyHEI0vm4AghhBDC6UiCI4QQQginIwmOEEIIIZyOJDhCCCGEcDqS4AghhBDC6UiCI4QQQginIwmOEEIIIZyOJDhCCCGEcDqS4AghhBDC6UiCI4QQQginIwmOEEIIIZyOJDhCCCGEcDqS4AghhBDC6UiCI4QQQginIwmOEEIIIZyOJDhCCCGEcDqS4AghCp3up7n4B3jgPmRQYYcihHASkuAIIQqdftGvAGhXxUBGRiFHI4RwBpLgCCEKleLKFTRbN5PVqAnKtFS0v60u7JCEEE5AEhwhRKHSLVuEwmIhdcIkLMEh6BctcDhvmPAh3nWr3ShIS0Nt0OPVvJG9SBEfj1+gJ5qNv9vL1Du249mpLX7hgfiWDcfttZdRpKbcuCcpEbfhL+FTqQx+of74VCuP22svO7St2bIJrzZN8Qv1///27j46qvrO4/hnMpMJEAiBQBLFEAWDQU2KDzypaAnErAyzRIJH3UUXNj6ArFvqUVuoRgUOom2PCvawodCl63oUDyh1GTkWUnkQQbTCJuBagZoSqiSQQDCJmce7f6SORjI2hMBkfnm/zsk5ye/em/l+73AvH373l6CUy4eq96M/lhoavtm+Y7sGpiYpftsWJd19hwZcnK5+o0co/p0yKRhU4pOPKSX7YvXPvUw9l794Wu9/r0YAHUfAARBVPV5fK3/ODxTMGibvlKlylv1etlP14e3+sdfL8edDstXUSJLiP3hfcjjk2F8h25enWsZ2vSfFxcl/7ShJkuP9XUqe5lYoNU2nVv2XGhYukXPz79Xn3x8If9/EkvmKf3+nGhc+rfo1b6hxfolks4W32//0ifreMVWh/ik69ZuX1PjofCW8vlZ9i+86rYc+D/9I/lFjVf+fLyt00WAlFd+t3j99WLaGBp1avkpe9xT1fmK+HB/uDh/TnhoBdJwj2gUA6L7iqg7L8eFuNT72lCTJe2uRev3Hi3K+tUHeO/5ZkuS/dpQsh0Px778nn7tQ8bvek/UPt8jauVOOD96XPy9f8bveUyAnV+rdW5LUe9ET8o8crS9/vTr8WqELLlRykVv2//tYweGXK37PH/XVv94rb2FReB/vbXeEP+/1yyUKXpShUy+tkex2SZLVr5+S7p0hxwfvKzBydHjf5tvu0Ff/9iNJUsOFg9R/3CjZDx1Q/esbWnq4abx6rH9dCW9tUOBvIaw9NQLoOGZwAERNwhvrJEnewqmSpMBV1yhwyZDwomNJUmKiAjm5LbM0apmtscaNk3/MdYrftTM85h99Xcv+TU1yfLhb3n+8VQoEwh/+0WNlxcfL8b97Wl7rihz1+tVS9fjNr2U/dOC02uI/+ki+Se5wuJEk7+Qpfwtbu1rt6xv3w/DnwUuGSJL8N9z4zQ5xcQpmXqy4Lz4/oxoBdBwBB0DU9HhjrQK5I2QlJclWf1K2+pPyFUxS/Patsh07Ft7PP+b6ljDj8yl+zx8VuuEG+ceMbQk9DQ1y7CuXf0xLwImrPylbMKg+P3lIAy/s/83HRQNk8/tl//yvkqSGJb+Q9xaXEn/5jPqPvUb9Ro9QwhvfrP+Jqzmq0MDU1gXb7bL69VfcyROthq2+fb/5wumUJIW+PSbJcjpl83rPqEYAHccjKgBRYT/wqRz7KyRJA7IGn7Y94c031Fx8nyTJP+Y69Sz9lZzbt8hyxEu5P5C/0a/eTz4m57vbZAsG5R89VpIUSuory2ZT0yPz5Jt482nfN5R+gSTJ6pusxsU/V+Pin8u+f596vfi8+sy+R4HLr1TwsmyFUtMVd/xY64ODQdlO1CmU3O+sem9vjQA6joADICoS1r0my27XqZdeldWzV6ttvX/2E/V4Y+03AWf0GMmy1HPpc/KPGq04u13By6+Q1aOnei5fpkDWMFkDBrQcnJiowDUjZT94QIGHf9quWoJXXKnGJxepx7rXZD/wqYKXZct/zTVyvvU/avzZE+HHVE7Pm7IFAi31nI0O1AjgzBBwAERFwvp18t80Xr6JBadta779n5T45M8UV3VYoYzBsvqnKHhZtpw7d6jhsSeVILX81NSo0UrY9La+umtGq+MbShYqeZpbeiBOXnehrN69FffXI0rY9LYa5z+u4NAsJU++Wd5JbgWyh0s2m3r+929l9UpU4OprJElNP35U/SbcoKR/uVPNM4oV9/nnSlz4hHzjJ7RaYNxR7akRQMexBgdAp9j/xSk9/+YePb56m0o3fay/1DVF3Nex9yM5/nxIzd/6qaVv806dJsXFhRchSwovIvaPuf70sVGtZ1QCY8bq5O82Kq72uPrMuU9977pdvV58XsELB4XX1fivHaUea15WUvHdSrp3hmx1tap/dZ1CFw6SJAWzh6v+lXWKO35MSTOnK3HJQnlvLVL9qpc6cHZO154aAXSczbIsK9pFnCt+f1AnT0a+ycay5ORexvbWHvTftfrf8qcabfjdDt25d6MuaapVRfIgrb16kubcfr2y0/p0+ut1tf7PN/qn/+7c/8CB7bun8IgKwFnxBUJa+4cKLX73txr8VctPFw1pPK4BzV/qtYHJKrnzLNerAEAH8IgKwFn5y4kmpZ04Gg43Xxtb+5mqjp5Qsz8YpcoAdGcEHABnpYfDri8dPfXdZ92NdqdsdrvscbY2jwOAc4mAA+CsXJTcQ84L0/WH1OzwmCXptSFjdHX2IMXbuc0AOP9YgwPgrNhsNs2afJV+0ezXu599rKE1ldo7+Ap5hw7TIzcOi3Z5ALopAg6As3ZRck89O2OcPjico9pGnyb166ERg/oqzsbjKQDRQcAB0CmcjjhdP6R/tMsAAEmswQEAAAYi4AAAAOMQcAAAgHEIOAAAwDgEHAAAYBwCDgAAMA4BBwAAGIeAAwAAjEPAAQAAxiHgAAAA4xBwAACAcQg4AADAOAQcAABgHAIOAAAwDgEHAAAYh4ADAACMQ8ABAADGIeAAAADjEHAAAIBxohJwNm7cKJfLpezsbFVUVLTaVlpaqvz8fBUUFGj79u3h8W3btqmgoED5+flasWLF+S4ZAADEkKgEnGHDhmnZsmUaOXJkq/GDBw/K4/HI4/Fo5cqVeuqppxQMBhUMBrVgwQKtXLlSHo9HGzZs0MGDB6NROgAAiAGOaLzo0KFD2xwvKyuTy+WS0+lURkaGMjMzVV5eLknKzMxURkaGJMnlcqmsrEyXXnrpeasZAADEji61Bqe6ulrp6enhr9PS0lRdXR1xHAAAoC3nbAZnxowZOn78+Gnjc+fO1cSJE9s8xrKs08ZsNptCoVCb43+P3W5TcnKvdlQbe+z2OGN7aw/6p3/6p//uqrv3317nLOCsXr36jI9JT0/X0aNHw19XV1crNTVVkiKOf59g0NLJk01nXEcsSE7uZWxv7UH/9E//9N9ddff+Bw7s0679utQjqry8PHk8Hvl8PlVVVamyslK5ubnKyclRZWWlqqqq5PP55PF4lJeXF+1yAQBAFxWVRcabNm3SwoULVVdXp/vvv1/Dhw/XqlWrlJWVpVtuuUWTJk2S3W5XSUmJ7Ha7JKmkpET33HOPgsGgioqKlJWVFY3SAQBADLBZbS18MYTfHzR2Gq+7T1HSP/3TP/13V929/5h8RAUAANAZCDgAAMA4BBwAAGAcAg4AADAOAQcAABiHgAMAAIxDwAEAAMYh4AAAAOMQcAAAgHEIOAAAwDgEHAAAYBwCDgAAMA4BBwAAGIeAAwAAjEPAAQAAxiHgAAAA4xBwAACAcQg4AADAOAQcAABgHAIOAAAwDgEHAAAYh4ADAACMQ8ABAADGIeAAAADjEHAAAIBxCDgAAMA4BBwAAGAcAg4AADAOAQcAABiHgAMAAIxDwAEAAMYh4AAAAOMQcAAAgHEIOAAAwDgEHAAAYBwCDgAAMA4BBwAAGIeAAwAAjEPAAQAAxiHgAAAA4xBwAACAcQg4AADAOAQcAABgHAIOAAAwDgEHAAAYh4ADAACMY7Msy4p2EQAAAJ2JGRwAAGAcAg4AADAOAQcAABiHgAMAAIxDwAEAAMYh4AAAAOMQcGLAxo0b5XK5lJ2drYqKivD4kSNHlJubqylTpmjKlCkqKSkJb9u3b5/cbrfy8/O1aNEixfJvA4jUvySVlpYqPz9fBQUF2r59e3h827ZtKigoUH5+vlasWHG+Sz5nli1bpnHjxoXf861bt4a3RToXpjH1vf0+eXl5crvdmjJliqZOnSpJOnnypGbOnKmbb75ZM2fOVH19fZSr7Dzz5s3T2LFjNXny5PBYpH4ty9KiRYuUn58vt9ut/fv3R6vsTtNW/1z7HWChyzt48KB16NAha/r06VZ5eXl4vKqqynK5XG0eU1RUZH300UdWKBSyiouLrS1btpyvcjtdpP4PHDhgud1uy+v1WocPH7YmTJhgBQIBKxAIWBMmTLAOHz5seb1ey+12WwcOHIhiB51n6dKl1sqVK08bj3QuTGPye/t9xo8fb9XW1rYae+aZZ6zS0lLLsiyrtLTUevbZZ6NR2jmxe/dua9++fa3ub5H63bJli1VcXGyFQiFrz5491rRp06JSc2dqq//ufu13BDM4MWDo0KEaMmRIu/evqalRQ0ODrrrqKtlsNhUWFqqsrOwcVnhuReq/rKxMLpdLTqdTGRkZyszMVHl5ucrLy5WZmamMjAw5nU65XK6Y7r89Ip0L03TH9zaSsrIyFRYWSpIKCwu1efPmKFfUeUaOHKm+ffu2GovU79fjNptNI0aM0KlTp1RTU3Pea+5MbfUfSXe59juCgBPjjhw5osLCQk2fPl0ffvihJKm6ulrp6enhfdLT01VdXR2tEs+Z7/aZlpam6urqiOOmePnll+V2uzVv3rzwNL3pPX+tu/TZluLiYk2dOlVr1qyRJNXW1io1NVWSlJqaqrq6umiWd85F6re73O+k7n3td4Qj2gWgxYwZM3T8+PHTxufOnauJEye2eUxqaqreeecd9evXT/v27dOcOXPk8XjaXG9js9k6vebO1JH+I/UZCoXaHI8V33cu7rzzTj3wwAOy2Wx64YUXtGTJEj399NMx+Z53RHfp87teeeUVpaWlqba2VjNnzjyjGV3TdZc/E9392u8IAk4XsXr16jM+xul0yul0SpKuvPJKDR48WJ999pnS09N19OjR8H5Hjx4N/8unq+pI/9/ts7q6OtxnpPFY0N5zcdttt2nWrFmSvv9cmKS79PldaWlpkqSUlBTl5+ervLxcKSkpqqmpUWpqqmpqatS/f/8oV3luReo3Fu93HTFgwIDw593x2u8IHlHFsLq6OgWDQUlSVVWVKisrlZGRodTUVCUmJmrv3r2yLEvr16/XhAkTolxt58vLy5PH45HP5wv3n5ubq5ycHFVWVqqqqko+n08ej0d5eXnRLrdTfHttwebNm5WVlSUp8rkwjcnvbSRNTU1qaGgIf75jxw5lZWUpLy9P69evlyRjr/Fvi9Tv1+OWZWnv3r3q06ePkX/Bd/drvyP438RjwKZNm7Rw4ULV1dUpKSlJw4cP16pVq/T2229r6dKlstvtstvtevDBB8M3+4qKCs2bN0/Nzc268cYb9fjjj8fstGWk/iVp+fLlWrdunex2u+bPn6+bbrpJkrR161YtXrxYwWBQRUVFmj17djRb6DSPPPKIPvnkE0nSoEGDtGDBgvDNPNK5MI2p720kVVVVmjNnjiQpGAxq8uTJmj17tk6cOKG5c+fqiy++0AUXXKAXXnhBycnJUa62czz00EPavXu3Tpw4oZSUFD344IOaOHFim/1alqUFCxZo+/bt6tmzpxYvXqycnJxot3BW2up/9+7d3f7aP1MEHAAAYBweUQEAAOMQcAAAgHEIOAAAwDgEHAAAYBwCDgAAMA4BBwAAGIeAAwAAjEPAARCzysvL5Xa75fV61dTUJJfLpU8//TTaZQHoAvhFfwBi2nPPPSefz6fm5malp6fr/vvvj3ZJALoAAg6AmObz+TRt2jQlJCTo1Vdfld1uj3ZJALoAHlEBiGn19fVqampSY2OjvF5vtMsB0EUwgwMgps2aNUsul0tHjhzRsWPHVFJSEu2SAHQBzOAAiFnr16+Xw+GQ2+3Wfffdp4qKCu3cuTPaZQHoApjBAQAAxmEGBwAAGIeAAwAAjEPAAQAAxiHgAAAA4xBwAACAcQg4AADAOAQcAABgHAIOAAAwzv8DOMGfT0A4R9kAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 648x648 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "word_vectors_plot(model, \"excellent\", [t[0] for t in model.wv.most_similar(positive=[\"excellent\"], \n",
    "                                                                             topn=16)][8:])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Eight Most Similar Words Vs. Ninth To Sixteenth Most Similar Words"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 648x648 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "word_vectors_plot(model, \"terrible\", [t[0] for t in model.wv.most_similar(positive=[\"terrible\"], \n",
    "                                                                             topn=16)][8:])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Eight Most Similar Words Vs. Ninth To Sixteenth Most Similar Words"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 648x648 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "word_vectors_plot(model, \"ipad\", [t[0] for t in model.wv.most_similar(positive=[\"ipad\"], \n",
    "                                                                             topn=16)][8:])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Wordcloud visualization of positive words in reviews"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "superb fantastic exceptional incredible excellent fabulous remarkable terrific impressive terms delivers aesthetic stellar aesthetics delight phenomenal soundcons superior wonderful amazing\n"
     ]
    }
   ],
   "source": [
    "pos_lst=[t[0] for t in model.wv.most_similar(positive=[\"outstanding\"],topn=20)]\n",
    "pos_wrd=' '.join(pos_lst)\n",
    "print(pos_wrd)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "wordcloud = WordCloud(background_color='white',\n",
    "                      max_words=200000,\n",
    "                      max_font_size=400).generate(pos_wrd)\n",
    "plt.figure()\n",
    "plt.imshow(wordcloud)\n",
    "plt.axis(\"off\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Wordcloud visualization of negative words in reviews"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "terrible horrible subpar worse suck oversensitive frustrating sucks weird quirky crappy crackling beeping stupid unreliable hears nows worst unresponsive unusable\n"
     ]
    }
   ],
   "source": [
    "neg_lst=[n[0] for n in model.wv.most_similar(positive=[\"awful\"],topn=20)]\n",
    "neg_wrd=' '.join(neg_lst)\n",
    "print(neg_wrd)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "wordcloud = WordCloud(background_color='black',\n",
    "                      max_words=100,\n",
    "                      max_font_size=50).generate(neg_wrd)\n",
    "plt.figure()\n",
    "plt.imshow(wordcloud)\n",
    "plt.axis(\"off\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### **Generating Feature vectors**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [],
   "source": [
    "\"\"\"\n",
    "Function to generate feature vectors\n",
    "\"\"\"\n",
    "\n",
    "def generate_feature_vectors(doc, model):\n",
    "    vec = np.zeros(150).reshape((1, 150))  \n",
    "    count = 0\n",
    "    for word in gensim.utils.simple_preprocess(doc):\n",
    "        if model.__contains__(word.strip()):\n",
    "            count = count + 1\n",
    "            vec += model[word.strip()]\n",
    "    vec = vec / count\n",
    "    return vec\n",
    "\n",
    "def generate_features(model, data):\n",
    "    features = np.concatenate([generate_feature_vectors(s, model) for s in data])\n",
    "    return features\n",
    "\n",
    "\"\"\"\n",
    "Generating train, test and validation vectors\n",
    "\"\"\"\n",
    "\n",
    "training_vectors = generate_features(model, x_train)\n",
    "test_vectors = generate_features(model, x_test)\n",
    "validation_vectors = generate_features(model, x_validation)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "####  Word2Vec Word Embedding Based Sentiment Analysis using LogisticRegression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "***** Word2Vec Word Embedding Based Sentiment Analysis using LogisticRegression *******\n",
      "\n",
      "LogisticRegression Performance : \n",
      "\n",
      "Train-Set Score : 0.9046\n",
      "Train-Set Accuracy : 0.9046\n",
      "\n",
      "Evaluation on Validation-Set : \n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.78      0.69      0.73       762\n",
      "           1       0.93      0.96      0.94      3312\n",
      "\n",
      "   micro avg       0.91      0.91      0.91      4074\n",
      "   macro avg       0.86      0.82      0.84      4074\n",
      "weighted avg       0.90      0.91      0.90      4074\n",
      "\n",
      "Confusion matrix:\n",
      " [[ 525  237]\n",
      " [ 146 3166]]\n",
      "Validation-Set Score : 0.9060\n",
      "Validation-Set Accuracy:0.9060\n",
      "\n",
      "Evaluation on Test-Set : \n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.82      0.67      0.73       833\n",
      "           1       0.92      0.96      0.94      3242\n",
      "\n",
      "   micro avg       0.90      0.90      0.90      4075\n",
      "   macro avg       0.87      0.81      0.84      4075\n",
      "weighted avg       0.90      0.90      0.90      4075\n",
      "\n",
      "Confusion matrix:\n",
      " [[ 554  279]\n",
      " [ 125 3117]]\n",
      "Test-Set Score : 0.9009\n",
      "Test-Set Accuracy:0.9009\n"
     ]
    }
   ],
   "source": [
    "lr = LogisticRegression()\n",
    "lr.fit(training_vectors, y_train)\n",
    "print(\"***** Word2Vec Word Embedding Based Sentiment Analysis using LogisticRegression *******\\n\")\n",
    "print(\"LogisticRegression Performance : \\n\")\n",
    "print('Train-Set Score : {:.4f}'.format(lr.score(training_vectors, y_train)))\n",
    "print('Train-Set Accuracy : {:.4f}'.format(accuracy_score(y_train,lr.predict(training_vectors))))\n",
    "\n",
    "print(\"\\nEvaluation on Validation-Set : \")\n",
    "pred_val = lr.predict(validation_vectors)\n",
    "print(\"Classification report:\\n {}\".format(classification_report(y_validation, pred_val)))\n",
    "print(\"Confusion matrix:\\n {}\".format(confusion_matrix(y_validation, pred_val)))\n",
    "print('Validation-Set Score : {:.4f}'.format(lr.score(validation_vectors, y_validation)))\n",
    "print('Validation-Set Accuracy:{:.4f}'.format(accuracy_score(y_validation, pred_val)))\n",
    "\n",
    "print(\"\\nEvaluation on Test-Set : \")\n",
    "pred = lr.predict(test_vectors)\n",
    "print(\"Classification report:\\n {}\".format(classification_report(y_test, pred)))\n",
    "print(\"Confusion matrix:\\n {}\".format(confusion_matrix(y_test, pred)))\n",
    "print('Test-Set Score : {:.4f}'.format(lr.score(test_vectors, y_test)))\n",
    "print('Test-Set Accuracy:{:.4f}'.format(accuracy_score(y_test, pred)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "####  Word2Vec Word Embedding Based Sentiment Analysis using SVC"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "***** Word2Vec Word Embedding Based Sentiment Analysis using SVC *******\n",
      "\n",
      "SVC with linear kernel Performance : \n",
      "\n",
      "Train-Set Score : 0.9050\n",
      "Train-Set Accuracy : 0.9050\n",
      "\n",
      "Evaluation on Validation-Set : \n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.79      0.70      0.74       762\n",
      "           1       0.93      0.96      0.94      3312\n",
      "\n",
      "   micro avg       0.91      0.91      0.91      4074\n",
      "   macro avg       0.86      0.83      0.84      4074\n",
      "weighted avg       0.91      0.91      0.91      4074\n",
      "\n",
      "Confusion matrix:\n",
      " [[ 533  229]\n",
      " [ 142 3170]]\n",
      "Validation-Set Score : 0.9089\n",
      "Validation-Set Accuracy:0.9089\n",
      "\n",
      "Evaluation on Test-Set : \n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.81      0.67      0.73       833\n",
      "           1       0.92      0.96      0.94      3242\n",
      "\n",
      "   micro avg       0.90      0.90      0.90      4075\n",
      "   macro avg       0.86      0.81      0.84      4075\n",
      "weighted avg       0.90      0.90      0.90      4075\n",
      "\n",
      "Confusion matrix:\n",
      " [[ 557  276]\n",
      " [ 131 3111]]\n",
      "Test-Set Score : 0.9001\n",
      "Test-Set Accuracy: 0.9001\n"
     ]
    }
   ],
   "source": [
    "svm = SVC(kernel='linear')\n",
    "svm.fit(training_vectors, y_train)\n",
    "print(\"***** Word2Vec Word Embedding Based Sentiment Analysis using SVC *******\\n\")\n",
    "print(\"SVC with linear kernel Performance : \\n\")\n",
    "print('Train-Set Score : {:.4f}'.format(svm.score(training_vectors, y_train)))\n",
    "print('Train-Set Accuracy : {:.4f}'.format(accuracy_score(y_train,svm.predict(training_vectors))))\n",
    "\n",
    "print(\"\\nEvaluation on Validation-Set : \")\n",
    "pred_val = svm.predict(validation_vectors)\n",
    "print(\"Classification report:\\n {}\".format(classification_report(y_validation, pred_val)))\n",
    "print(\"Confusion matrix:\\n {}\".format(confusion_matrix(y_validation, pred_val)))\n",
    "print('Validation-Set Score : {:.4f}'.format(svm.score(validation_vectors, y_validation)))\n",
    "print('Validation-Set Accuracy:{:.4f}'.format(accuracy_score(y_validation, pred_val)))\n",
    "\n",
    "print(\"\\nEvaluation on Test-Set : \")\n",
    "pred = svm.predict(test_vectors)\n",
    "print(\"Classification report:\\n {}\".format(classification_report(y_test, pred)))\n",
    "print(\"Confusion matrix:\\n {}\".format(confusion_matrix(y_test, pred)))\n",
    "print('Test-Set Score : {:.4f}'.format(svm.score(test_vectors, y_test)))\n",
    "print(\"Test-Set Accuracy: {:.4f}\".format(accuracy_score(y_test, pred)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Word2Vec Word Embedding Based Sentiment Analysis using XGBClassifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "***** Word2Vec Word Embedding Based Sentiment Analysis using XGBClassifier *******\n",
      "\n",
      "XGBClassifier Performance : \n",
      "\n",
      "Train-Set Score : 0.9058\n",
      "Train-Set Accuracy : 0.9058\n",
      "\n",
      "Evaluation on Validation-Set : \n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.79      0.68      0.73       762\n",
      "           1       0.93      0.96      0.94      3312\n",
      "\n",
      "   micro avg       0.91      0.91      0.91      4074\n",
      "   macro avg       0.86      0.82      0.84      4074\n",
      "weighted avg       0.90      0.91      0.90      4074\n",
      "\n",
      "Confusion matrix:\n",
      " [[ 515  247]\n",
      " [ 137 3175]]\n",
      "Validation-Set Score : 0.9057\n",
      "Validation-Set Accuracy:0.9057\n",
      "\n",
      "Evaluation on Test-Set : \n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.81      0.64      0.72       833\n",
      "           1       0.91      0.96      0.94      3242\n",
      "\n",
      "   micro avg       0.90      0.90      0.90      4075\n",
      "   macro avg       0.86      0.80      0.83      4075\n",
      "weighted avg       0.89      0.90      0.89      4075\n",
      "\n",
      "Confusion matrix:\n",
      " [[ 535  298]\n",
      " [ 123 3119]]\n",
      "Test-Set Score : 0.8967\n",
      "Test-Set Accuracy: 0.8967\n"
     ]
    }
   ],
   "source": [
    "xgb = XGBClassifier()\n",
    "xgb.fit(training_vectors, y_train)\n",
    "print(\"***** Word2Vec Word Embedding Based Sentiment Analysis using XGBClassifier *******\\n\")\n",
    "print(\"XGBClassifier Performance : \\n\")\n",
    "print('Train-Set Score : {:.4f}'.format(xgb.score(training_vectors, y_train)))\n",
    "print('Train-Set Accuracy : {:.4f}'.format(accuracy_score(y_train,xgb.predict(training_vectors))))\n",
    "\n",
    "print(\"\\nEvaluation on Validation-Set : \")\n",
    "pred_val = xgb.predict(validation_vectors)\n",
    "print(\"Classification report:\\n {}\".format(classification_report(y_validation, pred_val)))\n",
    "print(\"Confusion matrix:\\n {}\".format(confusion_matrix(y_validation, pred_val)))\n",
    "print('Validation-Set Score : {:.4f}'.format(xgb.score(validation_vectors, y_validation)))\n",
    "print('Validation-Set Accuracy:{:.4f}'.format(accuracy_score(y_validation, pred_val)))\n",
    "\n",
    "\n",
    "print(\"\\nEvaluation on Test-Set : \")\n",
    "pred = xgb.predict(test_vectors)\n",
    "print(\"Classification report:\\n {}\".format(classification_report(y_test, pred)))\n",
    "print(\"Confusion matrix:\\n {}\".format(confusion_matrix(y_test, pred)))\n",
    "print(\"Test-Set Score : {:.4f}\".format(xgb.score(test_vectors, y_test)))\n",
    "print(\"Test-Set Accuracy: {:.4f}\".format(accuracy_score(y_test, pred)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Sentiment Analysis using Keras Convolutional Neural Networks(CNN) \n",
    "[Keras reference](https://blog.keras.io/using-pre-trained-word-embeddings-in-a-keras-model.html)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "x_train[2] :  exactly described fast laptop one slimmest around extremely fast practical super great quality screen high resolution highly recommend\n",
      "\n",
      " X_token_train[2] :  [206, 114, 72, 269, 1, 58, 36, 58, 325, 253, 477, 98, 91, 320, 4010, 1674, 1579, 4140, 1363, 260, 831, 178, 3884, 165, 91, 123, 33, 220, 282, 180, 326, 16181, 477, 172, 2, 238, 165, 355, 11280, 513, 14, 41, 122, 13, 6, 1986, 122, 268, 10, 386, 310, 1, 753, 238, 477, 31, 1078, 94, 8, 1, 391, 122, 1078, 10, 6, 126, 593, 486, 130, 945, 6, 546, 30391, 86, 2315, 93, 5701, 922, 20420, 30392, 610, 5831, 58, 188, 198, 2, 87, 130, 1835, 366, 73, 2, 566, 162, 20, 9255, 513, 73, 129, 844, 195, 213, 2315, 585, 20420, 1409, 12356, 30393, 5107, 1472, 268, 4439, 31, 113, 996, 76, 1, 129, 346, 137]\n",
      "\n",
      " vocab_size :  73915\n"
     ]
    }
   ],
   "source": [
    "\"\"\"\n",
    "Create the tokenizer\n",
    "\"\"\"\n",
    "number_of_words=len(vocabulary)\n",
    "tokenizer = Tokenizer(num_words=number_of_words)\n",
    "\n",
    "\"\"\"\n",
    "Fit the tokenizer\n",
    "\"\"\"\n",
    "tokenizer.fit_on_texts(x_train)\n",
    "\n",
    "\"\"\"\n",
    "Sequence encode\n",
    "\"\"\"\n",
    "X_token_train = tokenizer.texts_to_sequences(x_train)\n",
    "X_token_test = tokenizer.texts_to_sequences(x_test)\n",
    "X_token_validation = tokenizer.texts_to_sequences(x_validation)\n",
    "\n",
    "\"\"\"\n",
    "Adding 1 because of reserved 0 index\n",
    "\"\"\"\n",
    "vocabulary_size = len(tokenizer.word_index) + 1  \n",
    "\n",
    "print(\"x_train[2] : \",x_train[2])\n",
    "print(\"\\n X_token_train[2] : \",X_token_train[2])\n",
    "print(\"\\n vocab_size : \",vocabulary_size)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "device : 142\n",
      "described : 1354\n",
      "laptop : 1\n",
      "resolution : 258\n",
      "highly : 346\n"
     ]
    }
   ],
   "source": [
    "\"\"\"\n",
    "Checking the index of each word by looking at the word_index dictionary of the Tokenizer object\n",
    "\"\"\"\n",
    "for word in ['device','described','laptop', 'resolution', 'highly']:\n",
    "    print('{} : {}'.format(word, tokenizer.word_index[word]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [],
   "source": [
    "\"\"\"\n",
    "Pad sequences\n",
    "\"\"\"\n",
    "max_length = 1000\n",
    "\n",
    "X_token_train = pad_sequences(X_token_train, padding='post', maxlen=max_length)\n",
    "X_token_test = pad_sequences(X_token_test, padding='post', maxlen=max_length)\n",
    "X_token_validation = pad_sequences(X_token_validation, padding='post', maxlen=max_length)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:From /opt/conda/lib/python3.6/site-packages/tensorflow/python/framework/op_def_library.py:263: colocate_with (from tensorflow.python.framework.ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Colocations handled automatically by placer.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "W0506 18:21:23.356515 139801131669312 deprecation.py:323] From /opt/conda/lib/python3.6/site-packages/tensorflow/python/framework/op_def_library.py:263: colocate_with (from tensorflow.python.framework.ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Colocations handled automatically by placer.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "embedding_1 (Embedding)      (None, 1000, 100)         7391500   \n",
      "_________________________________________________________________\n",
      "conv1d_1 (Conv1D)            (None, 996, 128)          64128     \n",
      "_________________________________________________________________\n",
      "global_max_pooling1d_1 (Glob (None, 128)               0         \n",
      "_________________________________________________________________\n",
      "dense_1 (Dense)              (None, 10)                1290      \n",
      "_________________________________________________________________\n",
      "dense_2 (Dense)              (None, 1)                 11        \n",
      "=================================================================\n",
      "Total params: 7,456,929\n",
      "Trainable params: 7,456,929\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n",
      "CPU times: user 140 ms, sys: 0 ns, total: 140 ms\n",
      "Wall time: 142 ms\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "\"\"\"\n",
    "Create model\n",
    "\"\"\"\n",
    "embedding_dimension = 100\n",
    "\n",
    "keras_cnn_model = Sequential()\n",
    "keras_cnn_model.add(Embedding(input_dim=vocabulary_size, \n",
    "                              output_dim=embedding_dimension, \n",
    "                              input_length=max_length))\n",
    "keras_cnn_model.add(Conv1D(128, 5, activation='relu'))\n",
    "keras_cnn_model.add(GlobalMaxPooling1D())\n",
    "keras_cnn_model.add(Dense(10, activation='relu'))\n",
    "keras_cnn_model.add(Dense(1, activation='sigmoid'))\n",
    "\n",
    "\"\"\"\n",
    "Compile network\n",
    "\"\"\"\n",
    "keras_cnn_model.compile(optimizer='adam',\n",
    "                        loss='binary_crossentropy',\n",
    "                        metrics=['accuracy'])\n",
    "keras_cnn_model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:From /opt/conda/lib/python3.6/site-packages/tensorflow/python/ops/math_ops.py:3066: to_int32 (from tensorflow.python.ops.math_ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Use tf.cast instead.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "W0506 18:21:23.534948 139801131669312 deprecation.py:323] From /opt/conda/lib/python3.6/site-packages/tensorflow/python/ops/math_ops.py:3066: to_int32 (from tensorflow.python.ops.math_ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Use tf.cast instead.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:From /opt/conda/lib/python3.6/site-packages/tensorflow/python/ops/math_grad.py:102: div (from tensorflow.python.ops.math_ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Deprecated in favor of operator or tf.math.divide.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "W0506 18:21:23.680222 139801131669312 deprecation.py:323] From /opt/conda/lib/python3.6/site-packages/tensorflow/python/ops/math_grad.py:102: div (from tensorflow.python.ops.math_ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Deprecated in favor of operator or tf.math.divide.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " **** Sentiment Analysis Using Keras Convolutional Neural Networks(CNN) ****\n",
      "\n",
      "Train-Set Accuracy: 0.9992\n",
      "\n",
      "Evaluation on Validation-Set : \n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.84      0.74      0.78       762\n",
      "           1       0.94      0.97      0.95      3312\n",
      "\n",
      "   micro avg       0.92      0.92      0.92      4074\n",
      "   macro avg       0.89      0.85      0.87      4074\n",
      "weighted avg       0.92      0.92      0.92      4074\n",
      "\n",
      "Confusion matrix:\n",
      " [[ 561  201]\n",
      " [ 109 3203]]\n",
      "Validation-Set Accuracy: 0.9239\n",
      "\n",
      "Evaluation on Test-Set : \n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.85      0.71      0.77       833\n",
      "           1       0.93      0.97      0.95      3242\n",
      "\n",
      "   micro avg       0.91      0.91      0.91      4075\n",
      "   macro avg       0.89      0.84      0.86      4075\n",
      "weighted avg       0.91      0.91      0.91      4075\n",
      "\n",
      "Confusion matrix:\n",
      " [[ 591  242]\n",
      " [ 106 3136]]\n",
      "Test-Set Accuracy:  0.9146\n",
      "CPU times: user 3h 43min 14s, sys: 12min 42s, total: 3h 55min 56s\n",
      "Wall time: 33min 55s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "\"\"\"\n",
    "Fit network\n",
    "\"\"\"\n",
    "keras_cnn_model.fit(X_token_train, y_train,\n",
    "                    epochs=5,\n",
    "                    verbose=False,\n",
    "                    validation_data=(X_token_validation, y_validation),\n",
    "                    batch_size=10)\n",
    "\n",
    "\"\"\"\n",
    "Evaluate\n",
    "\"\"\"\n",
    "print(\"\\n **** Sentiment Analysis Using Keras Convolutional Neural Networks(CNN) ****\\n\")\n",
    "loss, accuracy = keras_cnn_model.evaluate(X_token_train, y_train, verbose=False)\n",
    "print(\"Train-Set Accuracy: {:.4f}\".format(accuracy))\n",
    "\n",
    "\n",
    "print(\"\\nEvaluation on Validation-Set : \")\n",
    "pred_val=keras_cnn_model.predict_classes(X_token_validation)\n",
    "print(\"Classification report:\\n {}\".format(classification_report(y_validation, pred_val)))\n",
    "print(\"Confusion matrix:\\n {}\".format(confusion_matrix(y_validation, pred_val)))\n",
    "loss, accuracy = keras_cnn_model.evaluate(X_token_validation, y_validation, verbose=False)\n",
    "print(\"Validation-Set Accuracy: {:.4f}\".format(accuracy))\n",
    "\n",
    "print(\"\\nEvaluation on Test-Set : \")\n",
    "pred=keras_cnn_model.predict_classes(X_token_test)\n",
    "print(\"Classification report:\\n {}\".format(classification_report(y_test, pred)))\n",
    "print(\"Confusion matrix:\\n {}\".format(confusion_matrix(y_test, pred)))\n",
    "loss, accuracy = keras_cnn_model.evaluate(X_token_test, y_test, verbose=False)\n",
    "print(\"Test-Set Accuracy:  {:.4f}\".format(accuracy))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Sentiment Analysis Using Pre-trained Word2Vec Word Embedding To Keras CNN"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Vocabulary size : 33606\n"
     ]
    }
   ],
   "source": [
    "\"\"\"\n",
    "Vocabulary size\n",
    "\"\"\"\n",
    "num_of_words = list(model.wv.vocab)\n",
    "print('Vocabulary size : %d' % len(num_of_words))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "W0506 18:55:19.554864 139801131669312 smart_open_lib.py:385] this function is deprecated, use smart_open.open instead\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 6.86 s, sys: 124 ms, total: 6.98 s\n",
      "Wall time: 6.97 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "\"\"\"\n",
    "Save model in ASCII\n",
    "\"\"\"\n",
    "file_name = 'amazon_embedding_word2vec.txt'\n",
    "model.wv.save_word2vec_format(file_name, binary=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [],
   "source": [
    "\"\"\"\n",
    "Load word embedding\n",
    "\"\"\"\n",
    "def load_word_embedding(file_name):\n",
    "    word_embedding = dict()\n",
    "    file = open(file_name,'r')\n",
    "    lines = file.readlines()[1:]\n",
    "    file.close()\n",
    "    \n",
    "    \"\"\"\n",
    "    Mapping words to vectors\n",
    "    \"\"\"\n",
    "    for line in lines:\n",
    "        line_parts = line.split()\n",
    "        word_embedding[line_parts[0]] = np.asarray(line_parts[1:], dtype='float32')\n",
    "        \n",
    "    return word_embedding"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [],
   "source": [
    "\"\"\"\n",
    "Create a weight matrix for the Embedding layer\n",
    "\"\"\"\n",
    "def get_embedding_weight_matrix(wrd_embedding, vocabulary):\n",
    "    \n",
    "    vocabulary_size = len(vocabulary) + 1\n",
    "    \"\"\"\n",
    "    Define weight matrix dimensions with all 0\n",
    "    \"\"\"\n",
    "    embedding_weight_matrix = np.zeros((vocabulary_size, 150))\n",
    "    \"\"\"\n",
    "    Step vocab, store vectors using the Tokenizer's integer mapping\n",
    "    \"\"\"\n",
    "    for wrd, i in vocabulary.items():\n",
    "        vector = wrd_embedding.get(wrd)\n",
    "        if vector is not None:\n",
    "            embedding_weight_matrix[i] = vector\n",
    "    return embedding_weight_matrix"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Completed creation of raw word2vec word embedding\n",
      "Completed creation of embedding weight vectors\n"
     ]
    }
   ],
   "source": [
    "\"\"\"\n",
    "Load embedding from file\n",
    "\"\"\"\n",
    "raw_w2v_embedding = load_word_embedding('amazon_embedding_word2vec.txt')\n",
    "\n",
    "print('Completed creation of raw word2vec word embedding')\n",
    "\n",
    "\"\"\"\n",
    "Get weight vectors in the right order\n",
    "\"\"\"\n",
    "embedding_weight_vectors = get_embedding_weight_matrix(raw_w2v_embedding, tokenizer.word_index)\n",
    "\n",
    "print('Completed creation of embedding weight vectors')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Completed creation of embedding layer\n"
     ]
    }
   ],
   "source": [
    "\"\"\"\n",
    "Create the embedding layer\n",
    "\"\"\"\n",
    "embedding_layer = Embedding(vocabulary_size, \n",
    "                            150, \n",
    "                            weights=[embedding_weight_vectors], \n",
    "                            input_length=max_length, \n",
    "                            trainable=False)\n",
    "\n",
    "print('Completed creation of embedding layer')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "embedding_2 (Embedding)      (None, 1000, 150)         11087250  \n",
      "_________________________________________________________________\n",
      "conv1d_2 (Conv1D)            (None, 996, 128)          96128     \n",
      "_________________________________________________________________\n",
      "max_pooling1d_1 (MaxPooling1 (None, 498, 128)          0         \n",
      "_________________________________________________________________\n",
      "flatten_1 (Flatten)          (None, 63744)             0         \n",
      "_________________________________________________________________\n",
      "dense_3 (Dense)              (None, 1)                 63745     \n",
      "=================================================================\n",
      "Total params: 11,247,123\n",
      "Trainable params: 159,873\n",
      "Non-trainable params: 11,087,250\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "\"\"\"\n",
    "Create model\n",
    "\"\"\"\n",
    "keras_cnn_w2v_model = Sequential()\n",
    "keras_cnn_w2v_model.add(embedding_layer)\n",
    "keras_cnn_w2v_model.add(Conv1D(filters=128, kernel_size=5, activation='relu'))\n",
    "keras_cnn_w2v_model.add(MaxPooling1D(pool_size=2))\n",
    "keras_cnn_w2v_model.add(Flatten())\n",
    "keras_cnn_w2v_model.add(Dense(1, activation='sigmoid'))\n",
    "keras_cnn_w2v_model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " **** Sentiment Analysis Using Pre-trained Word2Vec Word Embedding To Keras CNN ****\n",
      "\n",
      "Train-Set Accuracy: 0.9690\n",
      "\n",
      "Evaluation on Validation-Set : \n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.73      0.71      0.72       762\n",
      "           1       0.93      0.94      0.94      3312\n",
      "\n",
      "   micro avg       0.90      0.90      0.90      4074\n",
      "   macro avg       0.83      0.83      0.83      4074\n",
      "weighted avg       0.90      0.90      0.90      4074\n",
      "\n",
      "Confusion matrix:\n",
      " [[ 543  219]\n",
      " [ 201 3111]]\n",
      "Validation-Set Accuracy: 0.8969\n",
      "\n",
      "Evaluation on Test-Set : \n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.75      0.69      0.72       833\n",
      "           1       0.92      0.94      0.93      3242\n",
      "\n",
      "   micro avg       0.89      0.89      0.89      4075\n",
      "   macro avg       0.84      0.81      0.82      4075\n",
      "weighted avg       0.89      0.89      0.89      4075\n",
      "\n",
      "Confusion matrix:\n",
      " [[ 571  262]\n",
      " [ 190 3052]]\n",
      "Test-Set Accuracy:  0.8891\n"
     ]
    }
   ],
   "source": [
    "\"\"\"\n",
    "Compile network\n",
    "\"\"\"\n",
    "keras_cnn_w2v_model.compile(loss='binary_crossentropy', \n",
    "                            optimizer='adam', \n",
    "                            metrics=['accuracy'])\n",
    "\n",
    "\"\"\"\n",
    "Fit network\n",
    "\"\"\"\n",
    "keras_cnn_w2v_model.fit(X_token_train, y_train,\n",
    "                        epochs=5,\n",
    "                        verbose=False,\n",
    "                        validation_data=(X_token_validation, y_validation),\n",
    "                        batch_size=10)\n",
    "\n",
    "\"\"\"\n",
    "Evaluate\n",
    "\"\"\"\n",
    "print(\"\\n **** Sentiment Analysis Using Pre-trained Word2Vec Word Embedding To Keras CNN ****\\n\")\n",
    "loss, accuracy = keras_cnn_w2v_model.evaluate(X_token_train, y_train, verbose=False)\n",
    "print(\"Train-Set Accuracy: {:.4f}\".format(accuracy))\n",
    "\n",
    "\n",
    "print(\"\\nEvaluation on Validation-Set : \")\n",
    "pred_val=keras_cnn_w2v_model.predict_classes(X_token_validation)\n",
    "print(\"Classification report:\\n {}\".format(classification_report(y_validation, pred_val)))\n",
    "print(\"Confusion matrix:\\n {}\".format(confusion_matrix(y_validation, pred_val)))\n",
    "loss, accuracy = keras_cnn_w2v_model.evaluate(X_token_validation, y_validation, verbose=False)\n",
    "print(\"Validation-Set Accuracy: {:.4f}\".format(accuracy))\n",
    "\n",
    "print(\"\\nEvaluation on Test-Set : \")\n",
    "pred=keras_cnn_w2v_model.predict_classes(X_token_test)\n",
    "print(\"Classification report:\\n {}\".format(classification_report(y_test, pred)))\n",
    "print(\"Confusion matrix:\\n {}\".format(confusion_matrix(y_test, pred)))\n",
    "loss, accuracy = keras_cnn_w2v_model.evaluate(X_token_test, y_test, verbose=False)\n",
    "print(\"Test-Set Accuracy:  {:.4f}\".format(accuracy))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Sentiment Analysis  Using Pre-trained Word2Vec Word Embedding To Keras CNN And Bidirectional LSTM"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:From /opt/conda/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py:3445: calling dropout (from tensorflow.python.ops.nn_ops) with keep_prob is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Please use `rate` instead of `keep_prob`. Rate should be set to `rate = 1 - keep_prob`.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "W0506 19:09:25.157642 139801131669312 deprecation.py:506] From /opt/conda/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py:3445: calling dropout (from tensorflow.python.ops.nn_ops) with keep_prob is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Please use `rate` instead of `keep_prob`. Rate should be set to `rate = 1 - keep_prob`.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "embedding_3 (Embedding)      (None, 1000, 150)         11087250  \n",
      "_________________________________________________________________\n",
      "conv1d_3 (Conv1D)            (None, 996, 128)          96128     \n",
      "_________________________________________________________________\n",
      "max_pooling1d_2 (MaxPooling1 (None, 498, 128)          0         \n",
      "_________________________________________________________________\n",
      "bidirectional_1 (Bidirection (None, 128)               98816     \n",
      "_________________________________________________________________\n",
      "dropout_1 (Dropout)          (None, 128)               0         \n",
      "_________________________________________________________________\n",
      "dense_4 (Dense)              (None, 1)                 129       \n",
      "=================================================================\n",
      "Total params: 11,282,323\n",
      "Trainable params: 195,073\n",
      "Non-trainable params: 11,087,250\n",
      "_________________________________________________________________\n",
      "CPU times: user 1.14 s, sys: 820 ms, total: 1.96 s\n",
      "Wall time: 779 ms\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "\"\"\"\n",
    "Create model\n",
    "\"\"\"\n",
    "keras_cnn_bidir_lstm_w2v_model = Sequential()\n",
    "keras_cnn_bidir_lstm_w2v_model.add(Embedding(vocabulary_size, \n",
    "                                             150, \n",
    "                                             weights=[embedding_weight_vectors],\n",
    "                                             input_length=max_length,  \n",
    "                                             trainable=False))\n",
    "keras_cnn_bidir_lstm_w2v_model.add(Conv1D(128, 5, activation='relu'))\n",
    "keras_cnn_bidir_lstm_w2v_model.add(MaxPooling1D(pool_size=2))\n",
    "keras_cnn_bidir_lstm_w2v_model.add(Bidirectional(LSTM(64)))\n",
    "keras_cnn_bidir_lstm_w2v_model.add(Dropout(0.5))\n",
    "keras_cnn_bidir_lstm_w2v_model.add(Dense(1, activation='sigmoid'))\n",
    "\n",
    "\"\"\"\n",
    "Compile network\n",
    "\"\"\"\n",
    "keras_cnn_bidir_lstm_w2v_model.compile(loss='binary_crossentropy', \n",
    "                                       optimizer='adam', \n",
    "                                       metrics=['accuracy'])\n",
    "keras_cnn_bidir_lstm_w2v_model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 8h 13min 38s, sys: 40min 46s, total: 8h 54min 24s\n",
      "Wall time: 1h 17min 17s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "\"\"\"\n",
    "Fit train data\n",
    "\"\"\"\n",
    "keras_cnn_bidir_lstm_w2v_model.fit(X_token_train, y_train,\n",
    "                                   epochs=5,\n",
    "                                   verbose=False,\n",
    "                                   validation_data=(X_token_validation, y_validation),\n",
    "                                   batch_size=10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "**** Sentiment Analysis  Using Pre-trained Word2Vec Word Embedding To Keras CNN And Bidirectional LSTM ****\n",
      "\n",
      "Train-Set Accuracy: 0.9567\n",
      "\n",
      "Evaluation on Validation-Set : \n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.81      0.75      0.78       762\n",
      "           1       0.94      0.96      0.95      3312\n",
      "\n",
      "   micro avg       0.92      0.92      0.92      4074\n",
      "   macro avg       0.88      0.86      0.87      4074\n",
      "weighted avg       0.92      0.92      0.92      4074\n",
      "\n",
      "Confusion matrix:\n",
      " [[ 572  190]\n",
      " [ 131 3181]]\n",
      "Validation-Set Accuracy: 0.9212\n",
      "\n",
      "Evaluation on Test-Set : \n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.82      0.74      0.78       833\n",
      "           1       0.93      0.96      0.95      3242\n",
      "\n",
      "   micro avg       0.91      0.91      0.91      4075\n",
      "   macro avg       0.88      0.85      0.86      4075\n",
      "weighted avg       0.91      0.91      0.91      4075\n",
      "\n",
      "Confusion matrix:\n",
      " [[ 615  218]\n",
      " [ 131 3111]]\n",
      "Test-Set Accuracy:  0.9144\n"
     ]
    }
   ],
   "source": [
    "\"\"\"\n",
    "Evaluate\n",
    "\"\"\"\n",
    "print(\"**** Sentiment Analysis  Using Pre-trained Word2Vec Word Embedding To Keras CNN And Bidirectional LSTM ****\\n\")\n",
    "loss, accuracy = keras_cnn_bidir_lstm_w2v_model.evaluate(X_token_train, y_train, verbose=False)\n",
    "print(\"Train-Set Accuracy: {:.4f}\".format(accuracy))\n",
    "\n",
    "print(\"\\nEvaluation on Validation-Set : \")\n",
    "pred_val=keras_cnn_bidir_lstm_w2v_model.predict_classes(X_token_validation)\n",
    "print(\"Classification report:\\n {}\".format(classification_report(y_validation, pred_val)))\n",
    "print(\"Confusion matrix:\\n {}\".format(confusion_matrix(y_validation, pred_val)))\n",
    "loss, accuracy = keras_cnn_bidir_lstm_w2v_model.evaluate(X_token_validation, y_validation, verbose=False)\n",
    "print(\"Validation-Set Accuracy: {:.4f}\".format(accuracy))\n",
    "\n",
    "print(\"\\nEvaluation on Test-Set : \")\n",
    "pred=keras_cnn_bidir_lstm_w2v_model.predict_classes(X_token_test)\n",
    "print(\"Classification report:\\n {}\".format(classification_report(y_test, pred)))\n",
    "print(\"Confusion matrix:\\n {}\".format(confusion_matrix(y_test, pred)))\n",
    "loss, accuracy = keras_cnn_bidir_lstm_w2v_model.evaluate(X_token_test, y_test, verbose=False)\n",
    "print(\"Test-Set Accuracy:  {:.4f}\".format(accuracy))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Sentiment Analysis Using Pre-trained Word2Vec Word Embedding To Keras Bidirectional LSTM"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "embedding_4 (Embedding)      (None, 1000, 150)         11087250  \n",
      "_________________________________________________________________\n",
      "bidirectional_2 (Bidirection (None, 128)               110080    \n",
      "_________________________________________________________________\n",
      "dropout_2 (Dropout)          (None, 128)               0         \n",
      "_________________________________________________________________\n",
      "dense_5 (Dense)              (None, 1)                 129       \n",
      "=================================================================\n",
      "Total params: 11,197,459\n",
      "Trainable params: 110,209\n",
      "Non-trainable params: 11,087,250\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "\"\"\"\n",
    "Create model\n",
    "\"\"\"\n",
    "\n",
    "keras_bidir_lstm_w2v_model = Sequential()\n",
    "keras_bidir_lstm_w2v_model.add(Embedding(vocabulary_size, \n",
    "                                         150, \n",
    "                                         weights=[embedding_weight_vectors],\n",
    "                                         input_length=max_length,  \n",
    "                                         trainable=False))\n",
    "keras_bidir_lstm_w2v_model.add(Bidirectional(LSTM(64)))\n",
    "keras_bidir_lstm_w2v_model.add(Dropout(0.5))\n",
    "keras_bidir_lstm_w2v_model.add(Dense(1, activation='sigmoid'))\n",
    "\n",
    "\n",
    "\"\"\"\n",
    "Compile network\n",
    "\"\"\"\n",
    "keras_bidir_lstm_w2v_model.compile(loss='binary_crossentropy', \n",
    "                                   optimizer='adam', \n",
    "                                   metrics=['accuracy'])\n",
    "keras_bidir_lstm_w2v_model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 12h 40min 42s, sys: 1h 21min 17s, total: 14h 2min\n",
      "Wall time: 2h 1min 12s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "\"\"\"\n",
    "Fit train data\n",
    "\"\"\"\n",
    "keras_bidir_lstm_w2v_model.fit(X_token_train, y_train,\n",
    "                               epochs=5,\n",
    "                               verbose=False,\n",
    "                               validation_data=(X_token_validation, y_validation),\n",
    "                               batch_size=10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "**** Sentiment Analysis Using Pre-trained Word2Vec Word Embedding To Keras Bidirectional LSTM ****\n",
      "\n",
      "Train-Set Accuracy: 0.9418\n",
      "\n",
      "Evaluation on Validation-Set : \n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.78      0.80      0.79       762\n",
      "           1       0.95      0.95      0.95      3312\n",
      "\n",
      "   micro avg       0.92      0.92      0.92      4074\n",
      "   macro avg       0.87      0.87      0.87      4074\n",
      "weighted avg       0.92      0.92      0.92      4074\n",
      "\n",
      "Confusion matrix:\n",
      " [[ 607  155]\n",
      " [ 169 3143]]\n",
      "Validation-Set Accuracy: 0.9205\n",
      "\n",
      "Evaluation on Test-Set : \n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.83      0.78      0.81       833\n",
      "           1       0.94      0.96      0.95      3242\n",
      "\n",
      "   micro avg       0.92      0.92      0.92      4075\n",
      "   macro avg       0.89      0.87      0.88      4075\n",
      "weighted avg       0.92      0.92      0.92      4075\n",
      "\n",
      "Confusion matrix:\n",
      " [[ 652  181]\n",
      " [ 133 3109]]\n",
      "Test-Set Accuracy:  0.9229\n"
     ]
    }
   ],
   "source": [
    "\"\"\"\n",
    "Evaluate\n",
    "\"\"\"\n",
    "print(\"**** Sentiment Analysis Using Pre-trained Word2Vec Word Embedding To Keras Bidirectional LSTM ****\\n\")\n",
    "loss, accuracy = keras_bidir_lstm_w2v_model.evaluate(X_token_train, y_train, verbose=False)\n",
    "print(\"Train-Set Accuracy: {:.4f}\".format(accuracy))\n",
    "\n",
    "print(\"\\nEvaluation on Validation-Set : \")\n",
    "pred_val=keras_bidir_lstm_w2v_model.predict_classes(X_token_validation)\n",
    "print(\"Classification report:\\n {}\".format(classification_report(y_validation, pred_val)))\n",
    "print(\"Confusion matrix:\\n {}\".format(confusion_matrix(y_validation, pred_val)))\n",
    "loss, accuracy = keras_bidir_lstm_w2v_model.evaluate(X_token_validation, y_validation, verbose=False)\n",
    "print(\"Validation-Set Accuracy: {:.4f}\".format(accuracy))\n",
    "\n",
    "print(\"\\nEvaluation on Test-Set : \")\n",
    "pred=keras_bidir_lstm_w2v_model.predict_classes(X_token_test)\n",
    "print(\"Classification report:\\n {}\".format(classification_report(y_test, pred)))\n",
    "print(\"Confusion matrix:\\n {}\".format(confusion_matrix(y_test, pred)))\n",
    "\n",
    "loss, accuracy = keras_bidir_lstm_w2v_model.evaluate(X_token_test, y_test, verbose=False)\n",
    "print(\"Test-Set Accuracy:  {:.4f}\".format(accuracy))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## About Doc2Vec\n",
    "\n",
    "* **Doc2Vec** is a generalization of the Word2Vec algorithm and applies at the document level. According to [Mikolov et al. (2014)](https://arxiv.org/pdf/1405.4053.pdf), paragraphs in a document are mapped to a vector representation called paragraph vector. This is then combined with the word vectors by averaging or concatenating to predict the next word in a context. The paragraph vector is just like another word vector but it represents the missing context of the the topic of the paragraph.\n",
    "* **PV-DM or DM:** is the Doc2Vec model analogous to CBOW in Word2Vec. The document vectors are obtained by training a neural network on the task of inferring a centre word based on context words and a context paragraph. See image below for architecture ([image credit: Google](https://www.google.com/search?newwindow=1&biw=1855&bih=952&tbm=isch&sa=1&ei=vQvOXO-dH8rj_AasiKywBw&q=doc2vec+dm&oq=doc2vec+dm&gs_l=img.3..0i24l3.2275.7633..7995...0.0..0.145.386.4j1......1....1..gws-wiz-img.......0j0i30.DJjU4rFKil8))\n",
    "\n",
    "<span style=\"display:block;text-align:center\">![DM](dm.png)\n",
    "\n",
    "* **PV-DBOW or DBOW:** is the Doc2Vec model analogous to Skip-gram in Word2Vec. The document vectors are obtained by training a neural network on the task of predicting a probability distribution of words in a paragraph given a randomly-sampled word from the document. See image below for architecture ([image credit: Google](https://www.google.com/search?newwindow=1&biw=1855&bih=952&tbm=isch&sa=1&ei=5QvOXPTaNsmIggfb6JGoAw&q=doc2vec+dbow&oq=doc2vec+dbow&gs_l=img.3..0i24l3.20722.21318..21542...0.0..0.69.203.3......0....1..gws-wiz-img.5MtfgvxyAEI))\n",
    "\n",
    "<span style=\"display:block;text-align:center\">![DBOW](dbow.png)\n",
    "\n",
    "Below we will see the usage gensim python library with comaprisons of each of these models and their combinations. That is:\n",
    "1. **DBOW (Distributed Bag of Words)**\n",
    "2. **DMC (Distributed Memory Concatenated)**\n",
    "3. **DMM (Distributed Memory Mean)**\n",
    "4. **DBOW + DMC**\n",
    "5. **DBOW + DMM**\n",
    "    \n",
    "[Doc2Vec reference](https://radimrehurek.com/gensim/models/doc2vec.html)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {},
   "outputs": [],
   "source": [
    "\"\"\"\n",
    "Function to labelize the reviews\n",
    "\"\"\"\n",
    "def labelize_review(reviews,label):\n",
    "    labelized_review = []\n",
    "    prefix = label\n",
    "    for indx, rvw in zip(reviews.index, reviews):\n",
    "        labelized_review.append(LabeledSentence(rvw.split(), [prefix + '_%s' % indx]))\n",
    "    return labelized_review"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {},
   "outputs": [],
   "source": [
    "\"\"\"\n",
    "labelize the reviews\n",
    "\"\"\"\n",
    "all_reviews_d2v = labelize_review(all_reviews, 'all')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### **Distributed Bag Of Words (DBOW)**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "W0506 22:35:11.768792 139801131669312 base_any2vec.py:723] consider setting layer size to a multiple of 4 for greater performance\n",
      "100%|██████████| 40744/40744 [00:00<00:00, 2390542.65it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 2.81 s, sys: 20 ms, total: 2.83 s\n",
      "Wall time: 2.83 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "\"\"\"\n",
    "Create doc2vec DBOW model\n",
    "\"\"\"\n",
    "dbow_model = Doc2Vec(dm=0, \n",
    "                     size=150, \n",
    "                     negative=5, \n",
    "                     min_count=2, \n",
    "                     workers=10, \n",
    "                     alpha=0.065, \n",
    "                     min_alpha=0.065)\n",
    "\n",
    "dbow_model.build_vocab([review for review in tqdm(all_reviews_d2v)])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 40744/40744 [00:00<00:00, 2476634.33it/s]\n",
      "100%|██████████| 40744/40744 [00:00<00:00, 2600908.94it/s]\n",
      "100%|██████████| 40744/40744 [00:00<00:00, 2527849.27it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 26.7 s, sys: 1.09 s, total: 27.8 s\n",
      "Wall time: 10.8 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "\n",
    "\"\"\"\n",
    "Train the model\n",
    "\"\"\"\n",
    "\n",
    "for epoch in range(3):\n",
    "    dbow_model.train(utils.shuffle([review for review in tqdm(all_reviews_d2v)]), \n",
    "                     total_examples=len(all_reviews_d2v), \n",
    "                     epochs=1)\n",
    "    dbow_model.alpha -= 0.002\n",
    "    dbow_model.min_alpha = dbow_model.alpha"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {},
   "outputs": [],
   "source": [
    "\"\"\"\n",
    "Function to generate vectors from corpus\n",
    "\"\"\"\n",
    "\n",
    "def generate_vectors(model, corpus, size):\n",
    "    vectors = np.zeros((len(corpus), size))\n",
    "    n = 0\n",
    "    for indx in corpus.index:\n",
    "        prefix = 'all_' + str(indx)\n",
    "        vectors[n] = model.docvecs[prefix]\n",
    "        n += 1\n",
    "    return vectors"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {},
   "outputs": [],
   "source": [
    "\"\"\"\n",
    "Generating train, test and validation vectors\n",
    "\"\"\"\n",
    "\n",
    "train_vectors_dbow = generate_vectors(dbow_model, x_train, 150)\n",
    "test_vectors_dbow = generate_vectors(dbow_model, x_test, 150)\n",
    "validation_vectors_dbow = generate_vectors(dbow_model, x_validation, 150)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Doc2Vec Distributed Bag Of Words(DBOW) Based Sentiment Analysis using LogisticRegression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "**** Doc2Vec Distributed Bag Of Words(DBOW) Based Sentiment Analysis using LogisticRegression ****\n",
      "\n",
      "LogisticRegression Performance : \n",
      "\n",
      "Train-Set Score : 0.8982\n",
      "Train-Set Accuracy : 0.8982\n",
      "\n",
      "Evaluation on Validation-Set : \n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.80      0.69      0.74       762\n",
      "           1       0.93      0.96      0.95      3312\n",
      "\n",
      "   micro avg       0.91      0.91      0.91      4074\n",
      "   macro avg       0.87      0.82      0.84      4074\n",
      "weighted avg       0.91      0.91      0.91      4074\n",
      "\n",
      "Confusion matrix:\n",
      " [[ 524  238]\n",
      " [ 131 3181]]\n",
      "Validation-Set Score : 0.9094\n",
      "Validation-Set Accuracy:0.9094\n",
      "\n",
      "Evaluation on Test-Set : \n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.81      0.64      0.71       833\n",
      "           1       0.91      0.96      0.94      3242\n",
      "\n",
      "   micro avg       0.90      0.90      0.90      4075\n",
      "   macro avg       0.86      0.80      0.82      4075\n",
      "weighted avg       0.89      0.90      0.89      4075\n",
      "\n",
      "Confusion matrix:\n",
      " [[ 530  303]\n",
      " [ 123 3119]]\n",
      "Test-Set Score : 0.8955\n",
      "Test-Set Accuracy:0.8955\n"
     ]
    }
   ],
   "source": [
    "logreg_dbow = LogisticRegression()\n",
    "logreg_dbow.fit(train_vectors_dbow, y_train)\n",
    "print(\"**** Doc2Vec Distributed Bag Of Words(DBOW) Based Sentiment Analysis using LogisticRegression ****\\n\")\n",
    "print(\"LogisticRegression Performance : \\n\")\n",
    "print('Train-Set Score : {:.4f}'.format(logreg_dbow.score(train_vectors_dbow, y_train)))\n",
    "print('Train-Set Accuracy : {:.4f}'.format(accuracy_score(y_train,logreg_dbow.predict(train_vectors_dbow))))\n",
    "\n",
    "print(\"\\nEvaluation on Validation-Set : \")\n",
    "pred_val = logreg_dbow.predict(validation_vectors_dbow)\n",
    "print(\"Classification report:\\n {}\".format(classification_report(y_validation, pred_val)))\n",
    "print(\"Confusion matrix:\\n {}\".format(confusion_matrix(y_validation, pred_val)))\n",
    "print('Validation-Set Score : {:.4f}'.format(logreg_dbow.score(validation_vectors_dbow, y_validation)))\n",
    "print('Validation-Set Accuracy:{:.4f}'.format(accuracy_score(y_validation, pred_val)))\n",
    "\n",
    "print(\"\\nEvaluation on Test-Set : \")\n",
    "pred = logreg_dbow.predict(test_vectors_dbow)\n",
    "print(\"Classification report:\\n {}\".format(classification_report(y_test, pred)))\n",
    "print(\"Confusion matrix:\\n {}\".format(confusion_matrix(y_test, pred)))\n",
    "print('Test-Set Score : {:.4f}'.format(logreg_dbow.score(test_vectors_dbow, y_test)))\n",
    "print('Test-Set Accuracy:{:.4f}'.format(accuracy_score(y_test, pred)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Doc2Vec Distributed Bag Of Words(DBOW) Based Sentiment Analysis using SVC"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "**** Doc2Vec Distributed Bag Of Words(DBOW) Based Sentiment Analysis using SVC ****\n",
      "\n",
      "SVC with linear kernel Performance : \n",
      "\n",
      "Train-Set Score : 0.8985\n",
      "Train-Set Accuracy : 0.8985\n",
      "\n",
      "Evaluation on Validation-Set : \n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.79      0.68      0.73       762\n",
      "           1       0.93      0.96      0.94      3312\n",
      "\n",
      "   micro avg       0.91      0.91      0.91      4074\n",
      "   macro avg       0.86      0.82      0.84      4074\n",
      "weighted avg       0.90      0.91      0.90      4074\n",
      "\n",
      "Confusion matrix:\n",
      " [[ 520  242]\n",
      " [ 136 3176]]\n",
      "Validation-Set Score : 0.9072\n",
      "Validation-Set Accuracy:0.9072\n",
      "\n",
      "Evaluation on Test-Set : \n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.82      0.64      0.72       833\n",
      "           1       0.91      0.96      0.94      3242\n",
      "\n",
      "   micro avg       0.90      0.90      0.90      4075\n",
      "   macro avg       0.86      0.80      0.83      4075\n",
      "weighted avg       0.89      0.90      0.89      4075\n",
      "\n",
      "Confusion matrix:\n",
      " [[ 533  300]\n",
      " [ 120 3122]]\n",
      "Test-Set Score : 0.8969\n",
      "Test-Set Accuracy: 0.8969\n"
     ]
    }
   ],
   "source": [
    "svm_dbow = SVC(kernel='linear')\n",
    "svm_dbow.fit(train_vectors_dbow, y_train)\n",
    "print(\"**** Doc2Vec Distributed Bag Of Words(DBOW) Based Sentiment Analysis using SVC ****\\n\")\n",
    "print(\"SVC with linear kernel Performance : \\n\")\n",
    "print('Train-Set Score : {:.4f}'.format(svm_dbow.score(train_vectors_dbow, y_train)))\n",
    "print('Train-Set Accuracy : {:.4f}'.format(accuracy_score(y_train,svm_dbow.predict(train_vectors_dbow))))\n",
    "\n",
    "print(\"\\nEvaluation on Validation-Set : \")\n",
    "pred_val = svm_dbow.predict(validation_vectors_dbow)\n",
    "print(\"Classification report:\\n {}\".format(classification_report(y_validation, pred_val)))\n",
    "print(\"Confusion matrix:\\n {}\".format(confusion_matrix(y_validation, pred_val)))\n",
    "print('Validation-Set Score : {:.4f}'.format(svm_dbow.score(validation_vectors_dbow, y_validation)))\n",
    "print('Validation-Set Accuracy:{:.4f}'.format(accuracy_score(y_validation, pred_val)))\n",
    "\n",
    "print(\"\\nEvaluation on Test-Set : \")\n",
    "pred = svm_dbow.predict(test_vectors_dbow)\n",
    "print(\"Classification report:\\n {}\".format(classification_report(y_test, pred)))\n",
    "print(\"Confusion matrix:\\n {}\".format(confusion_matrix(y_test, pred)))\n",
    "print('Test-Set Score : {:.4f}'.format(svm_dbow.score(test_vectors_dbow, y_test)))\n",
    "print(\"Test-Set Accuracy: {:.4f}\".format(accuracy_score(y_test, pred)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Doc2Vec Distributed Bag Of Words(DBOW) Based Sentiment Analysis using XGBClassifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "**** Doc2Vec Distributed Bag Of Words(DBOW) Based Sentiment Analysis using XGBClassifier ****\n",
      "\n",
      "XGBClassifier Performance : \n",
      "\n",
      "Train-Set Score : 0.9008\n",
      "Train-Set Accuracy : 0.9008\n",
      "\n",
      "Evaluation on Validation-Set : \n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.80      0.60      0.68       762\n",
      "           1       0.91      0.97      0.94      3312\n",
      "\n",
      "   micro avg       0.90      0.90      0.90      4074\n",
      "   macro avg       0.86      0.78      0.81      4074\n",
      "weighted avg       0.89      0.90      0.89      4074\n",
      "\n",
      "Confusion matrix:\n",
      " [[ 455  307]\n",
      " [ 115 3197]]\n",
      "Validation-Set Score : 0.8964\n",
      "Validation-Set Accuracy:0.8964\n",
      "\n",
      "Evaluation on Test-Set : \n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.82      0.56      0.67       833\n",
      "           1       0.90      0.97      0.93      3242\n",
      "\n",
      "   micro avg       0.88      0.88      0.88      4075\n",
      "   macro avg       0.86      0.76      0.80      4075\n",
      "weighted avg       0.88      0.88      0.88      4075\n",
      "\n",
      "Confusion matrix:\n",
      " [[ 467  366]\n",
      " [ 103 3139]]\n",
      "Test-Set Score : 0.8849\n",
      "Test-Set Accuracy: 0.8849\n"
     ]
    }
   ],
   "source": [
    "xgb_dbow = XGBClassifier()\n",
    "xgb_dbow.fit(train_vectors_dbow, y_train)\n",
    "\n",
    "print(\"**** Doc2Vec Distributed Bag Of Words(DBOW) Based Sentiment Analysis using XGBClassifier ****\\n\")\n",
    "print(\"XGBClassifier Performance : \\n\")\n",
    "print('Train-Set Score : {:.4f}'.format(xgb_dbow.score(train_vectors_dbow, y_train)))\n",
    "print('Train-Set Accuracy : {:.4f}'.format(accuracy_score(y_train,xgb_dbow.predict(train_vectors_dbow))))\n",
    "\n",
    "print(\"\\nEvaluation on Validation-Set : \")\n",
    "pred_val = xgb_dbow.predict(validation_vectors_dbow)\n",
    "print(\"Classification report:\\n {}\".format(classification_report(y_validation, pred_val)))\n",
    "print(\"Confusion matrix:\\n {}\".format(confusion_matrix(y_validation, pred_val)))\n",
    "print('Validation-Set Score : {:.4f}'.format(xgb_dbow.score(validation_vectors_dbow, y_validation)))\n",
    "print('Validation-Set Accuracy:{:.4f}'.format(accuracy_score(y_validation, pred_val)))\n",
    "\n",
    "\n",
    "print(\"\\nEvaluation on Test-Set : \")\n",
    "pred = xgb_dbow.predict(test_vectors_dbow)\n",
    "print(\"Classification report:\\n {}\".format(classification_report(y_test, pred)))\n",
    "print(\"Confusion matrix:\\n {}\".format(confusion_matrix(y_test, pred)))\n",
    "print(\"Test-Set Score : {:.4f}\".format(xgb_dbow.score(test_vectors_dbow, y_test)))\n",
    "print(\"Test-Set Accuracy: {:.4f}\".format(accuracy_score(y_test, pred)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### **Distributed Momory (concatenated)**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "W0506 22:39:05.669103 139801131669312 base_any2vec.py:723] consider setting layer size to a multiple of 4 for greater performance\n",
      "100%|██████████| 40744/40744 [00:00<00:00, 2501943.11it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 2.49 s, sys: 0 ns, total: 2.49 s\n",
      "Wall time: 2.49 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "\"\"\"\n",
    "Create Doc2Vec DMC model\n",
    "\"\"\"\n",
    "\n",
    "dmc_model = Doc2Vec(dm=1, \n",
    "                    dm_concat=1, \n",
    "                    size=150, \n",
    "                    window=10, \n",
    "                    negative=5, \n",
    "                    min_count=2, \n",
    "                    workers=10,\n",
    "                    alpha=0.065,\n",
    "                    min_alpha=0.065)\n",
    "\n",
    "dmc_model.build_vocab([review for review in tqdm(all_reviews_d2v)])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 40744/40744 [00:00<00:00, 2419891.28it/s]\n",
      "100%|██████████| 40744/40744 [00:00<00:00, 1688463.05it/s]\n",
      "100%|██████████| 40744/40744 [00:00<00:00, 2154309.08it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 3min 49s, sys: 4.36 s, total: 3min 53s\n",
      "Wall time: 39.1 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "\n",
    "\"\"\"\n",
    "Train the model\n",
    "\"\"\"\n",
    "\n",
    "for epoch in range(3):\n",
    "    dmc_model.train(utils.shuffle([review for review in tqdm(all_reviews_d2v)]), \n",
    "                    total_examples=len(all_reviews_d2v), \n",
    "                    epochs=1)\n",
    "    dmc_model.alpha -= 0.002\n",
    "    dmc_model.min_alpha = dmc_model.alpha"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {},
   "outputs": [],
   "source": [
    "\"\"\"\n",
    "Generating train, test and validation vectors\n",
    "\"\"\"\n",
    "train_vectors_dmc = generate_vectors(dmc_model, x_train, 150)\n",
    "test_vectors_dmc = generate_vectors(dmc_model, x_test, 150)\n",
    "validation_vectors_dmc = generate_vectors(dmc_model, x_validation, 150)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Doc2Vec DMC Based Sentiment Analysis using LogisticRegression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "**** Doc2Vec Distributed Momory (concatenated) Based Sentiment Analysis using LogisticRegression ****\n",
      "\n",
      "LogisticRegression Performance : \n",
      "\n",
      "Train-Set Score : 0.8088\n",
      "Train-Set Accuracy : 0.8088\n",
      "\n",
      "Evaluation on Validation-Set : \n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.80      0.01      0.01       762\n",
      "           1       0.81      1.00      0.90      3312\n",
      "\n",
      "   micro avg       0.81      0.81      0.81      4074\n",
      "   macro avg       0.81      0.50      0.45      4074\n",
      "weighted avg       0.81      0.81      0.73      4074\n",
      "\n",
      "Confusion matrix:\n",
      " [[   4  758]\n",
      " [   1 3311]]\n",
      "Validation-Set Score : 0.8137\n",
      "Validation-Set Accuracy:0.8137\n",
      "\n",
      "Evaluation on Test-Set : \n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       1.00      0.00      0.00       833\n",
      "           1       0.80      1.00      0.89      3242\n",
      "\n",
      "   micro avg       0.80      0.80      0.80      4075\n",
      "   macro avg       0.90      0.50      0.45      4075\n",
      "weighted avg       0.84      0.80      0.71      4075\n",
      "\n",
      "Confusion matrix:\n",
      " [[   2  831]\n",
      " [   0 3242]]\n",
      "Test-Set Score : 0.7961\n",
      "Test-Set Accuracy:0.7961\n"
     ]
    }
   ],
   "source": [
    "logreg_dmc = LogisticRegression()\n",
    "logreg_dmc.fit(train_vectors_dmc, y_train)\n",
    "print(\"**** Doc2Vec Distributed Momory (concatenated) Based Sentiment Analysis using LogisticRegression ****\\n\")\n",
    "print(\"LogisticRegression Performance : \\n\")\n",
    "print('Train-Set Score : {:.4f}'.format(logreg_dmc.score(train_vectors_dmc, y_train)))\n",
    "print('Train-Set Accuracy : {:.4f}'.format(accuracy_score(y_train,logreg_dmc.predict(train_vectors_dmc))))\n",
    "\n",
    "print(\"\\nEvaluation on Validation-Set : \")\n",
    "pred_val = logreg_dmc.predict(validation_vectors_dmc)\n",
    "print(\"Classification report:\\n {}\".format(classification_report(y_validation, pred_val)))\n",
    "print(\"Confusion matrix:\\n {}\".format(confusion_matrix(y_validation, pred_val)))\n",
    "print('Validation-Set Score : {:.4f}'.format(logreg_dmc.score(validation_vectors_dmc, y_validation)))\n",
    "print('Validation-Set Accuracy:{:.4f}'.format(accuracy_score(y_validation, pred_val)))\n",
    "\n",
    "print(\"\\nEvaluation on Test-Set : \")\n",
    "pred = logreg_dmc.predict(test_vectors_dmc)\n",
    "print(\"Classification report:\\n {}\".format(classification_report(y_test, pred)))\n",
    "print(\"Confusion matrix:\\n {}\".format(confusion_matrix(y_test, pred)))\n",
    "print('Test-Set Score : {:.4f}'.format(logreg_dmc.score(test_vectors_dmc, y_test)))\n",
    "print('Test-Set Accuracy:{:.4f}'.format(accuracy_score(y_test, pred)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Doc2Vec DMC Based Sentiment Analysis using SVC"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "**** Doc2Vec Distributed Momory (concatenated) Based Sentiment Analysis using SVC ****\n",
      "\n",
      "SVC with linear kernel Performance : \n",
      "\n",
      "Train-Set Score : 0.8086\n",
      "Train-Set Accuracy : 0.8086\n",
      "\n",
      "Evaluation on Validation-Set : \n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.00      0.00      0.00       762\n",
      "           1       0.81      1.00      0.90      3312\n",
      "\n",
      "   micro avg       0.81      0.81      0.81      4074\n",
      "   macro avg       0.41      0.50      0.45      4074\n",
      "weighted avg       0.66      0.81      0.73      4074\n",
      "\n",
      "Confusion matrix:\n",
      " [[   0  762]\n",
      " [   0 3312]]\n",
      "Validation-Set Score : 0.8130\n",
      "Validation-Set Accuracy:0.8130\n",
      "\n",
      "Evaluation on Test-Set : \n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.00      0.00      0.00       833\n",
      "           1       0.80      1.00      0.89      3242\n",
      "\n",
      "   micro avg       0.80      0.80      0.80      4075\n",
      "   macro avg       0.40      0.50      0.44      4075\n",
      "weighted avg       0.63      0.80      0.71      4075\n",
      "\n",
      "Confusion matrix:\n",
      " [[   0  833]\n",
      " [   0 3242]]\n",
      "Test-Set Score : 0.7956\n",
      "Test-Set Accuracy:0.7956\n"
     ]
    }
   ],
   "source": [
    "svm_dmc = SVC(kernel='linear')\n",
    "svm_dmc.fit(train_vectors_dmc, y_train)\n",
    "\n",
    "print(\"**** Doc2Vec Distributed Momory (concatenated) Based Sentiment Analysis using SVC ****\\n\")\n",
    "print(\"SVC with linear kernel Performance : \\n\")\n",
    "print('Train-Set Score : {:.4f}'.format(svm_dmc.score(train_vectors_dmc, y_train)))\n",
    "print('Train-Set Accuracy : {:.4f}'.format(accuracy_score(y_train,svm_dmc.predict(train_vectors_dmc))))\n",
    "\n",
    "print(\"\\nEvaluation on Validation-Set : \")\n",
    "pred_val = svm_dmc.predict(validation_vectors_dmc)\n",
    "print(\"Classification report:\\n {}\".format(classification_report(y_validation, pred_val)))\n",
    "print(\"Confusion matrix:\\n {}\".format(confusion_matrix(y_validation, pred_val)))\n",
    "print('Validation-Set Score : {:.4f}'.format(svm_dmc.score(validation_vectors_dmc, y_validation)))\n",
    "print('Validation-Set Accuracy:{:.4f}'.format(accuracy_score(y_validation, pred_val)))\n",
    "\n",
    "print(\"\\nEvaluation on Test-Set : \")\n",
    "pred = svm_dmc.predict(test_vectors_dmc)\n",
    "print(\"Classification report:\\n {}\".format(classification_report(y_test, pred)))\n",
    "print(\"Confusion matrix:\\n {}\".format(confusion_matrix(y_test, pred)))\n",
    "print('Test-Set Score : {:.4f}'.format(svm_dmc.score(test_vectors_dmc, y_test)))\n",
    "print('Test-Set Accuracy:{:.4f}'.format(accuracy_score(y_test, pred)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Doc2Vec DMC Based Sentiment Analysis using XGBClassifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "**** Doc2Vec Distributed Momory (concatenated) Based Sentiment Analysis using XGBClassifier ****\n",
      "\n",
      "XGBClassifier Performance : \n",
      "\n",
      "Train-Set Score : 0.8137\n",
      "Train-Set Accuracy : 0.8137\n",
      "\n",
      "Evaluation on Validation-Set : \n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.76      0.05      0.09       762\n",
      "           1       0.82      1.00      0.90      3312\n",
      "\n",
      "   micro avg       0.82      0.82      0.82      4074\n",
      "   macro avg       0.79      0.52      0.50      4074\n",
      "weighted avg       0.81      0.82      0.75      4074\n",
      "\n",
      "Confusion matrix:\n",
      " [[  38  724]\n",
      " [  12 3300]]\n",
      "Validation-Set Score : 0.8193\n",
      "Validation-Set Accuracy:0.8193\n",
      "\n",
      "Evaluation on Test-Set : \n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.61      0.03      0.06       833\n",
      "           1       0.80      0.99      0.89      3242\n",
      "\n",
      "   micro avg       0.80      0.80      0.80      4075\n",
      "   macro avg       0.70      0.51      0.48      4075\n",
      "weighted avg       0.76      0.80      0.72      4075\n",
      "\n",
      "Confusion matrix:\n",
      " [[  28  805]\n",
      " [  18 3224]]\n",
      "Test-Set Score : 0.7980\n",
      "Test-Set Accuracy:0.7980\n"
     ]
    }
   ],
   "source": [
    "xgb_dmc = XGBClassifier()\n",
    "xgb_dmc.fit(train_vectors_dmc, y_train)\n",
    "\n",
    "print(\"**** Doc2Vec Distributed Momory (concatenated) Based Sentiment Analysis using XGBClassifier ****\\n\")\n",
    "print(\"XGBClassifier Performance : \\n\")\n",
    "print('Train-Set Score : {:.4f}'.format(xgb_dmc.score(train_vectors_dmc, y_train)))\n",
    "print('Train-Set Accuracy : {:.4f}'.format(accuracy_score(y_train,xgb_dmc.predict(train_vectors_dmc))))\n",
    "\n",
    "print(\"\\nEvaluation on Validation-Set : \")\n",
    "pred_val = xgb_dmc.predict(validation_vectors_dmc)\n",
    "print(\"Classification report:\\n {}\".format(classification_report(y_validation, pred_val)))\n",
    "print(\"Confusion matrix:\\n {}\".format(confusion_matrix(y_validation, pred_val)))\n",
    "print('Validation-Set Score : {:.4f}'.format(xgb_dmc.score(validation_vectors_dmc, y_validation)))\n",
    "print('Validation-Set Accuracy:{:.4f}'.format(accuracy_score(y_validation, pred_val)))\n",
    "\n",
    "print(\"\\nEvaluation on Test-Set : \")\n",
    "pred = xgb_dmc.predict(test_vectors_dmc)\n",
    "print(\"Classification report:\\n {}\".format(classification_report(y_test, pred)))\n",
    "print(\"Confusion matrix:\\n {}\".format(confusion_matrix(y_test, pred)))\n",
    "print('Test-Set Score : {:.4f}'.format(xgb_dmc.score(test_vectors_dmc, y_test)))\n",
    "print('Test-Set Accuracy:{:.4f}'.format(accuracy_score(y_test, pred)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Distributed Memory (mean)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "W0506 22:44:19.779132 139801131669312 base_any2vec.py:723] consider setting layer size to a multiple of 4 for greater performance\n",
      "100%|██████████| 40744/40744 [00:00<00:00, 2511171.03it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 2.52 s, sys: 20 ms, total: 2.54 s\n",
      "Wall time: 2.53 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "\"\"\"\n",
    "Create doc2vec DMM model\n",
    "\"\"\"\n",
    "dmm_model = Doc2Vec(dm=1, \n",
    "                    dm_mean=1, \n",
    "                    size=150, \n",
    "                    window=10, \n",
    "                    negative=5, \n",
    "                    min_count=2, \n",
    "                    workers=10, \n",
    "                    alpha=0.065, \n",
    "                    min_alpha=0.065)\n",
    "\n",
    "dmm_model.build_vocab([review for review in tqdm(all_reviews_d2v)])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 40744/40744 [00:00<00:00, 2377141.77it/s]\n",
      "100%|██████████| 40744/40744 [00:00<00:00, 2491256.50it/s]\n",
      "100%|██████████| 40744/40744 [00:00<00:00, 2422017.66it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 41.8 s, sys: 3.66 s, total: 45.4 s\n",
      "Wall time: 14.3 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "\n",
    "\"\"\"\n",
    "Train the model\n",
    "\"\"\"\n",
    "\n",
    "for epoch in range(3):\n",
    "    dmm_model.train(utils.shuffle([review for review in tqdm(all_reviews_d2v)]), \n",
    "                    total_examples=len(all_reviews_d2v), \n",
    "                    epochs=1)\n",
    "    dmm_model.alpha -= 0.002\n",
    "    dmm_model.min_alpha = dmm_model.alpha"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "metadata": {},
   "outputs": [],
   "source": [
    "\"\"\"\n",
    "Generating train, test and validation vectors\n",
    "\"\"\"\n",
    "\n",
    "train_vectors_dmm = generate_vectors(dmm_model, x_train, 150)\n",
    "test_vectors_dmm = generate_vectors(dmm_model, x_test, 150)\n",
    "validation_vectors_dmm = generate_vectors(dmm_model, x_validation, 150)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Doc2Vec DMM Based Sentiment Analysis using LogisticRegression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 90,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "**** Doc2Vec Distributed Memory(mean) Based Sentiment Analysis using LogisticRegression ****\n",
      "\n",
      "LogisticRegression Performance : \n",
      "\n",
      "Train-Set Score : 0.8472\n",
      "Train-Set Accuracy : 0.8472\n",
      "\n",
      "Evaluation on Validation-Set : \n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.74      0.36      0.49       762\n",
      "           1       0.87      0.97      0.92      3312\n",
      "\n",
      "   micro avg       0.86      0.86      0.86      4074\n",
      "   macro avg       0.81      0.67      0.70      4074\n",
      "weighted avg       0.85      0.86      0.84      4074\n",
      "\n",
      "Confusion matrix:\n",
      " [[ 276  486]\n",
      " [  96 3216]]\n",
      "Validation-Set Score : 0.8571\n",
      "Validation-Set Accuracy:0.8571\n",
      "\n",
      "Evaluation on Test-Set : \n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.70      0.30      0.42       833\n",
      "           1       0.84      0.97      0.90      3242\n",
      "\n",
      "   micro avg       0.83      0.83      0.83      4075\n",
      "   macro avg       0.77      0.63      0.66      4075\n",
      "weighted avg       0.81      0.83      0.80      4075\n",
      "\n",
      "Confusion matrix:\n",
      " [[ 246  587]\n",
      " [ 103 3139]]\n",
      "Test-Set Score : 0.8307\n",
      "Test-Set Accuracy:0.8307\n"
     ]
    }
   ],
   "source": [
    "logreg_dmm = LogisticRegression()\n",
    "logreg_dmm.fit(train_vectors_dmm, y_train)\n",
    "print(\"**** Doc2Vec Distributed Memory(mean) Based Sentiment Analysis using LogisticRegression ****\\n\")\n",
    "print(\"LogisticRegression Performance : \\n\")\n",
    "print('Train-Set Score : {:.4f}'.format(logreg_dmm.score(train_vectors_dmm, y_train)))\n",
    "print('Train-Set Accuracy : {:.4f}'.format(accuracy_score(y_train,logreg_dmm.predict(train_vectors_dmm))))\n",
    "\n",
    "print(\"\\nEvaluation on Validation-Set : \")\n",
    "pred_val = logreg_dmm.predict(validation_vectors_dmm)\n",
    "print(\"Classification report:\\n {}\".format(classification_report(y_validation, pred_val)))\n",
    "print(\"Confusion matrix:\\n {}\".format(confusion_matrix(y_validation, pred_val)))\n",
    "print('Validation-Set Score : {:.4f}'.format(logreg_dmm.score(validation_vectors_dmm, y_validation)))\n",
    "print('Validation-Set Accuracy:{:.4f}'.format(accuracy_score(y_validation, pred_val)))\n",
    "\n",
    "print(\"\\nEvaluation on Test-Set : \")\n",
    "pred = logreg_dmm.predict(test_vectors_dmm)\n",
    "print(\"Classification report:\\n {}\".format(classification_report(y_test, pred)))\n",
    "print(\"Confusion matrix:\\n {}\".format(confusion_matrix(y_test, pred)))\n",
    "print('Test-Set Score : {:.4f}'.format(logreg_dmm.score(test_vectors_dmm, y_test)))\n",
    "print('Test-Set Accuracy:{:.4f}'.format(accuracy_score(y_test, pred)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Doc2Vec DMM Based Sentiment Analysis using SVC"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 91,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "**** Doc2Vec Distributed Memory(mean) Based Sentiment Analysis using SVC ****\n",
      "\n",
      "SVC With Linear Kernel Performance : \n",
      "\n",
      "Train-Set Score : 0.8428\n",
      "Train-Set Accuracy : 0.8428\n",
      "\n",
      "Evaluation on Validation-Set : \n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.80      0.30      0.44       762\n",
      "           1       0.86      0.98      0.92      3312\n",
      "\n",
      "   micro avg       0.86      0.86      0.86      4074\n",
      "   macro avg       0.83      0.64      0.68      4074\n",
      "weighted avg       0.85      0.86      0.83      4074\n",
      "\n",
      "Confusion matrix:\n",
      " [[ 231  531]\n",
      " [  58 3254]]\n",
      "Validation-Set Score : 0.8554\n",
      "Validation-Set Accuracy:0.8554\n",
      "\n",
      "Evaluation on Test-Set : \n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.74      0.24      0.36       833\n",
      "           1       0.83      0.98      0.90      3242\n",
      "\n",
      "   micro avg       0.83      0.83      0.83      4075\n",
      "   macro avg       0.79      0.61      0.63      4075\n",
      "weighted avg       0.82      0.83      0.79      4075\n",
      "\n",
      "Confusion matrix:\n",
      " [[ 201  632]\n",
      " [  69 3173]]\n",
      "Test-Set Score : 0.8280\n",
      "Test-Set Accuracy:0.8280\n"
     ]
    }
   ],
   "source": [
    "svm_dmm = SVC(kernel='linear')\n",
    "svm_dmm.fit(train_vectors_dmm, y_train)\n",
    "print(\"**** Doc2Vec Distributed Memory(mean) Based Sentiment Analysis using SVC ****\\n\")\n",
    "print(\"SVC With Linear Kernel Performance : \\n\")\n",
    "print('Train-Set Score : {:.4f}'.format(svm_dmm.score(train_vectors_dmm, y_train)))\n",
    "print('Train-Set Accuracy : {:.4f}'.format(accuracy_score(y_train,svm_dmm.predict(train_vectors_dmm))))\n",
    "\n",
    "print(\"\\nEvaluation on Validation-Set : \")\n",
    "pred_val = svm_dmm.predict(validation_vectors_dmm)\n",
    "print(\"Classification report:\\n {}\".format(classification_report(y_validation, pred_val)))\n",
    "print(\"Confusion matrix:\\n {}\".format(confusion_matrix(y_validation, pred_val)))\n",
    "print('Validation-Set Score : {:.4f}'.format(svm_dmm.score(validation_vectors_dmm, y_validation)))\n",
    "print('Validation-Set Accuracy:{:.4f}'.format(accuracy_score(y_validation, pred_val)))\n",
    "\n",
    "print(\"\\nEvaluation on Test-Set : \")\n",
    "pred = svm_dmm.predict(test_vectors_dmm)\n",
    "print(\"Classification report:\\n {}\".format(classification_report(y_test, pred)))\n",
    "print(\"Confusion matrix:\\n {}\".format(confusion_matrix(y_test, pred)))\n",
    "print('Test-Set Score : {:.4f}'.format(svm_dmm.score(test_vectors_dmm, y_test)))\n",
    "print('Test-Set Accuracy:{:.4f}'.format(accuracy_score(y_test, pred)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Doc2Vec DMM Based Sentiment Analysis using XGBClassifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "**** Doc2Vec Distributed Memory(mean) Based Sentiment Analysis using XGBClassifier ****\n",
      "\n",
      "XGBClassifier Performance : \n",
      "\n",
      "Train-Set Score : 0.8494\n",
      "Train-Set Accuracy : 0.8494\n",
      "\n",
      "Evaluation on Validation-Set : \n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.80      0.28      0.42       762\n",
      "           1       0.86      0.98      0.92      3312\n",
      "\n",
      "   micro avg       0.85      0.85      0.85      4074\n",
      "   macro avg       0.83      0.63      0.67      4074\n",
      "weighted avg       0.85      0.85      0.82      4074\n",
      "\n",
      "Confusion matrix:\n",
      " [[ 216  546]\n",
      " [  55 3257]]\n",
      "Validation-Set Score : 0.8525\n",
      "Validation-Set Accuracy:0.8525\n",
      "\n",
      "Evaluation on Test-Set : \n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.76      0.23      0.36       833\n",
      "           1       0.83      0.98      0.90      3242\n",
      "\n",
      "   micro avg       0.83      0.83      0.83      4075\n",
      "   macro avg       0.80      0.61      0.63      4075\n",
      "weighted avg       0.82      0.83      0.79      4075\n",
      "\n",
      "Confusion matrix:\n",
      " [[ 193  640]\n",
      " [  60 3182]]\n",
      "Test-Set Score : 0.8282\n",
      "Test-Set Accuracy:0.8282\n"
     ]
    }
   ],
   "source": [
    "xgb_dmm = XGBClassifier()\n",
    "xgb_dmm.fit(train_vectors_dmm, y_train)\n",
    "\n",
    "print(\"**** Doc2Vec Distributed Memory(mean) Based Sentiment Analysis using XGBClassifier ****\\n\")\n",
    "print(\"XGBClassifier Performance : \\n\")\n",
    "print('Train-Set Score : {:.4f}'.format(xgb_dmm.score(train_vectors_dmm, y_train)))\n",
    "print('Train-Set Accuracy : {:.4f}'.format(accuracy_score(y_train,xgb_dmm.predict(train_vectors_dmm))))\n",
    "\n",
    "print(\"\\nEvaluation on Validation-Set : \")\n",
    "pred_val = xgb_dmm.predict(validation_vectors_dmm)\n",
    "print(\"Classification report:\\n {}\".format(classification_report(y_validation, pred_val)))\n",
    "print(\"Confusion matrix:\\n {}\".format(confusion_matrix(y_validation, pred_val)))\n",
    "print('Validation-Set Score : {:.4f}'.format(xgb_dmm.score(validation_vectors_dmm, y_validation)))\n",
    "print('Validation-Set Accuracy:{:.4f}'.format(accuracy_score(y_validation, pred_val)))\n",
    "\n",
    "print(\"\\nEvaluation on Test-Set : \")\n",
    "pred = xgb_dmm.predict(test_vectors_dmm)\n",
    "print(\"Classification report:\\n {}\".format(classification_report(y_test, pred)))\n",
    "print(\"Confusion matrix:\\n {}\".format(confusion_matrix(y_test, pred)))\n",
    "print('Test-Set Score : {:.4f}'.format(xgb_dmm.score(test_vectors_dmm, y_test)))\n",
    "print('Test-Set Accuracy:{:.4f}'.format(accuracy_score(y_test, pred)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Combination of Doc2Vec Distributed Bag Of Words (DBOW) And Distributed Memory(Concatenated)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now, I have the document vectors from three different models, now I can concatenate them in combination to see how it affects the performance. Below I defined a function to concatenate document vectors from different models"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 93,
   "metadata": {},
   "outputs": [],
   "source": [
    "\"\"\"\n",
    "Function to concatenate document vectors from different models\n",
    "\"\"\"\n",
    "\n",
    "def generate_concat_vectors(model1,model2, corpus, size):\n",
    "    vectors = np.zeros((len(corpus), size))\n",
    "    n = 0\n",
    "    for indx in corpus.index:\n",
    "        prefix = 'all_' + str(indx)\n",
    "        vectors[n] = np.append(model1.docvecs[prefix],model2.docvecs[prefix])\n",
    "        n += 1\n",
    "    return vectors"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 94,
   "metadata": {},
   "outputs": [],
   "source": [
    "\"\"\"\n",
    "Generating train, test and validation document vectors\n",
    "\"\"\"\n",
    "train_vectors_dbow_dmc = generate_concat_vectors(dbow_model,dmc_model, x_train, 300)\n",
    "test_vectors_dbow_dmc = generate_concat_vectors(dbow_model,dmc_model, x_test, 300)\n",
    "validation_vectors_dbow_dmc = generate_concat_vectors(dbow_model,dmc_model, x_validation, 300)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Combination of Doc2Vec DBOW And DMC Based Sentiment Analysis using LogisticRegression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "**** Combination of Doc2Vec DBOW And DMC Based Sentiment Analysis using LogisticRegression ****\n",
      "\n",
      "LogisticRegression Performance : \n",
      "\n",
      "Train-Set Score : 0.8982\n",
      "Train-Set Accuracy : 0.8982\n",
      "\n",
      "Evaluation on Validation-Set : \n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.80      0.69      0.74       762\n",
      "           1       0.93      0.96      0.95      3312\n",
      "\n",
      "   micro avg       0.91      0.91      0.91      4074\n",
      "   macro avg       0.87      0.83      0.84      4074\n",
      "weighted avg       0.91      0.91      0.91      4074\n",
      "\n",
      "Confusion matrix:\n",
      " [[ 526  236]\n",
      " [ 131 3181]]\n",
      "Validation-Set Score : 0.9099\n",
      "Validation-Set Accuracy:0.9099\n",
      "\n",
      "Evaluation on Test-Set : \n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.81      0.63      0.71       833\n",
      "           1       0.91      0.96      0.94      3242\n",
      "\n",
      "   micro avg       0.89      0.89      0.89      4075\n",
      "   macro avg       0.86      0.80      0.82      4075\n",
      "weighted avg       0.89      0.89      0.89      4075\n",
      "\n",
      "Confusion matrix:\n",
      " [[ 524  309]\n",
      " [ 123 3119]]\n",
      "Test-Set Score : 0.8940\n",
      "Test-Set Accuracy:0.8940\n"
     ]
    }
   ],
   "source": [
    "logreg_dbow_dmc = LogisticRegression()\n",
    "logreg_dbow_dmc.fit(train_vectors_dbow_dmc, y_train)\n",
    "print(\"**** Combination of Doc2Vec DBOW And DMC Based Sentiment Analysis using LogisticRegression ****\\n\")\n",
    "print(\"LogisticRegression Performance : \\n\")\n",
    "print('Train-Set Score : {:.4f}'.format(logreg_dbow_dmc.score(train_vectors_dbow_dmc, y_train)))\n",
    "print('Train-Set Accuracy : {:.4f}'.format(accuracy_score(y_train,logreg_dbow_dmc.predict(train_vectors_dbow_dmc))))\n",
    "\n",
    "print(\"\\nEvaluation on Validation-Set : \")\n",
    "pred_val = logreg_dbow_dmc.predict(validation_vectors_dbow_dmc)\n",
    "print(\"Classification report:\\n {}\".format(classification_report(y_validation, pred_val)))\n",
    "print(\"Confusion matrix:\\n {}\".format(confusion_matrix(y_validation, pred_val)))\n",
    "print('Validation-Set Score : {:.4f}'.format(logreg_dbow_dmc.score(validation_vectors_dbow_dmc, y_validation)))\n",
    "print('Validation-Set Accuracy:{:.4f}'.format(accuracy_score(y_validation, pred_val)))\n",
    "\n",
    "print(\"\\nEvaluation on Test-Set : \")\n",
    "pred = logreg_dbow_dmc.predict(test_vectors_dbow_dmc)\n",
    "print(\"Classification report:\\n {}\".format(classification_report(y_test, pred)))\n",
    "print(\"Confusion matrix:\\n {}\".format(confusion_matrix(y_test, pred)))\n",
    "print('Test-Set Score : {:.4f}'.format(logreg_dbow_dmc.score(test_vectors_dbow_dmc, y_test)))\n",
    "print('Test-Set Accuracy:{:.4f}'.format(accuracy_score(y_test, pred)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Combination of Doc2Vec DBOW And DMC Based Sentiment Analysis using SVC"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 96,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "**** Combination of Doc2Vec DBOW And DMC Based Sentiment Analysis using SVC ****\n",
      "\n",
      "SVC with linear kernel Performance : \n",
      "\n",
      "Train-Set Score : 0.8990\n",
      "Train-Set Accuracy : 0.8990\n",
      "\n",
      "Evaluation on Validation-Set : \n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.79      0.68      0.73       762\n",
      "           1       0.93      0.96      0.94      3312\n",
      "\n",
      "   micro avg       0.91      0.91      0.91      4074\n",
      "   macro avg       0.86      0.82      0.84      4074\n",
      "weighted avg       0.90      0.91      0.90      4074\n",
      "\n",
      "Confusion matrix:\n",
      " [[ 520  242]\n",
      " [ 136 3176]]\n",
      "Validation-Set Score : 0.9072\n",
      "Validation-Set Accuracy:0.9072\n",
      "\n",
      "Evaluation on Test-Set : \n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.82      0.64      0.72       833\n",
      "           1       0.91      0.97      0.94      3242\n",
      "\n",
      "   micro avg       0.90      0.90      0.90      4075\n",
      "   macro avg       0.87      0.80      0.83      4075\n",
      "weighted avg       0.89      0.90      0.89      4075\n",
      "\n",
      "Confusion matrix:\n",
      " [[ 529  304]\n",
      " [ 113 3129]]\n",
      "Test-Set Score : 0.8977\n",
      "Test-Set Accuracy:0.8977\n"
     ]
    }
   ],
   "source": [
    "svm_dbow_dmc = SVC(kernel='linear')\n",
    "svm_dbow_dmc.fit(train_vectors_dbow_dmc, y_train)\n",
    "print(\"**** Combination of Doc2Vec DBOW And DMC Based Sentiment Analysis using SVC ****\\n\")\n",
    "print(\"SVC with linear kernel Performance : \\n\")\n",
    "print('Train-Set Score : {:.4f}'.format(svm_dbow_dmc.score(train_vectors_dbow_dmc, y_train)))\n",
    "print('Train-Set Accuracy : {:.4f}'.format(accuracy_score(y_train,svm_dbow_dmc.predict(train_vectors_dbow_dmc))))\n",
    "\n",
    "print(\"\\nEvaluation on Validation-Set : \")\n",
    "pred_val = svm_dbow_dmc.predict(validation_vectors_dbow_dmc)\n",
    "print(\"Classification report:\\n {}\".format(classification_report(y_validation, pred_val)))\n",
    "print(\"Confusion matrix:\\n {}\".format(confusion_matrix(y_validation, pred_val)))\n",
    "print('Validation-Set Score : {:.4f}'.format(svm_dbow_dmc.score(validation_vectors_dbow_dmc, y_validation)))\n",
    "print('Validation-Set Accuracy:{:.4f}'.format(accuracy_score(y_validation, pred_val)))\n",
    "\n",
    "print(\"\\nEvaluation on Test-Set : \")\n",
    "pred = svm_dbow_dmc.predict(test_vectors_dbow_dmc)\n",
    "print(\"Classification report:\\n {}\".format(classification_report(y_test, pred)))\n",
    "print(\"Confusion matrix:\\n {}\".format(confusion_matrix(y_test, pred)))\n",
    "print('Test-Set Score : {:.4f}'.format(svm_dbow_dmc.score(test_vectors_dbow_dmc, y_test)))\n",
    "print('Test-Set Accuracy:{:.4f}'.format(accuracy_score(y_test, pred)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Combination of Doc2Vec DBOW And DMC Based Sentiment Analysis using XGBClassifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 97,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "**** Combination of Doc2Vec DBOW And DMC Based Sentiment Analysis using XGBClassifier ****\n",
      "\n",
      "XGBClassifier Performance : \n",
      "\n",
      "Train-Set Score : 0.9000\n",
      "Train-Set Accuracy : 0.9000\n",
      "\n",
      "Evaluation on Validation-Set : \n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.80      0.60      0.69       762\n",
      "           1       0.91      0.97      0.94      3312\n",
      "\n",
      "   micro avg       0.90      0.90      0.90      4074\n",
      "   macro avg       0.86      0.78      0.81      4074\n",
      "weighted avg       0.89      0.90      0.89      4074\n",
      "\n",
      "Confusion matrix:\n",
      " [[ 457  305]\n",
      " [ 115 3197]]\n",
      "Validation-Set Score : 0.8969\n",
      "Validation-Set Accuracy:0.8969\n",
      "\n",
      "Evaluation on Test-Set : \n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.81      0.57      0.67       833\n",
      "           1       0.90      0.97      0.93      3242\n",
      "\n",
      "   micro avg       0.88      0.88      0.88      4075\n",
      "   macro avg       0.86      0.77      0.80      4075\n",
      "weighted avg       0.88      0.88      0.88      4075\n",
      "\n",
      "Confusion matrix:\n",
      " [[ 472  361]\n",
      " [ 108 3134]]\n",
      "Test-Set Score : 0.8849\n",
      "Test-Set Accuracy:0.8849\n"
     ]
    }
   ],
   "source": [
    "xgb_dbow_dmc =XGBClassifier()\n",
    "xgb_dbow_dmc.fit(train_vectors_dbow_dmc, y_train)\n",
    "print(\"**** Combination of Doc2Vec DBOW And DMC Based Sentiment Analysis using XGBClassifier ****\\n\")\n",
    "print(\"XGBClassifier Performance : \\n\")\n",
    "print('Train-Set Score : {:.4f}'.format(xgb_dbow_dmc.score(train_vectors_dbow_dmc, y_train)))\n",
    "print('Train-Set Accuracy : {:.4f}'.format(accuracy_score(y_train,xgb_dbow_dmc.predict(train_vectors_dbow_dmc))))\n",
    "\n",
    "print(\"\\nEvaluation on Validation-Set : \")\n",
    "pred_val = xgb_dbow_dmc.predict(validation_vectors_dbow_dmc)\n",
    "print(\"Classification report:\\n {}\".format(classification_report(y_validation, pred_val)))\n",
    "print(\"Confusion matrix:\\n {}\".format(confusion_matrix(y_validation, pred_val)))\n",
    "print('Validation-Set Score : {:.4f}'.format(xgb_dbow_dmc.score(validation_vectors_dbow_dmc, y_validation)))\n",
    "print('Validation-Set Accuracy:{:.4f}'.format(accuracy_score(y_validation, pred_val)))\n",
    "\n",
    "print(\"\\nEvaluation on Test-Set : \")\n",
    "pred = xgb_dbow_dmc.predict(test_vectors_dbow_dmc)\n",
    "print(\"Classification report:\\n {}\".format(classification_report(y_test, pred)))\n",
    "print(\"Confusion matrix:\\n {}\".format(confusion_matrix(y_test, pred)))\n",
    "print('Test-Set Score : {:.4f}'.format(xgb_dbow_dmc.score(test_vectors_dbow_dmc, y_test)))\n",
    "print('Test-Set Accuracy:{:.4f}'.format(accuracy_score(y_test, pred)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Combination of Doc2Vec Distributed Bag Of Words (DBOW) And Distributed Memory(Mean)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 98,
   "metadata": {},
   "outputs": [],
   "source": [
    "\"\"\"\n",
    "Generating train, test and validation document vectors\n",
    "\"\"\"\n",
    "train_vectors_dbow_dmm = generate_concat_vectors(dbow_model,dmm_model, x_train, 300)\n",
    "test_vectors_dbow_dmm = generate_concat_vectors(dbow_model,dmm_model, x_test, 300)\n",
    "validation_vectors_dbow_dmm = generate_concat_vectors(dbow_model,dmm_model, x_validation, 300)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Combination of Doc2Vec DBOW And DMM Based Sentiment Analysis using LogisticRegression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 99,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "**** Combination of Doc2Vec DBOW And DMM Based Sentiment Analysis using LogisticRegression ****\n",
      "\n",
      "LogisticRegression Performance : \n",
      "\n",
      "Train-Set Score : 0.9024\n",
      "Train-Set Accuracy : 0.9024\n",
      "\n",
      "Evaluation on Validation-Set : \n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.80      0.69      0.74       762\n",
      "           1       0.93      0.96      0.95      3312\n",
      "\n",
      "   micro avg       0.91      0.91      0.91      4074\n",
      "   macro avg       0.87      0.82      0.84      4074\n",
      "weighted avg       0.91      0.91      0.91      4074\n",
      "\n",
      "Confusion matrix:\n",
      " [[ 522  240]\n",
      " [ 130 3182]]\n",
      "Validation-Set Score : 0.9092\n",
      "Validation-Set Accuracy:0.9092\n",
      "\n",
      "Evaluation on Test-Set : \n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.82      0.65      0.72       833\n",
      "           1       0.91      0.96      0.94      3242\n",
      "\n",
      "   micro avg       0.90      0.90      0.90      4075\n",
      "   macro avg       0.87      0.81      0.83      4075\n",
      "weighted avg       0.90      0.90      0.89      4075\n",
      "\n",
      "Confusion matrix:\n",
      " [[ 538  295]\n",
      " [ 115 3127]]\n",
      "Test-Set Score : 0.8994\n",
      "Test-Set Accuracy:0.8994\n"
     ]
    }
   ],
   "source": [
    "logreg_dbow_dmm = LogisticRegression()\n",
    "logreg_dbow_dmm.fit(train_vectors_dbow_dmm, y_train)\n",
    "print(\"**** Combination of Doc2Vec DBOW And DMM Based Sentiment Analysis using LogisticRegression ****\\n\")\n",
    "print(\"LogisticRegression Performance : \\n\")\n",
    "print('Train-Set Score : {:.4f}'.format(logreg_dbow_dmm.score(train_vectors_dbow_dmm, y_train)))\n",
    "print('Train-Set Accuracy : {:.4f}'.format(accuracy_score(y_train,logreg_dbow_dmm.predict(train_vectors_dbow_dmm))))\n",
    "\n",
    "print(\"\\nEvaluation on Validation-Set : \")\n",
    "pred_val = logreg_dbow_dmm.predict(validation_vectors_dbow_dmm)\n",
    "print(\"Classification report:\\n {}\".format(classification_report(y_validation, pred_val)))\n",
    "print(\"Confusion matrix:\\n {}\".format(confusion_matrix(y_validation, pred_val)))\n",
    "print('Validation-Set Score : {:.4f}'.format(logreg_dbow_dmm.score(validation_vectors_dbow_dmm, y_validation)))\n",
    "print('Validation-Set Accuracy:{:.4f}'.format(accuracy_score(y_validation, pred_val)))\n",
    "\n",
    "print(\"\\nEvaluation on Test-Set : \")\n",
    "pred = logreg_dbow_dmm.predict(test_vectors_dbow_dmm)\n",
    "print(\"Classification report:\\n {}\".format(classification_report(y_test, pred)))\n",
    "print(\"Confusion matrix:\\n {}\".format(confusion_matrix(y_test, pred)))\n",
    "print('Test-Set Score : {:.4f}'.format(logreg_dbow_dmm.score(test_vectors_dbow_dmm, y_test)))\n",
    "print('Test-Set Accuracy:{:.4f}'.format(accuracy_score(y_test, pred)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Combination of Doc2Vec DBOW And DMM Based Sentiment Analysis using SVC"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 100,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "**** Combination of Doc2Vec DBOW And DMM Based Sentiment Analysis using SVC ****\n",
      "\n",
      "SVC with linear kernel Performance : \n",
      "\n",
      "Train-Set Score : 0.9033\n",
      "Train-Set Accuracy : 0.9033\n",
      "\n",
      "Evaluation on Validation-Set : \n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.80      0.69      0.74       762\n",
      "           1       0.93      0.96      0.95      3312\n",
      "\n",
      "   micro avg       0.91      0.91      0.91      4074\n",
      "   macro avg       0.86      0.82      0.84      4074\n",
      "weighted avg       0.91      0.91      0.91      4074\n",
      "\n",
      "Confusion matrix:\n",
      " [[ 523  239]\n",
      " [ 131 3181]]\n",
      "Validation-Set Score : 0.9092\n",
      "Validation-Set Accuracy:0.9092\n",
      "\n",
      "Evaluation on Test-Set : \n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.83      0.65      0.73       833\n",
      "           1       0.91      0.97      0.94      3242\n",
      "\n",
      "   micro avg       0.90      0.90      0.90      4075\n",
      "   macro avg       0.87      0.81      0.83      4075\n",
      "weighted avg       0.90      0.90      0.90      4075\n",
      "\n",
      "Confusion matrix:\n",
      " [[ 538  295]\n",
      " [ 112 3130]]\n",
      "Test-Set Score : 0.9001\n",
      "Test-Set Accuracy:0.9001\n"
     ]
    }
   ],
   "source": [
    "svm_dbow_dmm = SVC(kernel='linear')\n",
    "svm_dbow_dmm.fit(train_vectors_dbow_dmm, y_train)\n",
    "print(\"**** Combination of Doc2Vec DBOW And DMM Based Sentiment Analysis using SVC ****\\n\")\n",
    "print(\"SVC with linear kernel Performance : \\n\")\n",
    "print('Train-Set Score : {:.4f}'.format(svm_dbow_dmm.score(train_vectors_dbow_dmm, y_train)))\n",
    "print('Train-Set Accuracy : {:.4f}'.format(accuracy_score(y_train,svm_dbow_dmm.predict(train_vectors_dbow_dmm))))\n",
    "\n",
    "print(\"\\nEvaluation on Validation-Set : \")\n",
    "pred_val = svm_dbow_dmm.predict(validation_vectors_dbow_dmm)\n",
    "print(\"Classification report:\\n {}\".format(classification_report(y_validation, pred_val)))\n",
    "print(\"Confusion matrix:\\n {}\".format(confusion_matrix(y_validation, pred_val)))\n",
    "print('Validation-Set Score : {:.4f}'.format(svm_dbow_dmm.score(validation_vectors_dbow_dmm, y_validation)))\n",
    "print('Validation-Set Accuracy:{:.4f}'.format(accuracy_score(y_validation, pred_val)))\n",
    "\n",
    "print(\"\\nEvaluation on Test-Set : \")\n",
    "pred = svm_dbow_dmm.predict(test_vectors_dbow_dmm)\n",
    "print(\"Classification report:\\n {}\".format(classification_report(y_test, pred)))\n",
    "print(\"Confusion matrix:\\n {}\".format(confusion_matrix(y_test, pred)))\n",
    "print('Test-Set Score : {:.4f}'.format(svm_dbow_dmm.score(test_vectors_dbow_dmm, y_test)))\n",
    "print('Test-Set Accuracy:{:.4f}'.format(accuracy_score(y_test, pred)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Combination of Doc2Vec DBOW And DMM Based Sentiment Analysis using XGBClassifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 101,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "**** Combination of Doc2Vec DBOW And DMM Based Sentiment Analysis using XGBClassifier ****\n",
      "\n",
      "XGBClassifier Performance : \n",
      "\n",
      "Train-Set Score : 0.9016\n",
      "Train-Set Accuracy : 0.9016\n",
      "\n",
      "Evaluation on Validation-Set : \n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.81      0.60      0.69       762\n",
      "           1       0.91      0.97      0.94      3312\n",
      "\n",
      "   micro avg       0.90      0.90      0.90      4074\n",
      "   macro avg       0.86      0.78      0.81      4074\n",
      "weighted avg       0.89      0.90      0.89      4074\n",
      "\n",
      "Confusion matrix:\n",
      " [[ 457  305]\n",
      " [ 108 3204]]\n",
      "Validation-Set Score : 0.8986\n",
      "Validation-Set Accuracy:0.8986\n",
      "\n",
      "Evaluation on Test-Set : \n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.82      0.56      0.67       833\n",
      "           1       0.90      0.97      0.93      3242\n",
      "\n",
      "   micro avg       0.89      0.89      0.89      4075\n",
      "   macro avg       0.86      0.76      0.80      4075\n",
      "weighted avg       0.88      0.89      0.88      4075\n",
      "\n",
      "Confusion matrix:\n",
      " [[ 465  368]\n",
      " [  99 3143]]\n",
      "Test-Set Score : 0.8854\n",
      "Test-Set Accuracy:0.8854\n"
     ]
    }
   ],
   "source": [
    "xgb_dbow_dmm =XGBClassifier()\n",
    "xgb_dbow_dmm.fit(train_vectors_dbow_dmm, y_train)\n",
    "print(\"**** Combination of Doc2Vec DBOW And DMM Based Sentiment Analysis using XGBClassifier ****\\n\")\n",
    "print(\"XGBClassifier Performance : \\n\")\n",
    "print('Train-Set Score : {:.4f}'.format(xgb_dbow_dmm.score(train_vectors_dbow_dmm, y_train)))\n",
    "print('Train-Set Accuracy : {:.4f}'.format(accuracy_score(y_train,xgb_dbow_dmm.predict(train_vectors_dbow_dmm))))\n",
    "\n",
    "print(\"\\nEvaluation on Validation-Set : \")\n",
    "pred_val = xgb_dbow_dmm.predict(validation_vectors_dbow_dmm)\n",
    "print(\"Classification report:\\n {}\".format(classification_report(y_validation, pred_val)))\n",
    "print(\"Confusion matrix:\\n {}\".format(confusion_matrix(y_validation, pred_val)))\n",
    "print('Validation-Set Score : {:.4f}'.format(xgb_dbow_dmm.score(validation_vectors_dbow_dmm, y_validation)))\n",
    "print('Validation-Set Accuracy:{:.4f}'.format(accuracy_score(y_validation, pred_val)))\n",
    "\n",
    "print(\"\\nEvaluation on Test-Set : \")\n",
    "pred = xgb_dbow_dmm.predict(test_vectors_dbow_dmm)\n",
    "print(\"Classification report:\\n {}\".format(classification_report(y_test, pred)))\n",
    "print(\"Confusion matrix:\\n {}\".format(confusion_matrix(y_test, pred)))\n",
    "print('Test-Set Score : {:.4f}'.format(xgb_dbow_dmm.score(test_vectors_dbow_dmm, y_test)))\n",
    "print('Test-Set Accuracy:{:.4f}'.format(accuracy_score(y_test, pred)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Sentiment Analysis Using Combination of Doc2Vec DBOW And DMC Document Embedding and Keras Neural Network"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 102,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "dense_6 (Dense)              (None, 128)               38528     \n",
      "_________________________________________________________________\n",
      "dense_7 (Dense)              (None, 128)               16512     \n",
      "_________________________________________________________________\n",
      "dense_8 (Dense)              (None, 1)                 129       \n",
      "=================================================================\n",
      "Total params: 55,169\n",
      "Trainable params: 55,169\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "\"\"\"\n",
    "Create model\n",
    "\"\"\"\n",
    "keras_d2v_combo_dbow_dmc_model = Sequential()\n",
    "keras_d2v_combo_dbow_dmc_model.add(Dense(128, activation='relu', input_dim=300))\n",
    "keras_d2v_combo_dbow_dmc_model.add(Dense(128, activation='relu'))\n",
    "keras_d2v_combo_dbow_dmc_model.add(Dense(1, activation='sigmoid'))\n",
    "\n",
    "\"\"\"\n",
    "Compile network\n",
    "\"\"\"\n",
    "keras_d2v_combo_dbow_dmc_model.compile(optimizer='adam',\n",
    "                                       loss='binary_crossentropy',\n",
    "                                       metrics=['accuracy'])\n",
    "keras_d2v_combo_dbow_dmc_model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 103,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "**** Sentiment Analysis Using Combination of Doc2Vec DBOW And DMC Document Embedding and Keras Neural Network ****\n",
      "\n",
      "Train-Set Accuracy: 0.9312\n",
      "\n",
      "Evaluation on Validation-Set : \n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.78      0.68      0.73       762\n",
      "           1       0.93      0.95      0.94      3312\n",
      "\n",
      "   micro avg       0.90      0.90      0.90      4074\n",
      "   macro avg       0.85      0.82      0.83      4074\n",
      "weighted avg       0.90      0.90      0.90      4074\n",
      "\n",
      "Confusion matrix:\n",
      " [[ 521  241]\n",
      " [ 150 3162]]\n",
      "Validation-Set Accuracy: 0.9040\n",
      "\n",
      "Evaluation on Test-Set : \n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.82      0.68      0.74       833\n",
      "           1       0.92      0.96      0.94      3242\n",
      "\n",
      "   micro avg       0.90      0.90      0.90      4075\n",
      "   macro avg       0.87      0.82      0.84      4075\n",
      "weighted avg       0.90      0.90      0.90      4075\n",
      "\n",
      "Confusion matrix:\n",
      " [[ 564  269]\n",
      " [ 125 3117]]\n",
      "Test-Set Accuracy:  0.9033\n"
     ]
    }
   ],
   "source": [
    "\"\"\"\n",
    "Fit network\n",
    "\"\"\"\n",
    "keras_d2v_combo_dbow_dmc_model.fit(train_vectors_dbow_dmc, y_train,\n",
    "                                   validation_data=(validation_vectors_dbow_dmc, y_validation),\n",
    "                                   epochs=5, \n",
    "                                   batch_size=10, \n",
    "                                   verbose=False)\n",
    "\n",
    "\"\"\"\n",
    "Evaluate\n",
    "\"\"\"\n",
    "print(\"**** Sentiment Analysis Using Combination of Doc2Vec DBOW And DMC Document Embedding and Keras Neural Network ****\\n\")\n",
    "loss, accuracy = keras_d2v_combo_dbow_dmc_model.evaluate(train_vectors_dbow_dmc, y_train, verbose=False)\n",
    "print(\"Train-Set Accuracy: {:.4f}\".format(accuracy))\n",
    "\n",
    "print(\"\\nEvaluation on Validation-Set : \")\n",
    "pred_val=keras_d2v_combo_dbow_dmc_model.predict_classes(validation_vectors_dbow_dmc)\n",
    "print(\"Classification report:\\n {}\".format(classification_report(y_validation, pred_val)))\n",
    "print(\"Confusion matrix:\\n {}\".format(confusion_matrix(y_validation, pred_val)))\n",
    "loss, accuracy = keras_d2v_combo_dbow_dmc_model.evaluate(validation_vectors_dbow_dmc, y_validation, verbose=False)\n",
    "print(\"Validation-Set Accuracy: {:.4f}\".format(accuracy))\n",
    "\n",
    "print(\"\\nEvaluation on Test-Set : \")\n",
    "pred=keras_d2v_combo_dbow_dmc_model.predict_classes(test_vectors_dbow_dmc)\n",
    "print(\"Classification report:\\n {}\".format(classification_report(y_test, pred)))\n",
    "print(\"Confusion matrix:\\n {}\".format(confusion_matrix(y_test, pred)))\n",
    "\n",
    "loss, accuracy = keras_d2v_combo_dbow_dmc_model.evaluate(test_vectors_dbow_dmc, y_test, verbose=False)\n",
    "print(\"Test-Set Accuracy:  {:.4f}\".format(accuracy))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Sentiment Analysis Using Combination of Doc2Vec DBOW And DMM Document Embedding and Keras Neural Network"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "dense_9 (Dense)              (None, 128)               38528     \n",
      "_________________________________________________________________\n",
      "dense_10 (Dense)             (None, 128)               16512     \n",
      "_________________________________________________________________\n",
      "dense_11 (Dense)             (None, 1)                 129       \n",
      "=================================================================\n",
      "Total params: 55,169\n",
      "Trainable params: 55,169\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "\"\"\"\n",
    "Create model\n",
    "\"\"\"\n",
    "keras_d2v_combo_dbow_dmm_model = Sequential()\n",
    "keras_d2v_combo_dbow_dmm_model.add(Dense(128, activation='relu', input_dim=300))\n",
    "keras_d2v_combo_dbow_dmm_model.add(Dense(128, activation='relu'))\n",
    "keras_d2v_combo_dbow_dmm_model.add(Dense(1, activation='sigmoid'))\n",
    "\n",
    "\"\"\"\n",
    "Compile network\n",
    "\"\"\"\n",
    "keras_d2v_combo_dbow_dmm_model.compile(optimizer='adam',\n",
    "                                       loss='binary_crossentropy',\n",
    "                                       metrics=['accuracy'])\n",
    "keras_d2v_combo_dbow_dmm_model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 105,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "**** Sentiment Analysis Using Combination of Doc2Vec DBOW And DMM Document Embedding and Keras Neural Network ****\n",
      "\n",
      "Train-Set Accuracy: 0.9475\n",
      "\n",
      "Evaluation on Validation-Set : \n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.80      0.74      0.77       762\n",
      "           1       0.94      0.96      0.95      3312\n",
      "\n",
      "   micro avg       0.92      0.92      0.92      4074\n",
      "   macro avg       0.87      0.85      0.86      4074\n",
      "weighted avg       0.91      0.92      0.91      4074\n",
      "\n",
      "Confusion matrix:\n",
      " [[ 563  199]\n",
      " [ 145 3167]]\n",
      "Validation-Set Accuracy: 0.9156\n",
      "\n",
      "Evaluation on Test-Set : \n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.79      0.69      0.73       833\n",
      "           1       0.92      0.95      0.94      3242\n",
      "\n",
      "   micro avg       0.90      0.90      0.90      4075\n",
      "   macro avg       0.86      0.82      0.84      4075\n",
      "weighted avg       0.89      0.90      0.90      4075\n",
      "\n",
      "Confusion matrix:\n",
      " [[ 572  261]\n",
      " [ 153 3089]]\n",
      "Test-Set Accuracy:  0.8984\n"
     ]
    }
   ],
   "source": [
    "\"\"\"\n",
    "Fit network\n",
    "\"\"\"\n",
    "keras_d2v_combo_dbow_dmm_model.fit(train_vectors_dbow_dmm, y_train,\n",
    "                                   validation_data=(validation_vectors_dbow_dmm, y_validation),\n",
    "                                   epochs=5, \n",
    "                                   batch_size=10, \n",
    "                                   verbose=False)\n",
    "\n",
    "\"\"\"\n",
    "Evaluate\n",
    "\"\"\"\n",
    "print(\"**** Sentiment Analysis Using Combination of Doc2Vec DBOW And DMM Document Embedding and Keras Neural Network ****\\n\")\n",
    "loss, accuracy = keras_d2v_combo_dbow_dmm_model.evaluate(train_vectors_dbow_dmm, y_train, verbose=False)\n",
    "print(\"Train-Set Accuracy: {:.4f}\".format(accuracy))\n",
    "\n",
    "print(\"\\nEvaluation on Validation-Set : \")\n",
    "pred_val=keras_d2v_combo_dbow_dmm_model.predict_classes(validation_vectors_dbow_dmm)\n",
    "print(\"Classification report:\\n {}\".format(classification_report(y_validation, pred_val)))\n",
    "print(\"Confusion matrix:\\n {}\".format(confusion_matrix(y_validation, pred_val)))\n",
    "loss, accuracy = keras_d2v_combo_dbow_dmm_model.evaluate(validation_vectors_dbow_dmm, y_validation, verbose=False)\n",
    "print(\"Validation-Set Accuracy: {:.4f}\".format(accuracy))\n",
    "\n",
    "print(\"\\nEvaluation on Test-Set : \")\n",
    "pred=keras_d2v_combo_dbow_dmm_model.predict_classes(test_vectors_dbow_dmm)\n",
    "print(\"Classification report:\\n {}\".format(classification_report(y_test, pred)))\n",
    "print(\"Confusion matrix:\\n {}\".format(confusion_matrix(y_test, pred)))\n",
    "\n",
    "loss, accuracy = keras_d2v_combo_dbow_dmm_model.evaluate(test_vectors_dbow_dmm, y_test, verbose=False)\n",
    "print(\"Test-Set Accuracy:  {:.4f}\".format(accuracy))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 106,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Time elapsed :  4.87  in hours\n"
     ]
    }
   ],
   "source": [
    "print(\"Time elapsed : \",(round(((time.time()-program_start_time)/3600),2)),\" in hours\")"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}