SentimentAnalysis.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Lab 8: Implement Your Machine Learning Project Plan"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In this lab assignment, you will implement the machine learning project plan you created in the written assignment. You will:\n",
    "\n",
    "1. Load your data set and save it to a Pandas DataFrame.\n",
    "2. Perform exploratory data analysis on your data to determine which feature engineering and data preparation techniques you will use.\n",
    "3. Prepare your data for your model and create features and a label.\n",
    "4. Fit your model to the training data and evaluate your model.\n",
    "5. Improve your model by performing model selection and/or feature selection techniques to find best model for your problem."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Import Packages\n",
    "\n",
    "Before you get started, import a few packages."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import os \n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<b>Task:</b> In the code cell below, import additional packages that you have used in this course that you will need for this task."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.feature_extraction.text import TfidfVectorizer\n",
    "from sklearn.model_selection import train_test_split, cross_val_score\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.model_selection import GridSearchCV\n",
    "from sklearn.metrics import classification_report, accuracy_score, confusion_matrix\n",
    "from sklearn.svm import SVC\n",
    "from sklearn.ensemble import RandomForestClassifier, VotingClassifier, GradientBoostingClassifier\n",
    "from sklearn.naive_bayes import MultinomialNB\n",
    "from sklearn.neighbors import KNeighborsClassifier\n",
    "from sklearn.feature_selection import SelectKBest, chi2\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Part 1: Load the Data Set\n",
    "\n",
    "\n",
    "You have chosen to work with one of four data sets. The data sets are located in a folder named \"data.\" The file names of the three data sets are as follows:\n",
    "\n",
    "* The \"adult\" data set that contains Census information from 1994 is located in file `adultData.csv`\n",
    "* The airbnb NYC \"listings\" data set is located in file  `airbnbListingsData.csv`\n",
    "* The World Happiness Report (WHR) data set is located in file `WHR2018Chapter2OnlineData.csv`\n",
    "* The book review data set is located in file `bookReviewsData.csv`\n",
    "\n",
    "\n",
    "\n",
    "<b>Task:</b> In the code cell below, use the same method you have been using to load your data using `pd.read_csv()` and save it to DataFrame `df`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "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",
       "      <th>Positive Review</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>This was perhaps the best of Johannes Steinhof...</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>This very fascinating book is a story written ...</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>The four tales in this collection are beautifu...</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>The book contained more profanity than I expec...</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>We have now entered a second time of deep conc...</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1968</th>\n",
       "      <td>I purchased the book with the intention of tea...</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1969</th>\n",
       "      <td>There are so many design books, but the Graphi...</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1970</th>\n",
       "      <td>I am thilled to see this book being available ...</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1971</th>\n",
       "      <td>As many have stated before me the book starts ...</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1972</th>\n",
       "      <td>I love this book! It is a terrific blend of ha...</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>1973 rows × 2 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                 Review  Positive Review\n",
       "0     This was perhaps the best of Johannes Steinhof...             True\n",
       "1     This very fascinating book is a story written ...             True\n",
       "2     The four tales in this collection are beautifu...             True\n",
       "3     The book contained more profanity than I expec...            False\n",
       "4     We have now entered a second time of deep conc...             True\n",
       "...                                                 ...              ...\n",
       "1968  I purchased the book with the intention of tea...             True\n",
       "1969  There are so many design books, but the Graphi...             True\n",
       "1970  I am thilled to see this book being available ...             True\n",
       "1971  As many have stated before me the book starts ...            False\n",
       "1972  I love this book! It is a terrific blend of ha...             True\n",
       "\n",
       "[1973 rows x 2 columns]"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# print(os.getcwd())\n",
    "\n",
    "df=pd.read_csv('data/bookReviewsData.csv')\n",
    "df"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Part 2: Exploratory Data Analysis\n",
    "\n",
    "The next step is to inspect and analyze your data set with your machine learning problem and project plan in mind. \n",
    "\n",
    "This step will help you determine data preparation and feature engineering techniques you will need to apply to your data to build a balanced modeling data set for your problem and model. These data preparation techniques may include:\n",
    "* addressing missingness, such as replacing missing values with means\n",
    "* renaming features and labels\n",
    "* finding and replacing outliers\n",
    "* performing winsorization if needed\n",
    "* performing one-hot encoding on categorical features\n",
    "* performing vectorization for an NLP problem\n",
    "* addressing class imbalance in your data sample to promote fair AI\n",
    "\n",
    "\n",
    "Think of the different techniques you have used to inspect and analyze your data in this course. These include using Pandas to apply data filters, using the Pandas `describe()` method to get insight into key statistics for each column, using the Pandas `dtypes` property to inspect the data type of each column, and using Matplotlib and Seaborn to detect outliers and visualize relationships between features and labels. If you are working on a classification problem, use techniques you have learned to determine if there is class imbalance.\n",
    "\n",
    "\n",
    "<b>Task</b>: Use the techniques you have learned in this course to inspect and analyze your data. \n",
    "\n",
    "<b>Note</b>: You can add code cells if needed by going to the <b>Insert</b> menu and clicking on <b>Insert Cell Below</b> in the drop-drown menu."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'using the describe method, it turns out there are only 1865 \\nunique reviews out of 1973, indicating that there are some\\nduplicate reviews.In an effort to prepare the data through preprocessing, \\nwe must remove duplicates'"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.describe()\n",
    "\n",
    "'''using the describe method, it turns out there are only 1865 \n",
    "unique reviews out of 1973, indicating that there are some\n",
    "duplicate reviews.In an effort to prepare the data through preprocessing, \n",
    "we must remove duplicates'''"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "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",
       "      <th>Positive Review</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>1865</td>\n",
       "      <td>1865</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>unique</th>\n",
       "      <td>1865</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>top</th>\n",
       "      <td>So many disbelievers out there! I feel so sorr...</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>freq</th>\n",
       "      <td>1</td>\n",
       "      <td>979</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                   Review Positive Review\n",
       "count                                                1865            1865\n",
       "unique                                               1865               2\n",
       "top     So many disbelievers out there! I feel so sorr...            True\n",
       "freq                                                    1             979"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df=df.drop_duplicates(subset=\"Review\", keep=\"first\")\n",
    "df.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "Int64Index: 1865 entries, 0 to 1972\n",
      "Data columns (total 2 columns):\n",
      " #   Column           Non-Null Count  Dtype \n",
      "---  ------           --------------  ----- \n",
      " 0   Review           1865 non-null   object\n",
      " 1   Positive Review  1865 non-null   bool  \n",
      "dtypes: bool(1), object(1)\n",
      "memory usage: 31.0+ KB\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "'the results indicate that there are no nulls in our file'"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.info()\n",
    "'''the results indicate that there are no nulls in our file'''"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1865, 2)"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "True     979\n",
      "False    886\n",
      "Name: Positive Review, dtype: int64\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "'there are 979 positive reviews, and 886 negative reviews'"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print(df['Positive Review'].value_counts())\n",
    "'''there are 979 positive reviews, and 886 negative reviews'''"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Review             0\n",
       "Positive Review    0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "'''checking for missing values'''\n",
    "df.isnull().sum()\n",
    "\n",
    "#there are no missing values"
   ]
  },
  {
   "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>Review_Text</th>\n",
       "      <th>Is_Positive</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>This was perhaps the best of Johannes Steinhof...</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>This very fascinating book is a story written ...</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>The four tales in this collection are beautifu...</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>The book contained more profanity than I expec...</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>We have now entered a second time of deep conc...</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1968</th>\n",
       "      <td>I purchased the book with the intention of tea...</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1969</th>\n",
       "      <td>There are so many design books, but the Graphi...</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1970</th>\n",
       "      <td>I am thilled to see this book being available ...</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1971</th>\n",
       "      <td>As many have stated before me the book starts ...</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1972</th>\n",
       "      <td>I love this book! It is a terrific blend of ha...</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>1865 rows × 2 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                            Review_Text  Is_Positive\n",
       "0     This was perhaps the best of Johannes Steinhof...         True\n",
       "1     This very fascinating book is a story written ...         True\n",
       "2     The four tales in this collection are beautifu...         True\n",
       "3     The book contained more profanity than I expec...        False\n",
       "4     We have now entered a second time of deep conc...         True\n",
       "...                                                 ...          ...\n",
       "1968  I purchased the book with the intention of tea...         True\n",
       "1969  There are so many design books, but the Graphi...         True\n",
       "1970  I am thilled to see this book being available ...         True\n",
       "1971  As many have stated before me the book starts ...        False\n",
       "1972  I love this book! It is a terrific blend of ha...         True\n",
       "\n",
       "[1865 rows x 2 columns]"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#renaming review (feature) and positive review (label/ target)\n",
    "\n",
    "df=df.rename(columns={\"Review\": \"Review_Text\", \"Positive Review\": \"Is_Positive\"})\n",
    "df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#want to visualize the data pertaining to the frequency of positives (true)\n",
    "#and negative reviews (false)\n",
    "\n",
    "sns.countplot(x=\"Is_Positive\", data=df)\n",
    "plt.title(\"Count of Positive vs. Negative Reviews\")\n",
    "plt.show()"
   ]
  },
  {
   "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>Review_Text</th>\n",
       "      <th>Is_Positive_0</th>\n",
       "      <th>Is_Positive_1</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>This was perhaps the best of Johannes Steinhof...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>This very fascinating book is a story written ...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>The four tales in this collection are beautifu...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>The book contained more profanity than I expec...</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>We have now entered a second time of deep conc...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1968</th>\n",
       "      <td>I purchased the book with the intention of tea...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1969</th>\n",
       "      <td>There are so many design books, but the Graphi...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1970</th>\n",
       "      <td>I am thilled to see this book being available ...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1971</th>\n",
       "      <td>As many have stated before me the book starts ...</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1972</th>\n",
       "      <td>I love this book! It is a terrific blend of ha...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>1865 rows × 3 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                            Review_Text  Is_Positive_0  \\\n",
       "0     This was perhaps the best of Johannes Steinhof...              0   \n",
       "1     This very fascinating book is a story written ...              0   \n",
       "2     The four tales in this collection are beautifu...              0   \n",
       "3     The book contained more profanity than I expec...              1   \n",
       "4     We have now entered a second time of deep conc...              0   \n",
       "...                                                 ...            ...   \n",
       "1968  I purchased the book with the intention of tea...              0   \n",
       "1969  There are so many design books, but the Graphi...              0   \n",
       "1970  I am thilled to see this book being available ...              0   \n",
       "1971  As many have stated before me the book starts ...              1   \n",
       "1972  I love this book! It is a terrific blend of ha...              0   \n",
       "\n",
       "      Is_Positive_1  \n",
       "0                 1  \n",
       "1                 1  \n",
       "2                 1  \n",
       "3                 0  \n",
       "4                 1  \n",
       "...             ...  \n",
       "1968              1  \n",
       "1969              1  \n",
       "1970              1  \n",
       "1971              0  \n",
       "1972              1  \n",
       "\n",
       "[1865 rows x 3 columns]"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "'''one-hot encoding'''\n",
    "\n",
    "df['Is_Positive']=df['Is_Positive'].astype(int)\n",
    "df=pd.get_dummies(df, columns=['Is_Positive'])\n",
    "df"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Part 3: Implement Your Project Plan\n",
    "\n",
    "<b>Task:</b> Use the rest of this notebook to carry out your project plan. You will:\n",
    "\n",
    "1. Prepare your data for your model and create features and a label.\n",
    "2. Fit your model to the training data and evaluate your model.\n",
    "3. Improve your model by performing model selection and/or feature selection techniques to find best model for your problem.\n",
    "\n",
    "\n",
    "Add code cells below and populate the notebook with commentary, code, analyses, results, and figures as you see fit."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'I would like to implement a binary classication model on this data,\\nto predict whether a given review is positive or not.'"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "'''I would like to implement a binary classication model on this data,\n",
    "to predict whether a given review is positive or not.'''"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "X=df[\"Review_Text\"]\n",
    "y=df[\"Is_Positive_1\"]\n",
    "X_train, X_test, y_train, y_test=train_test_split(X,y, test_size=0.2, random_state=1234)\n",
    "\n",
    "tfidf_vectorizer=TfidfVectorizer(max_features=5000)\n",
    "X_train_tfidf=tfidf_vectorizer.fit_transform(X_train)\n",
    "X_test_tfidf=tfidf_vectorizer.transform(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LogisticRegression(C=1.0, class_weight=None, dual=False, fit_intercept=True,\n",
       "                   intercept_scaling=1, l1_ratio=None, max_iter=100,\n",
       "                   multi_class='auto', n_jobs=None, penalty='l2',\n",
       "                   random_state=None, solver='lbfgs', tol=0.0001, verbose=0,\n",
       "                   warm_start=False)"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "LR=LogisticRegression()\n",
    "LR.fit(X_train_tfidf, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.7935656836461126\n",
      "[[137  45]\n",
      " [ 32 159]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.81      0.75      0.78       182\n",
      "           1       0.78      0.83      0.81       191\n",
      "\n",
      "    accuracy                           0.79       373\n",
      "   macro avg       0.80      0.79      0.79       373\n",
      "weighted avg       0.79      0.79      0.79       373\n",
      "\n"
     ]
    }
   ],
   "source": [
    "y_pred=LR.predict(X_test_tfidf)\n",
    "print(\"Accuracy:\", accuracy_score(y_test, y_pred))\n",
    "print(confusion_matrix(y_test, y_pred))\n",
    "print(classification_report(y_test, y_pred))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.6/dist-packages/sklearn/linear_model/_sag.py:330: ConvergenceWarning: The max_iter was reached which means the coef_ did not converge\n",
      "  \"the coef_ did not converge\", ConvergenceWarning)\n",
      "/usr/local/lib/python3.6/dist-packages/sklearn/linear_model/_sag.py:330: ConvergenceWarning: The max_iter was reached which means the coef_ did not converge\n",
      "  \"the coef_ did not converge\", ConvergenceWarning)\n",
      "/usr/local/lib/python3.6/dist-packages/sklearn/linear_model/_sag.py:330: ConvergenceWarning: The max_iter was reached which means the coef_ did not converge\n",
      "  \"the coef_ did not converge\", ConvergenceWarning)\n",
      "/usr/local/lib/python3.6/dist-packages/sklearn/linear_model/_sag.py:330: ConvergenceWarning: The max_iter was reached which means the coef_ did not converge\n",
      "  \"the coef_ did not converge\", ConvergenceWarning)\n",
      "/usr/local/lib/python3.6/dist-packages/sklearn/linear_model/_sag.py:330: ConvergenceWarning: The max_iter was reached which means the coef_ did not converge\n",
      "  \"the coef_ did not converge\", ConvergenceWarning)\n",
      "/usr/local/lib/python3.6/dist-packages/sklearn/linear_model/_sag.py:330: ConvergenceWarning: The max_iter was reached which means the coef_ did not converge\n",
      "  \"the coef_ did not converge\", ConvergenceWarning)\n",
      "/usr/local/lib/python3.6/dist-packages/sklearn/linear_model/_sag.py:330: ConvergenceWarning: The max_iter was reached which means the coef_ did not converge\n",
      "  \"the coef_ did not converge\", ConvergenceWarning)\n",
      "/usr/local/lib/python3.6/dist-packages/sklearn/linear_model/_sag.py:330: ConvergenceWarning: The max_iter was reached which means the coef_ did not converge\n",
      "  \"the coef_ did not converge\", ConvergenceWarning)\n",
      "/usr/local/lib/python3.6/dist-packages/sklearn/linear_model/_sag.py:330: ConvergenceWarning: The max_iter was reached which means the coef_ did not converge\n",
      "  \"the coef_ did not converge\", ConvergenceWarning)\n",
      "/usr/local/lib/python3.6/dist-packages/sklearn/linear_model/_sag.py:330: ConvergenceWarning: The max_iter was reached which means the coef_ did not converge\n",
      "  \"the coef_ did not converge\", ConvergenceWarning)\n",
      "/usr/local/lib/python3.6/dist-packages/sklearn/linear_model/_sag.py:330: ConvergenceWarning: The max_iter was reached which means the coef_ did not converge\n",
      "  \"the coef_ did not converge\", ConvergenceWarning)\n",
      "/usr/local/lib/python3.6/dist-packages/sklearn/linear_model/_sag.py:330: ConvergenceWarning: The max_iter was reached which means the coef_ did not converge\n",
      "  \"the coef_ did not converge\", ConvergenceWarning)\n",
      "/usr/local/lib/python3.6/dist-packages/sklearn/linear_model/_sag.py:330: ConvergenceWarning: The max_iter was reached which means the coef_ did not converge\n",
      "  \"the coef_ did not converge\", ConvergenceWarning)\n",
      "/usr/local/lib/python3.6/dist-packages/sklearn/linear_model/_sag.py:330: ConvergenceWarning: The max_iter was reached which means the coef_ did not converge\n",
      "  \"the coef_ did not converge\", ConvergenceWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best hyperparameters::\n",
      "{'C': 10, 'penalty': 'l2', 'solver': 'liblinear'}\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.6/dist-packages/sklearn/linear_model/_sag.py:330: ConvergenceWarning: The max_iter was reached which means the coef_ did not converge\n",
      "  \"the coef_ did not converge\", ConvergenceWarning)\n"
     ]
    }
   ],
   "source": [
    "'''model hyperparameter tuning'''\n",
    "param_grid={\n",
    "    'C': [0.01, 0.1, 1, 10, 100],\n",
    "    'penalty':['l1', 'l2'],\n",
    "    'solver': ['liblinear', 'saga']\n",
    "}\n",
    "\n",
    "grid=GridSearchCV(LogisticRegression(), param_grid, cv=5)\n",
    "grid.fit(X_train_tfidf, y_train)\n",
    "\n",
    "print(\"Best hyperparameters::\\n{}\".format(grid.best_params_))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LogisticRegression(C=10, class_weight=None, dual=False, fit_intercept=True,\n",
       "                   intercept_scaling=1, l1_ratio=None, max_iter=100,\n",
       "                   multi_class='auto', n_jobs=None, penalty='l2',\n",
       "                   random_state=None, solver='saga', tol=0.0001, verbose=0,\n",
       "                   warm_start=False)"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "'''training the final model'''\n",
    "optimal_params={'C':10, 'penalty':'l2', \"solver\": \"saga\"}\n",
    "LR_best=LogisticRegression(**optimal_params)\n",
    "LR_best.fit(X_train_tfidf,y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.8096514745308311\n",
      "[[143  39]\n",
      " [ 32 159]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.82      0.79      0.80       182\n",
      "           1       0.80      0.83      0.82       191\n",
      "\n",
      "    accuracy                           0.81       373\n",
      "   macro avg       0.81      0.81      0.81       373\n",
      "weighted avg       0.81      0.81      0.81       373\n",
      "\n"
     ]
    }
   ],
   "source": [
    "'''evaluating on the test set'''\n",
    "y_pred_best=LR_best.predict(X_test_tfidf)\n",
    "print(\"Accuracy:\", accuracy_score(y_test, y_pred_best))\n",
    "print(confusion_matrix(y_test, y_pred_best))\n",
    "print(classification_report(y_test, y_pred_best))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'I now want to explore different models, Random Forest being next'"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "'''I now want to explore different models, Random Forest being next'''"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Random Forest Accuracy: 0.7613941018766756\n",
      "[[137  45]\n",
      " [ 44 147]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.76      0.75      0.75       182\n",
      "           1       0.77      0.77      0.77       191\n",
      "\n",
      "    accuracy                           0.76       373\n",
      "   macro avg       0.76      0.76      0.76       373\n",
      "weighted avg       0.76      0.76      0.76       373\n",
      "\n"
     ]
    }
   ],
   "source": [
    "rf=RandomForestClassifier(n_estimators=100, random_state=42)\n",
    "rf.fit(X_train_tfidf, y_train)\n",
    "y_pred_rf=rf.predict(X_test_tfidf)\n",
    "\n",
    "print(\"Random Forest Accuracy:\", accuracy_score(y_test, y_pred_rf))\n",
    "print(confusion_matrix(y_test, y_pred_rf))\n",
    "print(classification_report(y_test, y_pred_rf))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 1.0, 'Top 20 Most Important Features')"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "'''Now, i want to investigate which features are the most influential \n",
    "in making a prediction'''\n",
    "\n",
    "importances=rf.feature_importances_\n",
    "feature_names=tfidf_vectorizer.get_feature_names()\n",
    "indices=np.argsort(importances)[-20:][::-1]\n",
    "top_features=[]\n",
    "for i in indices:\n",
    "    top_features.append(feature_names[i])\n",
    "top_importances=importances[indices]\n",
    "\n",
    "plt.figure(figsize=(12,10))\n",
    "plt.barh(top_features, top_importances, align='center')\n",
    "plt.gca().invert_yaxis()\n",
    "plt.xlabel('Relative Importance')\n",
    "plt.title('Top 20 Most Important Features')\n",
    "# plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.7908847184986595\n",
      "[[158  24]\n",
      " [ 54 137]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.75      0.87      0.80       182\n",
      "           1       0.85      0.72      0.78       191\n",
      "\n",
      "    accuracy                           0.79       373\n",
      "   macro avg       0.80      0.79      0.79       373\n",
      "weighted avg       0.80      0.79      0.79       373\n",
      "\n"
     ]
    }
   ],
   "source": [
    "'''Model Ensembling with Voting Classifier'''\n",
    "\n",
    "lr_best=LogisticRegression(**optimal_params)\n",
    "rf=RandomForestClassifier(n_estimators=100, random_state=42)\n",
    "voting_clf=VotingClassifier(estimators=[('lr', lr_best), ('rf', rf)], voting='hard')\n",
    "voting_clf.fit(X_train_tfidf, y_train)\n",
    "y_pred_voting=voting_clf.predict(X_test_tfidf)\n",
    "\n",
    "print(\"Accuracy:\", accuracy_score(y_test, y_pred_voting))\n",
    "print(confusion_matrix(y_test, y_pred_voting))\n",
    "print(classification_report(y_test, y_pred_voting))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA1gAAAGoCAYAAABbkkSYAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuNCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8QVMy6AAAACXBIWXMAAAsTAAALEwEAmpwYAABpd0lEQVR4nO3ddZitVdn48e8Nh24pUboslDpSIoJ026iAgIEoiJhgEkoYINiigKigIraAgN0BKqgoJsoPC/XV13gN5P79ca/tPGeY2DOzz9kzh+/nuvY1s2vt9fS6Vz2RmUiSJEmSZm6JYWdAkiRJkhYXBliSJEmSNCAGWJIkSZI0IAZYkiRJkjQgBliSJEmSNCAGWJIkSZI0IAZYknQ3FxGHRsQ1E7z/0Ii4eVHmaaoiYsOIyIiYN+y89CMiroqII4adD0nS4IX3wZKkwYmIv3aeLg/8E/hPe/70zLxkAL/xOWAH4A7gH8AXgGMz89czTbuln8BmmfmTQaS3KETEhsDPgaUy844x3r+F2h4bZebf2mtPBQ7LzF0Xct5OATbNzMMW5u+030rg70Dv4n5HZq46gDTn1P4gScNkC5YkDVBmrth7AL8EDuy8NuPgquO49hubA6sCrxtg2ourJYFnDzsTi8CWnX1u1WFnJiKWHHYeJGlRMsCSpEUgIpaJiHMj4lftcW5ELNPe2zUi/l9EvDgifh8Rt0TEof2km5l/BD4IbNHS2ikivhkRf25/d+rk4ciI+FlE/CUift77jfb6l9r/X2gfvyEi/hoRh/Ty194/MSIuH7Vs50XE69v/q0TEBRHx64i4LSJeOV4BOyK2i4ivRsSf2uffGBFLd97PiDgmIn7cPvOmiIj23pIR8dq2vn4G7N/H6noN8PyIWHWc/Nw3Iq6NiD9GxM0R8bjOe6tHxMcj4n/ben1lb5111sGt7f3rI+Kh7fV9gBcDh7T1eUN7/XMR8dS2X/wpIrbopLVmRPxfRKzVnh8QEd9pn/tKRDyoj2UdvWz3iogPRsTtbdsf33lv3O0wzv7w3/2lk0ZGxKbt/3dGxFsi4sqI+BuwWx+/f11bd7+NiHOmunySNJsYYEnSovESqlvfVsCWwHbASzvv3xNYA7g3cARwfkTcZ7JEI2IN4NHAtyPiHsAVwOuB1YFzgCtacLBCe33fzFwJ2An4zuj0MnOX9m+vFeT9oz7yPmC/iFip/f6SwOOAS9v776S6Lm4KbA3sBTx1nOz/B3hOW+4dgd2BZ476zAHAg4EHtd/Zu73+tPbe1sB84DHj/EbXdcDngOePfqOtn2vbcqwFPB54c0Tcv33kTcDfqO10RHt0fZPatvdoaXwgIpbNzE8CZwDvb+tzy+6XMvOfwIeAJ3Refhzw+cz8XURsDVwIPJ3apm8DPhYtOO9HRCwBfBy4gdq/dgdOiIjeuhx3O/SxP4znicDpwErAVyb5/fOA8zJzZWAT4LJ+l02SZiMDLElaNA4FTsvM32Xm7cCpwOGjPvOyzPxnZn6eCpQeNzqRjtdHxJ+oQuuvgedSrTg/zsx3Z+Ydmfle4IfAge07dwJbRMRymfnrzPz+VBciM38BfAt4ZHvp4cDfM/NrEbE2sB9wQmb+LTN/R3VdfPw4aV2fmV9reb2FCh4eNupjZ2XmnzLzl8BnqSAGat2cm5m3tla8M/tchJcDz4qINUe9fgBwS2Ze1PLzbapl8LEtiHw0cHJm/j0zbwIuHrUs78nMP7Tvng0sA0waIDeXsuA6eiIjAevRwNsy8+uZ+Z/MvJga17fDBOl9q7VG/am1LD4YWDMzT8vMf2Xmz4C3936zz+0wVR/NzC9n5p3AAyf6feDfwKYRsUZm/jUzvzbD35akoZoTsy1J0mLgXsAvOs9/0V7r+Z/e5AvjvD/a8Zn5ju4LETH6N3rp3Dsz/xYRh1CtNxdExJeB52XmD6e4HFCF/ycA72LBYGADYCng160nH1RF3q1jJRIRm1OtbPOpCSjmAdeP+thvOv//HVix/X+vUemOXu4xZeb3IuITwEnADzpvbQBs34LWnnnAu4E12//d31tgmSLi+cBTWr4SWJlqEerHZ4HlI2J74LdUEPnhTr6OiIhndT6/NBPvG9t0J6RoXR3vNWrZlgS+2N7vZztMVXf9bDDR71Pr7TTghxHxc+DUzPzEDH9fkobGFixJWjR+RRU0e9Zvr/Ws1rqpjff+dH6jl85tAJl5dWbuCaxDtWy9fYrp93wA2DUi1qVasnoB1q1U68oamblqe6ycmQ8YJ523tHxs1rqHvRiIcT472q+B9TrP159C/k+muhjeu/ParVS3vFU7jxUz8xnA7VS3x3U7n//vb7fxVi+kWtVWaxNL/LmzLBNO15uZ/6G6xT2hPT6RmX/p5Ov0UflavrVO9utW4Oej0lgpM/dr7091O/yNCsQAiIh7jrVY/f5+Zv44M59Adc18FXD5qGNBkuYUAyxJWjTeC7y0TWCwBtVV7T2jPnNqRCzdCuwHUIHMVFwJbB4RT4yIea3F6v7AJyJi7Yg4uBVc/wn8leoyOJbfAhuP9yOti+PngIuogvMP2uu/Bq4Bzo6IlSNiiYjYJCLG6262EvC/wF8j4r7AM6awrJcBx0fEuhGxGtUi1ZfWuvN+4PjOy5+g1t3hEbFUezw4Iu7XAqAPAadExPItr08atRx3UIHYvIh4OdWC1fNbYMM2Fmo8lwKHUF1JL+28/nbgmIjYPsoKEbF/bwxcn74B/CVqgpLloiYI2SIiHtzJ/0TbYfT+cAPwgIjYKiKWBU6Zye9HxGERsWbrTvin9p3x9k1JmvUMsCRp0XglNcnCjcB3qXFMr+y8/xvgf6hWqEuAY6bafS8z/0AFZs8D/kC1qhyQmb+nzvfPben/kRpjM15AcwpwcRvDM944sEuBPVgwGIAKPJYGbmrLcznVYjaW51NdDP9CBRL9TqBA+/zVVGH/W1QANBWnAf9tJWktRntR44J+RW2PV1FjqQCOA1Zpr7+bCpj/2d67Gvgk8COqq+I/WLCLXC9Q/kNEfGuszGTm16mWoXsBV3Vev45qbXsjtT5/Ahw5lQVtAeIBVNfDnwO/B97Rlgcm3w6n0NkfMvNH1Pr7FPBj4EtMoI/f3wf4ftQ95M4DHp+Z/zeVZZSk2cQbDUvSkEXErsB7MnPdST6qWSIiXgXcMzNHzyYoSbqbswVLkqRJRN0j60Gtm9521MQMH57se5Kkux9nEZQkaXIrUd0C70WNSTob+OhQcyRJmpXsIihJkiRJA2IXQUmSJEkakDnVRXCNNdbIDTfccNjZkCRJknQ3d/311/8+M9cc/fqcCrA23HBDrrvuumFnQ5IkSdLdXET8YqzX7SIoSZIkSQNigCVJkiRJA2KAJUmSJEkDYoAlSZIkSQNigCVJkiRJA2KAJUmSJEkDYoAlSZIkSQNigCVJkiRJA2KAJUmSJEkDYoAlSZIkSQNigCVJkiRJA2KAJUmSJEkDYoAlSZIkSQMyb9gZkCRJkrT42/CkK6b93VvO2n+AOVm4bMGSJEmSpAExwJIkSZKkATHAkiRJkqQBMcCSJEmSpAExwJIkSZKkATHAkiRJkqQBMcCSJEmSpAExwJIkSZKkATHAkiRJkqQBMcCSJEmSpAExwJIkSZKkATHAkiRJkqQBMcCSJEmSpAExwJIkSZKkATHAkiRJkqQBMcCSJEmSpAGZN+wMSJIkSZq9Njzpiml/95az9h9gTuYGW7AkSZIkaUBswZIkSZIWM7Y6DY8tWJIkSZI0IAZYkiRJkjQgBliSJEmSNCAGWJIkSZI0IAZYkiRJkjQgziIoSZIkzRLO/jf32YIlSZIkSQNigCVJkiRJA2KAJUmSJEkDMvQAKyKWjIhvR8Qnhp0XSZIkSZqJoQdYwLOBHww7E5IkSZI0U0MNsCJiXWB/4B3DzIckSZIkDcKwW7DOBV4I3DneByLi6Ii4LiKuu/322xdZxiRJkiRpqoYWYEXEAcDvMvP6iT6Xmedn5vzMnL/mmmsuotxJkiRJ0tQNswXrIcBBEXEL8D7g4RHxniHmR5IkSZJmZN6wfjgzXwS8CCAidgWen5mHDSs/kiRJunvZ8KQrpv3dW87af+DpaPEw7DFYkiRJkrTYGFoLVldmfg743JCzIUmSJEkzYguWJEmSJA2IAZYkSZIkDYgBliRJkiQNiAGWJEmSJA3IrJjkQpIkSerXdKdFd0p0LQq2YEmSJEnSgBhgSZIkSdKA2EVQkiRJC910u/WBXfs0t9iCJUmSJEkDYoAlSZIkSQNigCVJkiRJA2KAJUmSJEkDYoAlSZIkSQPiLIKSJEkal7P/SVNjC5YkSZIkDYgtWJIkSYsZW52k4THAkiRJmiUMjKS5zy6CkiRJkjQgBliSJEmSNCAGWJIkSZI0II7BkiRJd0uDHO803bQcNyUtfmzBkiRJkqQBMcCSJEmSpAGxi6AkSZpT7I4naTYzwJIkSeMa1Dgl7+8k6e7CAEuSpMWMwYwkDY8BliRJs4SBkSTNfQZYkqS7JafoliQtDM4iKEmSJEkDYoAlSZIkSQNigCVJkiRJA2KAJUmSJEkDYoAlSZIkSQNigCVJkiRJA2KAJUmSJEkDYoAlSZIkSQNigCVJkiRJA2KAJUmSJEkDYoAlSZIkSQNigCVJkiRJA2KAJUmSJEkDMrQAKyLWi4jPRsRNEfH9iHj2sPIiSZIkSYMwb4i/fQfwvMz8VkSsBFwfEddm5k1DzJMkSZIkTdvQWrAy89eZ+a32/1+AHwD3HlZ+JEmSJGmmZsUYrIjYENga+PoY7x0dEddFxHW33377Is+bJEmSJPVr6AFWRKwIfBA4ITP/d/T7mXl+Zs7PzPlrrrnmos+gJEmSJPVpqAFWRCxFBVeXZOaHhpkXSZIkSZqpYc4iGMAFwA8y85xh5UOSJEmSBmWYLVgPAQ4HHh4R32mP/YaYH0mSJEmakaFN056ZXwJiWL8vSZIkSYM29EkuJEmSJGlxYYAlSZIkSQNigCVJkiRJA2KAJUmSJEkDYoAlSZIkSQNigCVJkiRJA2KAJUmSJEkDYoAlSZIkSQNigCVJkiRJA2KAJUmSJEkDYoAlSZIkSQNigCVJkiRJA2KAJUmSJEkDYoAlSZIkSQNigCVJkiRJA2KAJUmSJEkD0neAFRHLL8yMSJIkSdJcN2mAFRE7RcRNwA/b8y0j4s0LPWeSJEmSNMf004L1OmBv4A8AmXkDsMvCzJQkSZIkzUV9dRHMzFtHvfSfhZAXSZIkSZrT5vXxmVsjYicgI2Ip4NnADxZutiRJkiRp7umnBesY4Fjg3sBtwFbtuSRJkiSpY9IWrMz8PXDoIsiLJEmSJM1pkwZYEXERkKNfz8wnL5QcSZIkSdIc1c8YrE90/l8WeCTwq4WTHUmSJEmau/rpIvjB7vOIeC/wpYWWI0mSJEmao/qapn2UzYC1Bp0RSZIkSZrr+hmD9RdqDFa0v78BTlzI+ZIkSZKkOaefLoIrLYqMSJIkSdJcN26AFRHbTPTFzPzW4LMjSZIkSXPXRC1YZ0/wXgIPH3BeJEmSJGlOGzfAyszdFmVGJEmSJGmu6+c+WETEFsD9qftgAZCZ71pYmZIkSZKkuaifWQRPBnalAqwrgX2p+2AZYEmSJElSRz/3wXoMsDvwm8w8CtgSWGWh5kqSJEmS5qB+Aqz/y8w7gTsiYmXgd8B6CzdbkiRJkjT39DMG67qIWBV4O3A98FfgqwszU5IkSZI0F010H6w3AZdm5jPbS2+NiE8CK2fmjYskd5IkSZI0h0zUgvUj4LURsQ5wGfDezPz2osmWJEmSJM09447ByszzMnNH4GHAH4ALI+KHEXFyRGy+yHIoSZIkSXPEpJNcZOYvMvNVmbk18ATgEcAPBvHjEbFPRNwcET+JiJMGkaYkSZIkDcukAVZEzIuIAyPiEuAq4GbgUTP94YhYEngTdV+t+wNPiIj7zzRdSZIkSRqWiSa52JNqsdoP+AbwPuDozPzbgH57O+Anmfmz9nvvAw4GbhpQ+pIkSZK0SEVmjv1GxGeAS4EPZub/DPyHIx4D7JOZT23PDwe2z8zjRn3uaOBogPXXX3/bX/ziF4POyoxseNIV0/reLWftP5B0Rqc1qHQWtzwtrGWbjXmaDcs2G/PkPmCeRqcjSdJMRMT1mTl/9OvjtmBl5sMXbpb6k5nnA+cDzJ8/f+xoUJIkSZJmgUnHYC1EtwHrdZ6v216TJEmSpDlpmAHWN4HNImKjiFgaeDzwsSHmR5IkSZJmpJ9ZBF/Vz2tTlZl3AMcBV1PTvl+Wmd+fabqSJEmSNCz9tGDtOcZr+w7ixzPzyszcPDM3yczTB5GmJEmSJA3LRNO0PwN4JrBxRNzYeWsl4MsLO2OSJEmSNNeMG2BRU7RfBZwJnNR5/S+Z+ceFmitJkiRJmoMmmqb9z8CfgSdExJLA2u3zK0bEipn5y0WUR0mSJEmaEyZqwQIgIo4DTgF+C9zZXk7gQQsvW5IkSZI090waYAEnAPfJzD8s5LxIkiRJ0pzWzyyCt1JdBSVJkiRJE+inBetnwOci4grgn70XM/OchZYrSZIkSZqD+gmwftkeS7eHJEmSJGkMkwZYmXkqQEQsn5l/X/hZkiRJkqS5adIxWBGxY0TcBPywPd8yIt680HMmSZIkSXNMP5NcnAvsDfwBIDNvAHZZiHmSJEmSpDmpnwCLzLx11Ev/WQh5kSRJkqQ5rZ9JLm6NiJ2AjIilgGcDP1i42ZIkSZKkuaefFqxjgGOBewO3AVu155IkSZKkjn5mEfw9cOgiyIskSZIkzWnjBlgR8cLMfHVEvAHI0e9n5vELNWeSJEmSNMdM1ILVG2d13aLIiCRJkiTNdeMGWJn58fb34kWXHUmSJEmau/q50fC1EbFq5/lqEXH1Qs2VJEmSJM1B/cwiuGZm/qn3JDP/B1hroeVIkiRJkuaofgKs/0TE+r0nEbEBY0x6IUmSJEl3d/3caPglwJci4vNAAA8Fjl6ouZIkSZKkOaif+2B9MiK2AXZoL53Q7o0lSZIkSeoYt4tgRNy3/d0GWB/4VXus316TJEmSJHVM1IL1XKor4NljvJfAwxdKjiRJkiRpjpoowLq2/X1KZv5sUWRGkiRJkuayiWYRfFH7e/miyIgkSZIkzXUTtWD9MSKuATaOiI+NfjMzD1p42ZIkSZKkuWeiAGs/YBvg3Yw9DkuSJEmS1DFRgHVBZh4eEW/PzM8vshxJkiRJ0hw10RisbSPiXsChEbFaRNyj+1hUGZQkSZKkuWKiFqy3Ap8GNgauB6LzXrbXJUmSJEnNuC1Ymfn6zLwfcGFmbpyZG3UeBleSJEmSNMpEXQQByMxnRMTOEXEUQESsEREbLfysSZIkSdLcMmmAFREnAycycl+spYH3LMxMSZIkSdJcNGmABTwSOAj4G0Bm/gpYaWFmSpIkSZLmon4CrH9lZlITWxARKyzcLEmSJEnS3NRPgHVZRLwNWDUingZ8Cnj7ws2WJEmSJM09E03TDkBmvjYi9gT+F7gP8PLMvHah50ySJEmS5phJA6zmRmCZ9v8NCykvkiRJkjSn9TOL4OOAbwCPBR4HfD0iHjOTH42I10TEDyPixoj4cESsOpP0JEmSJGk26GcM1kuAB2fmEZn5JGA74GUz/N1rgS0y80HAjxiZAl6SJEmS5qx+AqwlMvN3ned/6PN748rMazLzjvb0a8C6M0lPkiRJkmaDfsZgfTIirgbe254fAlw1wDw8GXj/eG9GxNHA0QDrr7/+AH9WkiRJkgarn1kEXxARjwJ2bi+dn5kfnux7EfEp4J5jvPWSzPxo+8xLgDuASyb4/fOB8wHmz5+fk/2uJEmSJA3LuAFWRGwKrJ2ZX87MDwEfaq/vHBGbZOZPJ0o4M/eY6P2IOBI4ANi93chYkiRJkua0icZSnUvd+2q0P7f3pi0i9gFeCByUmX+fSVqSJEmSNFtMFGCtnZnfHf1ie23DGf7uG4GVgGsj4jsR8dYZpidJkiRJQzfRGKxVJ3hvuZn8aGZuOpPvS5IkSdJsNFEL1nUR8bTRL0bEU4HrF16WJEmSJGlumqgF6wTgwxFxKCMB1XxgaeCRCzlfkiRJkjTnjBtgZeZvgZ0iYjdgi/byFZn5mUWSM0mSJEmaY/q5D9Zngc8ugrxIkiRJ0pw20RgsSZIkSdIUGGBJkiRJ0oAYYEmSJEnSgBhgSZIkSdKAGGBJkiRJ0oAYYEmSJEnSgBhgSZIkSdKAGGBJkiRJ0oAYYEmSJEnSgBhgSZIkSdKAGGBJkiRJ0oAYYEmSJEnSgBhgSZIkSdKAGGBJkiRJ0oDMG3YGJEmailvO2n/YWZAkaVy2YEmSJEnSgBhgSZIkSdKAGGBJkiRJ0oAYYEmSJEnSgBhgSZIkSdKAOIugJGmhc+Y/SdLdhS1YkiRJkjQgBliSJEmSNCAGWJIkSZI0II7BkiSNy7FTkiRNjS1YkiRJkjQgBliSJEmSNCB2EZSkxYzd+iRJGh5bsCRJkiRpQAywJEmSJGlADLAkSZIkaUAcgyVJs4RjpyRJmvsMsCRphgyMJElSj10EJUmSJGlADLAkSZIkaUAMsCRJkiRpQIY6Bisinge8FlgzM38/zLxIuntx3JQkSVoYhhZgRcR6wF7AL4eVB0lzj4GRJEmazYbZRfB1wAuBHGIeJEmSJGlghhJgRcTBwG2ZeUMfnz06Iq6LiOtuv/32RZA7SZIkSZqehdZFMCI+BdxzjLdeAryY6h44qcw8HzgfYP78+bZ2SZIkSZq1FlqAlZl7jPV6RDwQ2Ai4ISIA1gW+FRHbZeZvFlZ+JEmSJGlhW+STXGTmd4G1es8j4hZgvrMISpIkSZrrvA+WJEmSJA3IUO+DBZCZGw47D5IWLqdWlyRJdxe2YEmSJEnSgAy9BUvS7GXLkyRJ0tQYYEmLGYMiSZKk4bGLoCRJkiQNiC1Y0ixhy5MkSdLcZwuWJEmSJA2IAZYkSZIkDYgBliRJkiQNiAGWJEmSJA2IAZYkSZIkDYizCEoz5Ox/kiRJ6jHA0t2SQZEkSZIWBgMszSkGRpIkSZrNHIMlSZIkSQNigCVJkiRJA2KAJUmSJEkDYoAlSZIkSQNigCVJkiRJA2KAJUmSJEkDYoAlSZIkSQNigCVJkiRJA2KAJUmSJEkDMm/YGdDi75az9h92FiRJkqRFwgBL4zIwkiRJkqbGLoKSJEmSNCAGWJIkSZI0IAZYkiRJkjQgBliSJEmSNCAGWJIkSZI0IAZYkiRJkjQgBliSJEmSNCDeB2sx472rJEmSpOGxBUuSJEmSBsQAS5IkSZIGxABLkiRJkgbEAEuSJEmSBsQAS5IkSZIGxFkEZwln/5MkSZLmPluwJEmSJGlADLAkSZIkaUCGFmBFxLMi4ocR8f2IePWw8iFJkiRJgzKUMVgRsRtwMLBlZv4zItYaRj4kSZIkaZCG1YL1DOCszPwnQGb+bkj5kCRJkqSBGVaAtTnw0Ij4ekR8PiIePKR8SJIkSdLALLQughHxKeCeY7z1kva79wB2AB4MXBYRG2dmjpHO0cDRAOuvv/7Cyq4kSZIkzdhCC7Ayc4/x3ouIZwAfagHVNyLiTmAN4PYx0jkfOB9g/vz5dwnAJEmSJGm2GFYXwY8AuwFExObA0sDvh5QXSZIkSRqIocwiCFwIXBgR3wP+BRwxVvdASZIkSZpLhhJgZea/gMOG8duSJEmStLAM7UbDkiRJkrS4McCSJEmSpAExwJIkSZKkATHAkiRJkqQBMcCSJEmSpAEZ1jTti41bztp/2FmQJEmSNEvYgiVJkiRJA2KAJUmSJEkDYoAlSZIkSQNigCVJkiRJA2KAJUmSJEkDYoAlSZIkSQNigCVJkiRJA2KAJUmSJEkDYoAlSZIkSQNigCVJkiRJA2KAJUmSJEkDYoAlSZIkSQNigCVJkiRJA2KAJUmSJEkDEpk57Dz0LSJuB34x7HxMwRrA72dROoNMyzzNzXQGmZZ5mpvpDDKtxTlPi/OyDTKtxTlPi/OyDTKtxTlPi/OyDTKt2ZinRWWDzFxz9ItzKsCaayLiusycP1vSMU9zN0+L87KZp0WbjnlatOmYp7mbp8V52czTok3HPA0nrWGyi6AkSZIkDYgBliRJkiQNiAHWwnX+LEtnkGmZp7mZziDTMk9zM51BprU452lxXrZBprU452lxXrZBprU452lxXrZBpjUb8zRUjsGSJEmSpAGxBUuSJEmSBsQAS3NKRMSw8yD1uD9K0tzUO397HtfCYIA1TaMPSA/Q8UXEfSJi+RmmsQZAZqbrWrPIxsPOgKTpiYh1h52H0SJiuQGmNeuulbMsT/eD2VmuiIhZUz6PiI2GnYe5aNZswLkkIiLb4LWI2AvqAJ1uWhM9n0las0FE7Am8A7jHdPMXEUsB74iIt8DsORl28zAb8jPbRcQDZ/DdWbd+I2JF4N0R8aoZpjMvItZp/98/IpYcSAYHJCK2jIhNh52P0WZ4rly28//ACrSDNIh9PiI2mu7+NOhjbra1FkTEmsCbI+K4Yeelpx1nr46I+84wnXnt33kTfrC/tAZSRomIDSNit0FevyNi9ekEIlHmAR+NiHfDzMoVg9ynI2KniNgmM++cSZDVlnHF9v+KM0hjOeCTEXHKdPPS0lopIlZt/9+7s48utgywpqETXB0KnNZO1FPWAod12/97dQO3aaTVDfoeGxE7TSedUWku2f071fy0QsyBwNuB5YBdp3Miysx/A8cC942IU9trMzpJR8QWEbHhdL/fycP8iNhkuvkZ7zszrb2KiFVm8v1x0pzJyf5xwIUxjZbMUfv2hr1gpPfeDPK0fkSsPM3vLpGZfwUOA3aOiBOnmw/g4cATIuIc4J3AUjNI678i4oERcf+ZpgFcAPxnEHkahIjYPiKWm8G5ckVgrxbMPgY4aKZBbUTcNyL2b//PZJ/cPiLeDtOvtOukdSTwMmCZaXx3iVHXk91nmJfutW216abR+X8QQfFfgDcDu0fEk2aaWPe8NAMrUfl6ekRsMo083Csi1svMOyJiX6pi8sRoFcHTSK977t0CZrRfbktVSO0+iCArIjYGTsnMO6fx9SUy847M3AzYMSJeC9MrV4xaR1tHxHoRce9p5KlnR+CyiNhqhkHWrtS16SjggulcewEy8/+Ag4BHRsSLp5NGRCwD7NTSOB04BZhVFYkLgwHWNEXE9lSh/3mZefs0o/FtqADtdcAbmOaFBxYI+nYCjgG+N920IuJZEXEuVSB+YGZOp3C1ZGb+A/ggcCFwTWZ+dqon587J7t7A9cARMw2y2onmicAW7fmUjoPeb0bENsDbgK9GxGbTyU83WI+Il0bEsyNi/emcWDv52go4qf2dtkHkqaVzMPAC4KjM/PtUCrOjLl7PBa4A3hoRr4QZ7QMHApcD95rqd9vv9i7qWwI3As+Y7sUH+AKwF3AUcHo7bmakFUAfA9y/PZ/OdnsIcBbw9sz8+QwDhz2jamZnGvA9B3g5sHbntanma0kq6Hgn8Grgysz8zzT3o2j786OBh8L0CqCd374ZWHMA6+kQYCvgnMz8+xS/uwXwiIjoXY/2A347k/x0juFnARdHxHJTWd+jzgPHAW/rFYpnkKd/UNfcPwLHRsQRU02jc87djAoenjCdvPTSycxvAz8F7gkc34KIftNYEjicWjcHUcH1F1paB0fE0VPNV2edHw+c0g0cprH9PgicDpwXEQ+d6rl7jM/+Btih7VNT0ivTRMQ+wCeAZ0bE69t7U8pXZx2dALyJKsud0a57fYuIh0bEipl5dkvjnRGx9QyCrD8DJwOnAZdP9TwAI8uWmTcDjwAOnc51LjP/CfyOKpseBrwnM/85k2vKXGCA1afujtD+X42q1X1GRCzfaoymWrj+OvAv4GjgJZn5xxkWYvYC3gt8MjP/dzpBX0QcAxwMnAvclwoip5rGmsCH2+//DPg5sEZEbNne73u/aye7XYAPAZ8DzgAeFhGv7rw/1fX+d+BvwFPb8ynVgLXf3Ivq+ngK8CngqojYvN/8RNU0Ltf+fxbwLKrmcgPggxGx6TTztS/wKuDJwAsiYtuppNHJ37GDyFNUMPsbYHNqv2IqhdnOxWt7qkLiAOCFVAvEtPaBqEqI04GnZOYPI2K1iLjnFBarl86TqHV9MbUf7BvT60axNvBa4P3A1i0Q6RXcpnWObrWOfwSOaYWb6dTy/o7abtMOHAAi4gXAi6mA71URseM009kPOAR4XGbeEtWaufIUjrleIfbPVM+BtYCvAZu11/tevs52WbIV1i6ltv8eU1qoEb30/g+4jartn3Lw2Pn8Y4DHAstPI52dgcdRPQ6WB+4Elu2lM919sl1bDgVOaPvnCv1+t3MeeARVOXY+dR14c0yzNasFHC8FPgxcTQWVT5tKGm3fOwg4EwjgKdMJ1DrL9xzq3H0rsCFwQgve+knjP9Q55EvAScAHM/MC4JXUNWrrmEZvm6hW3sOAYzLztohYu5fnfveF9tkDqKD/t8BbImKvKRy78zrraLWIWL1dx58DrBcRy0/jWHksFRC9GdiXuqa8tZPfqVxT7k8Ft/sAJ1D71FFTvP4eAfyoBVnnAe8CLppqkNU5z30L+DJVaTMvItabQl666R0QEc/JzJ9T19++g6xunlvlwTXAZ4AtI+J+nW26eMYimeljkgfU/cLa//cBNm3/PwQ4h6opWm70Z/tMez7wbODjwG7UBXvK+eq89k7gK8Ay00mDKgitAfRaC+ZRtb2rTHG5lgd2B1Zozx8J/C+wW3u+xBTSehTw8vb/km0bfAd45RTztDlwUOf5hVRhbTrb7UwqKO49P50qGG082fJRrXFvAJ5OdZ08G9i+8/6JVKF9uSnm6T7Ad4FNqZaZt7a0HzjVfap9d7uZ5ImqrXoncCrwFODHwBET7b9j5YeRVqILe8cHsD5VOH7TVNZR++7DqZaLg4HjgS+2fD5kiuk8nQrSevvlg4Hv9/bVPtM4lrqQLk8VOF/XHvcFntTbPydJ4x7Aqu3/zYAndN57I/CkqezjbT+6L9VVce227l801fXcyc+H2/9nAR+hgollp7FPPoUKZB7WjrcvUgXR1aaYpye37b0j1Wr4dmCP9t7mvXU5znc3ANZv/28BHNc7vkbvD1PIzybtuN2HOjc8kDq/bTqN9b1q5/8LgKvo89rEgte5p7V1tE9L51Ht9WWADaeRr3ltm20LbN3W2/dp54N+9s22vT4IHNmeLw98kiokLz+NPJ0M7N/+X40KSD/X24Z9prFOW44t27Z7NPAx4JBp5GdlqiLxXu35g6nz1DnAJlPYdvemelb8rLetqOvM54Ct+shH73hbov19EXW9eyBV1vka8MUpLtv6VMvcQ4D1qGDku8DDJtv+1Pnth1Rr/DzgfVSAvT81ydA1wJbTWN+PBl7Web42VRn4ln7XUef5/YHPdp6v3vbLR/aR1jad/98M/AhYsT1/LnUu2Kq7TfrYdhu3bb4cFdReCjyzrb/7AJtPks4Snf/3buv4mPZ8I+AGplD+AnZpv7sSdb57PVXWXIWqwJs/1e03Fx6LZ9Q4YNnbc6sJ+FKq69zlVBPs1dTB9PKIWLb32clExGERcRKwPXUgvpfa4baKiBMi4qxJvt/tMnFIRDwpIh6emUcCvwDeH9XvdaI0lqIKFETE7lEz9W3Qlmk+FYzcQRVsDp9iy9PfqQvGjyNijcz8MNUicnlE7J0T1KiPUXP0D6pr4KqZ+Z+s5urrgEdFxOb95KflfU/gJRHxhqiuAT+mdcucbLuNkadb6dTAZuZLgF8Bl0bEahMtX/vc9VTh81DgAVShsecq4F9ZtbxTMQ+4Hbg9M39FdaXaGXhR9D9oerMYGRu463TzFBGPpgpR51LreBXg3VSL73Ng/HXeXddZbqBaeDajuoQslZm/BJ4A3D8i1uqzFnTVVtt9G9X6/DKqleflwC+ZYKzKBOk/NyKWzqo9/g7wbWDviFi9j/w8gSrgn9SOl39QXSn/RgW0r6QKbxOlMQ94C3BiRNyLupA9MyIubDXrv6HOT5Pu4y29g6iA7zVUQL0DsAfwuIg4ebLvj0prO2BV4M8RcTZVQDukHRu797OOmt44ufdQheoTqePnYdR+eb8p5OkhwOOpwsFXqSDtu8DjI+JtVIXEROe5Q4EvRMT6VEF2deqYP7Tl43GtVa2vbtUR8QDg19Rxsh217nduedqsfaavLrVRrc7nRcRpUeNwngL8nhrPscJE2797PQHIzLdT3cueQV2jTomIi6n1fUG7VkyUlwWOl3YduZ2qJHkF1VJ3NrXe1xgrb2Mcc0tRPT52b7Xff6e6LW1DBW9956f3MtWVeunM/B+qtv9fVEvkKn2mtzTwh8y8ITNvo4KYnwHHRcQjp5inv1OVm4cBZOY3qcBiD+Bp7bx8lzR62y5qtt75VK+Dl1MB8nltH1uHarH992R56myLVdvfd1HXzldTZYsDgNtbuhPqLOO/gRsy88uZeStV3vks1V10t4n2zcz8I3AZVS5Zg7quXElV3O1JVQadGBNM5DDO9v8/6nhduv3Ob1u+9omItcc7548qex0YEdtl5k3AP3vltsz8A/BPWvlqEkdF63qZmc8EPg98K6ol6xzgIuAjEfGgScoVtP3gAKol8xXUNe4X1LH2kPb3a9S+MNayPTQi7pXVYjY/aqzr1VTl2P4RcVxWS9ZjmOAc3l13EfF0Kig+kTr+/0pVAKxJ7aPvBf4w6Vqai4Yd4c2VBxVwfBVYuT0/h9pZlqNOgGcBa/SZ1rHUTn4CFcl/laqlOYIqhH4D2LrPtJ5LNbk+gyp07Ntev4w6gS09wXe3obrcvQP4YXutV9P0kvb8KOAmJqnxmOA39qG6CK7enj+dKswuz9gtcL0amL2oE8SzqULMS6mganOqAPKRfvNEFew2pQpE86huBS+nAp3babVofaSzI9XKeD/qgvVtqpvh6lRB9G3AB4BTJ0ijt3xPpi4SV7ZtcCvw1PbeoW2brjJJfnppLU21oCxB1ew9prcvtt/5PPCKPpbvOOAnVOH6dOD/AU+eSp46ab0YeH4nf8dQLTM7tP1y1bG2/6g0DqW63j2TKmQ/AbiWqvFaqn1mXp/5eQRVO/w+quZyHUZqCbegWmnGbMHq5rOt26fSakypwO/rbf86irpYrNlHftamCizHte+eCHwTeGt7fz3gnn0u22bUOILnMXJ+Oo7qJvQL4E+0mvpJ0lmdOi9tSZ2PdqOOswe3PP6Y1kLbR1p7M3Jeeyvwrd7yUK0j19HH+ZI6V15MtTZsMOq9RwM/ANbt4xhZgurm9iKq9vV5nfc2oCbjuQTYYpx01mGk5eq0tp42aM93a/vou6hC8vMn27fb9w6munM9uD1fsu2Lb6OC9e/0s67bd4+kei5sSBXwL6W1QAMfBT4w2Tpq/x9NnXcf2fKzPzVW8UiqVXzF3j7WZ3pPoioNDmzPNwJWav/vSgUkq0+SxkOoa9WaVO3369p2vE97f9netunz+H0WVYG0EnWuu7Qt12Oo88O4x29nn7lH57WLqVbQXovPkW0bvrXledzrXOf43aT9/3CqwuTwTn7fAaw1yTo/gGr5+CgVGD+Kuu69oe2TV9HKBn3uT8dQ3dxeSXXJnMdI74EDqAqAcfM0znr6FPDGzvOjqEqTXSZIp9uS8nKqMmLz9nxT6tx0AdWSuXYf2/8Zbf85jAqwz6TOIQ+jymOXTrT9R6V7PFUG6OXnAVQwdE1L6yYmaIUela9tWbAF7O0s2JL1TPo491KVNN+hrh+vaf9fQJ3bN6SO64eO890dqAq5janj7Fzq+rtse39PqkzQKxtOeu2lul6e3vKzOhXwfYIK8FagejlN2Do7lx9Dz8BsfcBdmoA3pgpo9+689pl2kC1B6wrXT5rctfvVi4F3tP/XYYJCLAsWGFYHLm3PX9J23GU6n30nsN4keXo1dQI+uvPaA9qJ42LqZH3/Ga7LfdvJolfoH3f52vv7UYXNg6gCda/QeQYVkHwFeHSfv30wVQB+PXXBmN9Zh48EzgOO663TMb7fu2g+hDq5v52qSduz7RNXUBeJn1JN8UcBL54kT4e29bsVdQE8mTrp/5LqIvA94AF9Lt9B7fffRRUY9mt5PJsqyPbW46eY+IJ4EBWcbUJdhE5s+8+vWh77zlNL7xHUxf7+ndc+S10Ux+weRhXget2ZnkUVYp/dtlGvsH5Ye33HKeTlwW0fuAdtjCIV4C9Ntfh8h07X0QnSOYFq8Ti5LcsxLZ1XUBUan6ePriptu5ze9qlbqe7BR1MFrU/RZxcs7lpIu7LlpXsO2Is6xnvB7lj7eO94uCdV4O8FaStRF8QXtOeTnuPa5w6jgphdO/vC2VTB47R+9yVGgoaNqKDhEuqCvCxVyLuZcQKiMdbPqu3vkm3/Pg94xFjH+jhpvZo6v27Qnp/R9puNO59ZmhofeFEfy/YAKqjvdTdfg5Huxcu1tN5NH4Xiti5eR7USHw98mgr8r2QkyLpXH+nsTgW+p1EBwulUwfpxVNe8A/s95jrHy+favv15arziOu2951NB94Mm2nZUwfK71Lnp51Tw/yCqgvOVwGZTyM9h7Tc/SJ0zH0kdN2+mrjPX0d/xu187Tt5OFWjv1/anL1Bd337a9vn39pZ3grSeTx3zX6YC0Qe09X0ddf78CWMcK1SB9e3t/3u0bb5Ne/7ktv22bvvEy+mUN/pcT59v6+aTtC5z1LFzONWqNu5x10ln37b9z6IK6qtTx/MlVOD9Q1pFMhNUSNCCjPb/S6lK0fuO+sxF1ARBE+Vn1/b7J1G9hnqt1ce27XfFePtj+373WN+ubaNeAPQQqgJwNWq/fyYTlJuoAGOt3nHX/l4DXNb5zFvasvZTtuyVU/aiyhV7URXuu1OVvpcyqkKaBc+PQVVYvZQqC7ycqnTqdeXrXZcvpCpcVpnkuF2Sun78mLper8xImfVl1PGzYb/75Fx9DD0Ds/ExasdbiZHxEe9tO9092nvPA54+hXQ3p7o5fAJ4Yef1BwEX9PH9VTv/9y7MH6EuxB+lFayolrAN+kjvwdRJ+FnthPNY2ngG6oK/DBOMR5jiOj2YKpAsweT9iE+nWtIObAfn+p33lmCkADhZC8i9qIvXClSN/lfaci3V+cwuVJ/5pUZ9dzkWDK5OBXZqz/enCtj7tOerUYXTvakL+ISFR6oA0yu0Lk2dkN9PFc42pP+W0AdSrZ0HUoWEH7Rl3aKldRFV87sjdQEfs+aZatn7JXBhe74MVWt5InVx3Jgxapkn21epAtDpVDB6EBXsjblsTD4urVsJcSwT1Fh3vtM72T+mpXsQFaht1F5fs+1nW4y1P7Fg7em2bRstwUjh8G0t3d7vTDquiCpYnMtIrefatFZmqrX3W3RqfftYtq3a9l6jLc8V1AWsW3O8DVUpsOI4aa3R+f9sqgDUu6Ae3ZZzSfpvLVyP6nr55s5rG1AFt0PoY2wR4wcNH6UKDavTqeyaJK1jqXNub9zdEi3N1zGFsTJt/7yUkXEtr6QqSjYc9bmvd/fdcdLanSrU70QVZq6gupxv3fnM24BnTpLOym1/OpIqEF/d2z+AW1oexxw3SZ2z5rX/n0wVhnv75U5UofMVVJD1qMnWNwseL5tT55951PHyhZbPs1ue96S1QI1KY8PO//dp63KDTh5/TFUCPbgtW1/nJarA+VlGxgQ/hxqf2BuDtSzjHB/t/V7rzYPbvrRnW5ZzqfPKStT58iXUNXV76to1bis0dZ3+dGdb39bSWLU95jNBYEwFm711807aOLn2/DXAe3vLNsm62ba7LagKoK2pysJrOvvIZtRxvFEf63sH6nr/QKoy4ovt9RXa81OAA8b57hqMtOruT7UqXs5IMHMSVfF3v1F5fgfjjH2kAsNvMdLzYD51PngtI2WeMXv7UMfSstT19RXttdXaNnsfdZxcTQXuT+tzf3wIdU47s62nXo+MK6jJSXqfO5cJWngYuQ6sNur11zNSuXIedSxOVpmxARXQ387INXLf9v23UhWKVzJOYMSocnP7uyZ1zT1l1PZ9AX1cw+f6Y+gZmG0ParDieu3/57UTzGXUyWx7qrb59e3g/EH3IJ8k3Rl1v2oH+VOp2vNjgM938titCTqKqiEet9tM+9xz2rLcuz1/GlWrtx9VOD+dKUxE0ec6uMsFjCpMr93+f0D7e2470XyOkcLMgW09zWOSwKqT9gbtxHAsFWD0gtKdad0AqO6dP6FzoW4nhRcyUuv9OqrA2KtpWqadeL7OSLe+1akT7pZ95OsRVGD8gM5r36RqeMftfkMViHonzQdQF9XTO++/gRqz06sAWJK6+N8wWb6oAtSvgMe350tQBZozmGaQTQW4x1HH0IcmykPbv4+kLnhPpWpOR1dCXDjF3+91u9meugB+g5ELx6Op43rSSTuo4O8eVOvbLlTt7jyqkPC9tn8tMdF+yUiw/ilqspdel9l57btHt203lQlJDmrLdBpVWN+Cau35OJ3CJ9Wl8hbG6PpCBXXXUJU0r2CkwH89dS75MW0CiD7ycxx1vB1HFbD+DJw4jf1msqDhVPqc1ICR1pPNqRbsbzFyzJ5EVSCMW7AeI703UYWq3nnpNKp1vrdfbdLW2ZjnXzrHN3W8fp0KPFeiKhEOZaSm9wNMUFlDqyGnjtMT27r6IVU43pfO+X2MY+2eVKvQ8u217ajZAruTCe1IdS96Wb/rp313I6pScsO2P32h7edHtH38jHG+tzp1fexVoC1PXZPWZ6QAehpwZu/9CfLQLewt2X77T71t314/gapMGLf7LBXk9dbzPanj923t+VLUdeL1VCDQC8IeyhitYdy1AmfXtr8cT1XybU3t32+kj+537f9r2zp9DnXe3Kq9vnPbVyedbIVq0b0nrfzR9sffA58Z9ZkXM0Ely6h87Upd53ajzlG942Xj8b7TWacvoo7/J1JB6u5Uxca1jJTNTgb+hwrYVmzr8IETpLte+/zrO69tSx2Dr2GCcgUj5+5NqQqVU6hjdLe27bdo+9iz6VyzJkqr/f9maszf3qM+8wla4N3n8bYvVX48HXh1e+0D1Dl9p5bnrfrYZntQlQZfps69y7bl2poqm3yM/loue2OuXkFVRNyTuk6+vPOZgZYtZ+tj6BmYLQ/qorME1Ur1DuqEeRVV0/E0qhC+JdV3+wlUU2pf3ROYYferUQflzdSFYt32/D4tvZupE+p3J0qrfWe/duLq1ebdq51gDmz5uZ5pzMozzfW+fTuoj6daUdagar1uohXOqELtj+h/rNRynf/fRnXZ6M30tXs74fQK3/fmrk3nS1GFhHvSaqKpgv8XGAm6lmnrsdvKslSf+Vu1nQxPb/k5gLpwj1tD3LbP46kBvUu39XQxVbO3Zedz51PBe69Ash2julNM8Bv7U92WukHWSgPYxsszQTcHpjYubVX6CLCpwuYPGGkpfBsVKD+8Pb7LOF2eqItSbx08i6phvKjtA6fQuoC2/J7POIUhuGsXtfb/x4EvjNq2W9P/+SSorsSfp2pSn0V1uehVVGxABZS9CoXVGKMWlArSf0IVCPenCjbvp7qvHEddKPsNrnqDs9elKiPOoLpg3Uan9rKPdCYLGj4x0XEyKq2VqMqo3jr6ZFvOr1P3ZIMJZiDs7Jf3AbbtvH4eVYDYoD0/E9i5/b/UBPvDAVSB/jW02mRGCm87tOXcqfP5iY6ZHanxdU+ngsevUkHEE6mW+u8wSWGIOi53ZmSWye2pCRJ6x9yS1PljzLEtExwvP6G6zx3VtuMZ7b0jqGNwvLEyy1AVbvMZOcYuo661vVaUE2gB1mTbrf2/PlU5sgQVOHycBVt6nsnErUyPpsbn9MahPKct3wG97Ue1GL+VkS5fD2ZU7fwYeVqq8/33M9K9701UQbbvmTGpMsufqW6Tb6K6lv0QOHiS73XLFQ+irkHbUoHu69s2vCdV2L6B/rr17kK1Fj6ICtK+x0g54+HUeXjCFnrq+n8qFdy/pvP6G6nzWm8s5MbjLEt3XfcqfY6ljrE/0qn0oXoATDrmrvP83m1dnDTq9SOooHqiboGjx4KdQlXw3Ejn/NLev5BJhne0z+3Qvr8lFQT1Kt5XoCqUPkwfwymoYPgb1LXlAVRr7/Es2N28n8rIo6iK8Qe2/ekN7fX1qOv5tGajnauPoWdgtjw6J7ylqcLUJ+mMpaGamG+mzwJH53sz6n7VTna9E8re7YTxNeoGkt3Pbc0YJ/ZRnxlvuuOvUEHICtQFbsJBzAth3b+dmkGtN8VxUDUf32/5/A59DNJv3z2QuuC8j5FBlL0C0dOpwO2/F8cxvt+dBvw8KhDetr32FqqQ32shiu7fKS5zr3XnM1QLwrh9vzvfWabtDxdRhaHV2rp7OQvW3k17zBxViL0VeMwi3gcGOS7tYCrwvLh97wSqNu4lVK3eZYx0DRprAPr+1HiPV7T9b5O2H51IjcG6s6X9Y/oIXqlWlA9QhYveYP+PANdOYf10u7Uu2/aFt1EX9i/RgrN23CzFSIFwonFF2zAyvmIJ6iJ4IRMMOh8nnZWpSql1qIvy1W29XUB1O/4pVcidrEvvjIKG8dKnAs4rGakcuYKqPFulj2U7iKpwurwda/drr59NFYQn7S7V2S43UMHal6lj/pFt3T2o7UvjnpdGpbU0Fch+hWqRO6jl50bqOrAG4wcxS4x6fnhbz73bVexAdRF61hS2f+94OaXtB5swMs7wnHa8nE8F23fp8UGdx3rX3/tQ17krqIrMVaiuVBdQ59/r6f888EIqoPpU2y83avvSR+ncymC8fYmR8/tKVODZ6z3wVCrQ3689X5L+J395LlXovails1xbT++juk1dQ/9jMLtBxSVU69f2VEXcLpMcE93vHkV1WXtV28+3oQrsr2z56avloqX1BEbGhr2UOl8+gLqu3MgEY10763sedZ07q/3+Lp3PvIMqwC/L5L0GJqr0Oa2PZekGRM9q+9/xVDnrB4xUAmzd1ltfvQ+o89vXGek99BLqOF6PCrxO6idfbR3t3NbtrizYUtirgJ/0Fg2MjJHfofPa/LbuX8gEQzKoc9d9O8+Pp1r6nty+P689lqLKPH0dJ4vLY+gZmA0P6gTfa+HYkQp6zqcKQet2PvcU6iK/AlNo4mQG3a9afi6kmsxvYKQm78uMdFV4Iu3+UpOktUr7u0xbtitb3pZoJ66+Jw4YwDrvnrx2oQoIl7WTVy/IuQcVoPY1y0w70fW6FLyXujCsT11Yj6cCmt1G//7oPFE1+u+jCmanUifW+e29i9q672s8Sh95XoGJ+/4HC14MN6C6Np3flnetlr8zuWut+JQDv/a9PRf1iZDBjUtbhepu+dC2X2/Z9oljO59ZdbL109bB94BL2vNlqMHsp1O1fUfQ31iiA6na5G3bspxLm1CGuiB+pI80lqRqkXekKkUua/vFJdSsT72Wqt3ask/YEkYVpg6jCla/pxUU23tvoXVdnuL2W6at68929ts/UwWGSVtBmUHQMEZax7XvXkgFaqtQhez7tuW+ZLz9iWrV6R0/D6aC19WpIPlXVAtBryvzG+h/tteTqa5E+7ft3iu8Poq6/vTGPvUThJ5KFVo3ps5xR7b83Qmc3Gd+7s/ImN39qYLoIe35zlRhfdXJ8jPqePkuI4XrZdq6fg7VGvIYxqj8owpfvZnGnkdVqqxJtXZ8lCqwL0MFXU+hz3uDUb0Lel1LP8jIZFArt/3+/VTgNF4A0rsWzKfO0c+kbpDb7Yr/OSaZ+IMFr3N7MtLC8AVGJm/ahgquPkmfgUwnze614ePAJ8d6b5LtdhkjZYNXUMHCVu35UkwwE/EY6W3d9snuOPVPM3mlVm9937cdF70JOk6nykk7dz7bT0taP5U+q/ezf7NgoPYHKhA9kCqPndw+01c3Yyqg/kjbn9eigq2TqePtw9S1aqs+0tmLOlc/kaoQuZGRMWoPpyolV5lsn2zPN6cmEXrPqNd3pI7BiVr5z2mf6c3oeSx1Tbqm85ljgOdOZb9eXB5Dz8BseFDNma+iZmL7SXttKepC/GY6TbVMs3WHGXS/omqq/49OqwJVEPhKOxn+lElaLZjhdMcLab0/jLpQ9SYZeAFVM7g51Z1mwtn4RqW1aTt5ntN57c3UyX6DKaTzUKor2CPb82WowtAbGekuOKWL4IDW1ZZU7e7q7ST9HKoAuRU1UcIFTGFGrdn2YJrj0jqf7V2gV6EqDnq1g0tSF8ifAs+ZYp4OpvrtH9J57WOMMzh7jO8/hLq4P6M9X77t15cyUonQ177ZtvOfqRaeXk36Tm0f/xBVYfM9JpkNsX3nB9TYlrOpYOZr1IV+Z6rQMK2KFqo73xep82nvXiyTLh8DChpaWs+ggqmNqdaOXheVM9r+Ne54RCqgeBML3rR5PlUIvY4qCH+UKgD2dWNMqkVhj7aM61Pn617h89tUq1hfgWP7zrpUy8d1bZs/nZGbAD+FcSqjqALr8e3/Z1K9A66hdR+iApJPM3Lj375vBD3qePkdI9e4JVueTmGCbmFU8PzV9t1e8LoSVVH2IaZROKMKwEdTlVGfZCSY3Ljla6Kxrr1zye4sOFb6WKolZMv2/Gm0yRgmSqf9f3Tbt45o2+zqTp564/emvM7b97pB1kdo43Am+c6Sbf++k1ZR23nvZCp4nLRXRfv8g6lrZm95XkkFN72KiqUZqRieKLg6gJHb1FxJHXv3oCrfXsfUW9ZnVOnTvjNeoHYR1Rr2jYn27XHSPJo69j9GjeU/pi3j5v2kRVWOXMBIt+RTqQBwE2pM7UTd37v75A7UOW359t3PA2eN+vx4M/7el5Hu3K+mKqQ3a+vpQqr3z1Jtf/8uM5yJeq4+hp6BoS78gjvbi6gg5tmd15algq53MYDggz67X40+CVH94I+jgo/dGOn+syx14Z5sQosjmcF0xwtjnVMFj5uo2sV3MXIRe2E7WH8IPHYK6W5GtQ5cxYJdCi6guuFNNpNSL1+nUd1BntB5b2nqAnI+i6j7JNVd4Oz2/65U7enF7aT8QOricTx1st+WPsd/zdYH0xiXNur73ZnwXt32995xcgDV2vCxqZ7o23d/RhUSH9EuFpO2qFIX5je3Y+3rdAorTKGA3j7fK6hcTHVv6d3rblmqcHoiVfjrTcIyXq38dlTf+h3a842plsLL2rq+gFFTl09xXS3T8tIbeN/XumaaQcOoddM7fk9u6+R5VDezZTqfWY7JbxGxblsve3ReO5k2OxhVOLqSPiozqIDjeqpV4Gqqm9HXqZkU79PW+baTpTNO2lu2fHyRdg/DCT7b63J9NdXS/f52vD2Vqrw7sn3uEdQ1ZtyWnT7yNVZF4l3OmSx47e3NyHYBFYT0JmdZkiow9u7jM2FrU/u/V5Dfsa2bqxiZpfM5VIDcz2yf+7Zt9z4WrLQ7BrhjKtuNKohfRnWD/hIL3u/oeVSQvdR01/mo4+DJ1HniLj0sxkqfagH5F6MqVajyUF9lHqpy88PU9fHsts3OY2R2vjFb0lhwfM/G1Hl2G6oy4Bm0adOpCsQz6XMs8ajfmFalz+h8MnagdizTGKNMnbcfzEhFy6HUebmfMU4bU+eNr1IVN0Fda15ElXM+yBgthaO3PVWZ/Rmq/HBJW0+bUufuN06Sh3lUWeutLNhl+lIq+NuOuvZd09Jb5BXSs+Ux9AzMhgdVc/sQqrvGO6iplHsDxtdsO1JfN/3s47f67n5FXcxf2h7zqBqzq6gC9fPp7+axA5vueIbL3Z14Yvu2HL1umU+kCsC9IGsdRmr1Jusysw1VE7MFVYB6JVPoUsBIwWzNzmsnUF1Bu9PDL800b7Y83fVFtTS8kyrg7UjVNB3bTowPpE7Uz6PPmsbZ/mAa49La9w6gCo+voQozy1ItXze14+SXVGvp+dM52VMFz/+0Y6afmz32avae0vbHZ1M1nnu1tL7ez/mks2+uzUjBcXtqvMthvX2b/rtP7tmWozd2YCmqK95rGWeQ+DTW1VJU5cCUzylMIWgY47ubU+fIi6na9/d11tmzqILxRF1Cl+z8/xSqALNre34o1QJ6LFXonrQmnQpg3ksVfh5NtRSu0vbVL1MtaQfP8HhZi+pC9zXGnzp5LUa675xJ1bh/qPP+YVRwc0x73te9zibJV9/jOKnz18ltPa1DFcx745XnUy1r/c4Y2esa2rsR6qupngdPoVrtbqS/7mXrUy1BO1GFzrePev8IYK8+83Tvti4uaM8voCrxHtvS+TYDLIBSlXGT9WZ5JDXu5wAqiH0UNbPpzn3+RreS9CBGbl+yHlVB+k3q3prjjiWizmnHM9KtbRPgqs7796Jas66mz5max/mdaVX6jJHOjAO1MdJcou2b351oH+is7y3bMfxiqjvoS1mwYnFZRioTRgdU3funbgt8vD0/m7qu9QL0+7S0J7ux9WrUdeNcRq5351CV5b3uzsvRR9C4OD+GnoFhP6gA6hwq4l6Kujhc0k6AL28nw777IA8wX8dTNRUPayfhXqHo2Yw0o0/Y/58BTnc8w2VZhbpgrdKe70cV9E5oz5eiBua+nbpI9tvvf29qcOY51PiIJ7QT95S6FDAyVfV72jpZlpELRV+D2Ae4rpZkpKvF8lSN8g2MTGKwMlVY+Ho74U67MDxbH0wyLm3UZ3dt62dTqvbui7RWaOqGnUdQQcgOVCF3WvfeaMfhhn18rjtRw2bUWIszqUL2tS2PW07hdw9u54H3A4e213ahgqzTqJtfP3SK6f2E1kLbluubVEF86PsSfQQN7XOjZ677MVVYOIUaJ9ELGI6kAu1xW5wYKXzsAhze/j+mnRMe3o7DI6mAab8+l2MFKnA5l2q16M1aunPbRzbt/vYM19m4rddtH7yWCvA/QwWL19G6C7bPPJUaJ7XKALfjpBWJVGvLV+m0UFKB8utbHm+YZB/oVgpsQxVU92tpvrOz/c+hAq2+Znhrz3sD+zcDrmj/P4yqvJvSGFcqgPkddb1ajao4fS91vVuktftUueLbVKvqxW0fXY0aI3cnk9y/rZPO/lQl5Fup1qtLO+9t0pb5cipQGqvlbFWqML8OI+PaPguc2vnMCW2/fR1VhpjW1N7MoNKnk8ZAArVRaS5PDZGYNICkurx+nrruv5eaufIzVBfYie6VFlRQ+ANGyhVbUefKV1EV3b3Xez0gxjyfjHGMrMpIJUYvyHpNS7OvsZKL+2PoGZgND6q59kyq9mx5Ksjq3Q9gq0WUh9FdXV7TTgzPo2p1l2FkpqVVmXxyjIFNdzyg5VunnVB7sxs9mir49Gavmkdd/Cc9cbWTxipUwbM3k9O2VBeDR1JB8xmMcSPLMdLqTlW9H1Wb+u7ONvgO0+wbP411tDTVwrEhFeC/hKoF+hydfvJUkHU8fV4MF+cHVRC+H3XBv47qWnI1dXFepX1mPlVg33IRbL/uRA0Ht33oRioADDrdYvpIrxv8XEhd2E9o723ZlvFh08jngVSN9eVU16UJx20NabtO2OWVsWeu690C46L23hupc3g/55S923lgt85rx1LdlPbs5on+C9bHtXNc7/sPa7+xqAvVr23buzcWcF+qwuvZnc8s0plj22+ewcjMib37cc1r57xH0/+EFru2Y+1xndeuA97Vz/7EyDV3P6pVoFvI35xqDd2HCvj27idPY/zGgYwaG8Mirrht55+LaYEvdT1+MW2MGzWj5JjXTDo9Xdo2uryzX6/Y0u12pVyqfeYuvT5YcEbU86iA+gFU2eRSqgL5wHbsPKmlvdArgvtYfzMO1Mbb9yb5zNpUJc192/PjqPPeSdS15mWT7UtU5fF11DVqGWpW2y8y0h23N7PhmGPAWLC74SFUOWV3qkL41VRFUq+74CsGuY7m8mPoGRjaglfNQfcmrdswEtX3aq/6vgHlAPO1ERXkfZgKrD7MyIX96fQxRoIB3CNlgMvTrWU8hprx5qHt+cFUQfSwKaa5DNXUfTo1Jq03YcCBwIcn23Ys2B1o9FTV61Ld8np5nHb3hGmur0PaheVmRoLH5amC4pvHWoa744NqCXgc1e3u3lQL9AbtvU9SAUmvm+kGDKiL7wT5mWyihpf3u80YKfAdQfVnP6Bd/J5MBVwvZYJ7t/T5Gwe180Bv5sb/Tks9Vx6MPXNd7xYYp1OB6SqTrWuqgPg+RgVS7f9nURUc495KY4K016a6iF5LBTk/pM/bTQx4PW1KFZ6/zchMgdu2fWpK594Z5GGsVox3AK8b9do+TNLllQULe0dQwfRXqaBxm857PwLeO97vj0pz/7Z+evcVurrz3uepHh99tV5O8BuL9BYYbf/v3Ztr77avf5TOfcSoCskPTrR+qOvPc9t5tNcF7SO0Lq7UdXMX2oQy7bUt2v6+Uee1lTr/70b1WtmHakV7SVv3a1Oz472OCrh6N6qe8vG3uDyoFsavAA9pz5dqx87lVCXbmBWtdGYg7myPL1MB8gFUC+pFVKVyX7dBab/3eapr8U+pSqNVqSDrAqYxTm5xfgw9A4tsQe/avLkRNUXxizqvHUUV+F/FIirActeuLt9uJ5crqOkue7M8HUk18044wJ4BTnc8qHXefrP3/yHtwBzdkrUO/U0r+4i2bl5PjYc4q3fipmpULmecu7KPOsE/lImnqu4Nal8kdxzvrJ95VB/o66gBo72bNC7flveiRZGf2fxgZCa8d7ULxIeoAtBTqQDnSyyiG2V38jTtiRrG2AeW77y2XFvOXov0JVSr07S6Oo76vb2oWdIeNdO0hrgvHMyCM9f1boFxJpO08o9K5/x2Pliysx16Ywlm0rVoBaoQ8gjajHNjnZsW0bo6sF0H9m7r7Yssgi7Q3LX2+yCqgL0G1Sr7wnbeO4wKiia9wWpL60gqcF21HfdnUIXFrTqf2bCPdFZsx9gD2nr5bHt8o73/YqbZcjXGby2yW2BQ48o/3Y6F66nJS3oTEPRaMx9FBVgTVia3dbwOFQitSl23f8rIPSL3onqUrNL5zjqd/5enAqXHUL1pbqLO2+dQXZ0voiqOtuh8Z9e2v265KNbXbH5QPZlexsiMy3tTXfHOZ/LW/ue2bXM4VVl0fTvPbUINNziOPsaXUwH7B9r/L6WC9V7AvWLbjoukbDlXHkPPwCJZyPHv7L091aJyYnvv8VRT50Kt7R6Vt7FuaroXNePRZ6kxFhdSNdeTDV4d2HTHA1y+A6iA6GPtJLFKOzn/kJF7Uk04oLKT1qpUzdkTqXEa36Km0T6Pmn3qBsbp7tQ5wT+ammJ0oUxVPZP9k2qJWYGqsXpsOyk+rL23cjuJbbeot+FsejD2THjPogLr3nYcWsDADCZqaN/fhwqiTmWkJf3i9ujV5g6sayhDuOfZQljnU7oFRud4W5eRsVEvoFqbejd136adtzYc9vINeF31pnH+Gn3esHeAv30CI7XfP6NaPe7TjtuLqVaofiYk6tXKX0nN+LpKe74TFUy8mkkmyOGuFa5rtrxcR12jAvgHbfzVWN+ZrQ8W7DXyZmqWwH06y3kg1ZLxCarlYtx1NSqt3alWj5Oo6dOfSI19fhUL9rgYs3Kaai37OhUYbNleO5QKik+lWpFP7WzPPVnEY6Bn66Odq17R1t3pbX1vR5Wrtpzge0GVK7uzor6vXUcm61bYLTcvS5VNPkCVkT/CyE2Mj6ImhpkTx8ci3W7DzsAiXdi73tn7TEbu7H0JNdvYIu0S1vI13k1NX00FKPegjy5BDKAWfcDLdR+qRnI+VXP1IkbuS/N0qtVhtX4OTCoYfgbwss5rB9HudUONo+s1oY83nW/vBP8ZRmY+GuhU1TNYV/u0vL2TCvzWoILIa6jC31+5G0932llPY82EdzBVe70iI11ihnayp8+JGsb43vZUIfOQto+eQ43H26ztF19jhjPPLa4Pptj9iipkXteO+/dSXWjeRY3/eA8VhDxy2Mu1kNbVWv1cTwbwO2swMrZqotrvpakC3ET3yuoW9rotIx8Bvtp5/jCqkD7pzJpUL4bHM9IdfGMq0FuLqqw8gzaT5Fx5jFpPz6DG6pxFVUDM77y3LFVeGLdyk5GAdmNGbnmxNdV75MVUpeWDqJayvir+2vn7fxjpmjyvnSvPbOt7sZgRdyFt25WpcsKJ1MQV27ftunbnM0uM+s4SVJnm6M5r96cmAvrM6H1mnP3oUOCo9v+rqUaJ3izbT6J6XY070cbd+TH0DCyyBZ34zt6708f9pBZy/g7mrjc1/SjT6K/NDGvRB7AsvRPzzsAHO68/kJoNbcf2vK9uTlTN5M1UjfI3qe4lvXFph1GtUf3OOrcHdQ+Ll7bnA5+qehrra3OqRe+hVGH6RKqGcQWqIPhC+pwW+O7wYOyZ8K6nz+nKF2E++743GdVl+cO0LstUpcpF7YLWm0FyzfbXmsKx12FfrXFUC//nqNr8Q4Gb2+vLUi1XBzEys5nrenrbYl0qWHlEO4+tRLUyn8tda7/7rjiiep58mQXHXH4I+GLnM/3cT+jBVEHxDKoVpleAfBdVJvgdrdZ/Lu4DjExa0JuY4iW0WVSpHi0T3uKFkWv43lTF87upFt6VqVno3kAFWRtMI2+PoLqG9s7fS1LX8UnvLefjv+twNyaYuImqcNqmXUe2A/5Em2CFqsB7EX20DlJB+o2MdJdek5q46XtUmelbUzl+726PoWdgkS7s2DeM+x9qAHrfs3stxPxN66am46Q1rVr0Gea/d1LuTfu5GlWQObLzmbfQalPoY5wbVUtzLSP3zHpFO7k/jJEga0pjJKiWrKFOVc2CNUQbMDJQv9f95U2MTMvdu5/PnJuIYCGuv1k/E94UlmUpanD3+XTGj1FdYi+jZsMb+vlpcXlQweyx1KQ7X2NkVrUdhp23xeHBSCvys6hKzd6NT19BjXftu/abBSu9HkC7f2I7Vt7QOVY+DVzT/p/whsTU1OE70cZVURWBf6YC6yWoAGLOdsWmxm1+hAqO1qKCrZOpHiMfoc8ZVamC+elt/exMtTK9ngqWt6K65j+Pasma0nWJmrHxeuCIYa+vufigGgo26DzvlicOp1rz301V0m3Stt932ms/Y5zJKDrHSFDlt48yMv63OznYPlSFsF04J3j0VubdRkRsRo1peiZVsD2caq7+5VAz1kTEI6hBp58AnpOZP5thektl5r8Hkbc+f29v6sJ5M9UFZzkqgPkrNXnDRdRJ9et9prcX1Rr3wsw8JyKWorqYrEdNp/7ZaebzQKqG9YtUX/v3ZObHp5PWdEXEw6jxYL+g9smXZOZF7b3TgP/NzNcuyjzNJRFxEDWw9pLMfE1EBEDOoZNaOx99kCoM/Z0qsKxGBdw3RsSqVEvvjcPL5dwWEZGZ2faPJakppy+lauMPysxft2PxHOCxMz3n3p1FxHrUGJ3PZeYHIuLpVBey91ETGzyVquC6irrH2JMy83vjpBW9Y7ldF5ejatJPjYiVqVaZ5al7Xl0fEetm5v+bKK2I2JdqRfsLFdy9KDN/HxEPpYYPPC0zLxjIyhiiiDiaan24leod8TMqsHwP8PvM/OMk3+/dK/P/ZeZD2mvbUmOY16C6rd+H2tZPz8zbp5HHg6jui3sAv8nMO6eahhYUEYdTZYpzqK63j6SCq1OplshVqCDq15Oks2Jm/jUi3kmN8/pYZt7R3tuO6hn1vwttQRYTd8cAaxlqzM0e1AnnsZl501AzNUq72P8iM28Zdl6mIiJ2pJqP307VMkLVDF5FdXv7A3W39o9OMd2Dqa4cp2fmpRExj5GC9fdnkN9HtXSelplf7V7QF5bOhX57Kqi6mSp4rErN6HQeddPUl1P3PPrcwszPXNcC8AupG6d+aNj5mY6IOJsa8H8AkFTBaD1qvOINw8zb4qIV5h5Hjfl4JdUV96VUS/GKVLepkzLzE0PL5GIgIpajWgY3oIKsj0TEMVRryAcz84p2zP4duC0zf95HmodT2+wT1LTsj8/MT0TEClQX2v+jKqf+OUk621P7wEeosY3bUufeyzPzjxGxCzXW6JppLPqsEhHLUtfgn7ZlO5QKbvfLzP8b5zvda1NQ66hX8fe69pntqfHUbwf+TXWnfNJ4gW0f+VxzOsGZSmebLZGZd0bEZ6iugetl5l8iYiOqu+A+wGsz8wt9pLkfdT58JDVz4TLUPeVujohDqMlNnpyZf1hYy7W4uNsFWFCtOsA9gTsz87Zh52dxEBHrU+OrPpGZp0fEalShcV/g+cDfqC59/5pOINMO+ldQhc53DjDf95isNm/QWg3QaVSr3I0RcRg1kPieVB/nH1BTBFvY60NE7EkVJOZMy0NEbAz8uXeRiohXUrPh7Ul1GTyWuofPtCsQVCLiftQkIWdQx9mxVE38alTrympUxc+nF0Uly+Ko23rcgqyjqIlDrhkVZH2MusfUmIX8MdLdnQqqXpWZ34+Ix1GFvhdn5scjYnlq/O3vJklnGari6jeZuV177XCqMPpzqrKudywuNvtARCxBbYsTqC7xY7YWdj5/MFW5dy3VDe1Wqovh2Zl5VvvMypn5vy2IW8kAaTh6QVX7f9XM/FP7/2PUTIwPa883oaa8vyozf9VHumsBZ2XmkyNiDaqVcWWqRWx9asiHPSr6MG/YGRiG1mXu1mHnYzHzL2rM2JMj4gOZ+aOIuJI6sd8/M7/RPjOtLlyZeWVruTorIq4BfpuZ/5lpphd1cNWsSrWg7kkNIH0fVbO6LFUIOLfXpWlxudAvTJl57bDzMJm270Zm/jsiVqFuFfG1iDg3M/+YmS+NiE2p2QN3B07LzH8NM8+Lg4h4INXK8fley3lE/JpqDdlhdAuxx9vUjerKd8/M/A3w5og4CtgnIsjMt7YC+Z7UbK0TptVJc2dqRt2dI+JnmXlZRCTwloj4T2ZeSbWGTZTWptQg/12Ar0fEizLzzMx8d0QsTc2MtwLVw2Jx2weWpW7R8rjM/MFEH2zdkR9PTaCwJ/DyzDwiIt4D3NgK9Ge04GqJzPwH1b1eQ9AJro4G9oyIP1IV3AdFxEci4lOZuUdm/jQiftHr4tc16th9ItXy/A9gvYh4UAukntqOodWpLqM2SvTpbhlgaeY6F68HUq0uN1B94n8OnBMRJ1MXtbWpLhwzlpkfi4ivzvUas8y8pnVPPDMifpWZ742I97e3v9M74S1mF/q7rVaI2wX4dUTcnwqwn0vVDB4TEW9rtedXUsfS5pn55WHld67rnJtWom6G+hdgi4jYkOqW9r6IeDg1AcBtVmTMTKeA9kzg8RHxG6og9twWDO0REUtn5rmtpv2vY6Uzajts0M6Np0bEr6jJjn4SEV/MGtv1b2ps0YT5ihpr+0pqnOvN1O1K3hkRd2bmqzLzgoi49+JaaMzMv0fEO/vcv/9NTRx0GnVrlUf1kqFafP97HU/HS80Kbf9+LjUb6u7A7hGxTmY+IiK+EREfzcyDqVubjP5uN7jagOpltC518/ndgcsi4gPUpCYXZZ/j5jXCAEvT0i5ee1M18TdSA19fT9UMr011Mfg6Nb7pu4MqxMz14KqnBYt3AK9ohY+LqYH3WsxkdYtdixrzsz41Xux7EfFcasD9ihHxe6r2+MicZWNC55JOcLUvNRPrc6nxBO+gbt5+TUT8mRrv9hawImO6RhXQ9qXGXh1CFcQvjYgLWzejVYD5EXFlrxvTWEYFakcAt7RA6AmtkuIQYOmI+HRmfqSP/O1AdXfbsz3Ob3k7Erg8IpZsLTKLZXDV0+/+nZl/i4jvUhOAPbu1fDyMGoe1f2b+0MqI2SMiHkLNiHpe1iQvP6AmjnliRFyQmdu1wGnMfaBzvB1HHR/XATdl5pvaMbsMNQv0o6jKck2RAZamJSLuQ11QH5uZ10X1Z384NfPQi4HfULVgNw8tk7PcqG6P1+JMSoudGOkn/x3qfPsD4PZWk/+jiDiWGjS+OTWJi8HVDLTg6uHA66hg9W8AEfEkalrvM4AvUIO0v22BcXpGBVcbU5MZfbTTDW2niPhC1HjT86nJI8ZsuRqV7u5UQPwo4I/A+RHxyczcp/WK2Iua7a+f7rP/jwoWtgKeTd2i5W1UofRwLDSO5QPUWOAXtgrUA6gKoR+ClRHDNOqYW5lqtVoF2DQivty6830iIo6nur1el5m/mCTNR1AVFwdRk7/0ymt/pG6R8GnqFgiahiWGnQHNLRGxRNTA4idS9yXZBiAz3w38lhp8/Hfqfgs/Ac5uLTSemMeQmR8DHpaZvzK4Wry0C+KdrabxddQF8UrgadTkClDjPt6amcdk5kcjarIAzcg+1OyA34mIIyLiUqr16mjqBpkJfMngavo6Bb1nUDOfbg48NiLW7nysdwP4/8vM/xkrnTH2979R96n8WWb+KTMfB9wREXtk5qlUJcSYY67GyOP/y8xvUrcJuSQzf0LdmuP+wNcy81qPtwVl5m+p+129kron3+FZMz+6noasc8zdP2uK9OupIRmXAs+JiL1awLQKVbnQj1Woa9PBwD+B49rrfwAeEBHLR02UommwBUt96RRGovXrPhe4A7h/ROyVNbXtVcB9WkB1W0ScB/w7Haw/ocWl26MW1FpTtqNqyz/cahPfFBEvAA5uXZiOogKC7/W+M7QMz1GdboHbU1Ou/xh4EXUz6q8CHwYOa+8dS9XSn0jdG+YuA7/Vn6ip758BHJCZv4yaEvprEfEcarD8dsCrJvh+t0b+McCPqC58B0fEJZl5Xfvoz6mZHsnM308jq98Fnh41e/CjqBaZW1t6Hm+jtFbfz496zfU0C0TdCud9EXEGNSHSEcCXqaEZL6DG0B2dNdFMP26huoD+KjMf2n7jeKq19/n9VmZobAZYmlSnALMPcFjrp/0ZavzIc4GXtgvk1sAregFVTnIzO+luYAOqBv1/o928MeumyI+juiodnZNMnayJtXNTb3rpK6lW9UuBCzLz5xHxIGrs22qZeUtEPJZqWTG4mpl7Ae9rwdWSmXly1AyNW1P3cTssJ7h1Qie4OgF4LDVe96aom6xfEREvp2rYH0K1kk3XldR4koOoFjAnkNGc08Yh3grcRk2d/xtqnPv+1P3I9gKWzknuBzfK9cBHgTsjYlfqPHkEdW+z3w4s83dTd8v7YGnq2iDmM6g7tz+Jup/My6jxDC+mCoufzQHeo0qaS3rdaFqBfyPgj5n559ZF8AzgbODabPcA6lRc2FVtBqKml34LVejYk5rNdK/M/H3UTW3fALwga2KZJXMAt3fQf68Jz6YmRLi5vfYo4F/Z5z382nHyTuCRWTfE7d0w9UAqUFsLePMgxiZGxLzMvMPjTXNNa7naB7iMauU9F/gQdW+qtwKnUJUHUz63RcQ6VOXDQVTXwNdk5ncHkvG7OQMsTSrq/iUvpO7XtAnVteZiatazk4GvAc+jLobvz8wvDSmr0lDEgjd93Jea6vhHVA3886lj42QqEPj4FGsZNYGIWAE4hyp4zAeOyJoBbRfqPkB3ZOZnLFgPVhto/wKqJ8yXqX39BOCJmfnjcb6zwDaIiPtS15U9WkDcC4JWbuNMBplft7/mpIhYlwqwjgHeTN2M/vbM/FBEPAX4XGb+dIa/sRT89z6xGgADLI2pU7u+MdVPd1Wqm8X7qZriHwOfpWpQdgPuTfVvvygzfzeMPEvDEBGbA7sC76ICqauAp1LHzR5U5cMB1L18ng08xmNksKKmGu5NL31t1PTSFwAHZ+b3h5u7xVer/T6Yqv3+M3Bm1mxmY322O+ZqPaqLE8ArqPtUvb+1Yh1B3Yfn6cA/DIqkEhFbUpOQrASsmZn3HXKWNAEDLN1Fp5vGAcCzgOdl3bfnXtTsXE8C7kvNOHNmtilce7WPQ8u4tIhFxJrANdREFjdT41LOyswndD7zSmoQ8ZtjMb6p6TC12eueRQWxN1AB7fMy84qhZuxuoo0PoZ8JjaLu//ZQapr0LwF3UjP7bUdNCX0o8IgcmfJdUhN1T8Xdqcq6x2fmLcPNkcZjgKX/iohlM/Mf7f/51FTrj8u6UfAKWTcifAPVTfABwAmZ+WG7XujuqtXgfxy4HNiMGo/4SeCazDyxfeaF1AQLL+p2JdRgta6C86kZ527LmqJbQzaq5WpP4KWZ+bCI+BJ1r54Tom6IujPV9enL43UxlFQiYim7881uziIo4L8Fxf0j4vLM/BMVRH0T+FNEPA/YNyJWpGZ0Wp+6Cd1PDa50d9X2/V9HxDXU+KqzM/O3bfrqj7aaxmuoKcKfA2BwtfCMNb20hmtUcHU0Nf3zhRHxdOqeVye1jy6RmZcMKZvSnGNwNft5AzEREatT46e+BWTr53slsC7wQeDv1P16fkTdFPfnvQGVBle6u+rs+7dQ9/vZLyIOybrHzh7UmJRNgBMz89PDyaU0PJ3g6pHUMfEV6kbbj8/MvTPzH63L4It6g+wlaXFgC9bdXJtaej/gQcA/qHEMfwDekZkP7927p91LZj411bR0t9WZAGZL6ka1l7fB+d8HXh4R/8nMy6kZ1Rb4zpCyLA1NRNwbeD3VbfaSiHg4cFu7H9ny1PjFw62Rl7Q4cQyWAGjdANcEfgds3v5+DPgOsAM1LfsJmfnxYeVRmi3aWJJ3A1dT4xEPzcybW039OdR9ly4fZh6l2aLdH+utVDD1DerGwrsBfwXOS2+2LWkxYwuWiIi9qWl2lwBup+4Ovjl134V51JTsh2TmddbE6+4uIu5Htfo+OjO/HBEnAh+JiEe0SV/mAb8dbi6l2aPdr+ffwGuBF2fm+cD5EbF0PzMPStJcY4B1N9cG4r8MODozb4qIY4F1gBupLoHLAK/OzOvAMVe6e2v3vHoTEMAHW4XDqyIigU9FxJ6Z+YH2WSsjpCYzPx4Rd1CB1TKZebnBlaTFlZNc6N9UoL1Ge34+1VXwQOAm4AOZ+Zch5U2aNdpMm4+n7rP0T2pa6TUBMvPVwFuAtXufN7iSFpSZVwFPpiZUkqTFlmOw1Lvx44rAh9oNhfcGjqXGXP1suLmThmfUNNNLUjfZXgdYlbox6lXAuzLz12N9R5Ik3f0YYImIWBc4hiowfhN4DHBsZn5qqBmTZoGI2AnYqM2AtiRwCLAxsDSwF3Wj4dfY3UmSJIFdBAVk5v+j7uPzWmqK9qMNrqT/Wg14RUQ8PjP/A3wAWA7YBvgi8GmDK0mS1OMkFwKgjbO6pj0kNZl5RUTcCZwVEUtk5qURcS010+Zb7EYrSZK6DLAkaRKZeVWbKfBdrcvgvsDTDa4kSdJojsGSpD5FxBbATsBNmfmlYedHkiTNPgZYkiRJkjQgTnIhSZIkSQNigCVJkiRJA2KAJUmSJEkDYoAlSZIkSQNigCVJkiRJA2KAJUla5CJi9Yj4Tnv8JiJu6zxfehrpHRkRt3fSeNc08/Xi6XxPkqQep2mXJA1VRJwC/DUzXzuDNI4E5mfmcTPMy18zc8UpfmdeZt4xk9+VJC0+bMGSJM0KEbF7RHw7Ir4bERdGxDLt9Vsi4tXt9W9ExKZTSPMFEfHNiLgxIk7tvP6RiLg+Ir4fEUe3184ClmstYJdExIYR8b3Od57fgkEi4nMRcW5EXAc8OyK2jYjPtzSvjoh12ueOj4ib2u+/byArSpI0q80bdgYkSQKWBd4J7J6ZP2pd/J4BnNve/3NmPjAintReO2CMNA6JiJ3b/+cBtwGbAdsBAXwsInbJzC8AT87MP0bEcsA3I+KDmXlSRByXmVsBRMSGk+R56cycHxFLAZ8HDs7M2yPiEOB04MnAScBGmfnPiFh1iutEkjQH2YIlSZoNlgR+npk/as8vBnbpvP/ezt8dx0nj/Zm5VXtcBOzVHt8GvgXclwq4AI6PiBuArwHrdV6five3v/cBtgCujYjvAC8F1m3v3QhcEhGHAXYjlKS7AVuwJElzQY7z/0QCODMz37bAixG7AnsAO2bm3yPic1QL2mh3sGBF5OjP/K3zO9/PzLECv/2pQPFA4CUR8UDHa0nS4s0WLEnSbPAfYMPO+KrDqW53PYd0/n61zzSvBp4cESsCRMS9I2ItYBXgf1pwdV9gh853/t26/AH8FlirzXi4DGN3SwS4GVgzInZsv7NURDwgIpYA1svMzwIntt+d0gQakqS5xxYsSdJs8A/gKOADETEP+Cbw1s77q0XEjcA/gSf0k2BmXhMR9wO+GhEAfwUOAz4JHBMRP6CCo691vnY+cGNEfCszD42I04BvUOO5fjjO7/wrIh4DvD4iVqGurecCPwLe014L4PWZ+ad+8i5Jmrucpl2SNKtFxC3UFOy/H3ZeJEmajF0EJUmSJGlAbMGSJEmSpAGxBUuSJEmSBsQAS5IkSZIGxABLkiRJkgbEAEuSJEmSBsQAS5IkSZIG5P8D5FKFdAdec2EAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 864x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "'''Model Interpretability for Logistic Regression'''\n",
    "lr_best.fit(X_train_tfidf, y_train)\n",
    "coef=lr_best.coef_.ravel()\n",
    "top_pos_coeff=np.argsort(coef)[-20:]\n",
    "top_neg_coeff=np.argsort(coef)[:20]\n",
    "top_coeff=np.hstack([top_neg_coeff, top_pos_coeff])\n",
    "\n",
    "feature_names=np.array(tfidf_vectorizer.get_feature_names())\n",
    "top_features=feature_names[top_coeff]\n",
    "top_coeff_values=coef[top_coeff]\n",
    "plt.figure(figsize=(12,6))\n",
    "plt.bar(top_features,top_coeff_values)\n",
    "plt.xticks(rotation=45, ha='right')\n",
    "plt.xlabel('Top Features')\n",
    "plt.ylabel('Coefficient Value')\n",
    "plt.title('Top Positive and Negative Features')\n",
    "plt.tight_layout()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.6/dist-packages/sklearn/linear_model/_sag.py:330: ConvergenceWarning: The max_iter was reached which means the coef_ did not converge\n",
      "  \"the coef_ did not converge\", ConvergenceWarning)\n",
      "/usr/local/lib/python3.6/dist-packages/sklearn/linear_model/_sag.py:330: ConvergenceWarning: The max_iter was reached which means the coef_ did not converge\n",
      "  \"the coef_ did not converge\", ConvergenceWarning)\n",
      "/usr/local/lib/python3.6/dist-packages/sklearn/linear_model/_sag.py:330: ConvergenceWarning: The max_iter was reached which means the coef_ did not converge\n",
      "  \"the coef_ did not converge\", ConvergenceWarning)\n",
      "/usr/local/lib/python3.6/dist-packages/sklearn/linear_model/_sag.py:330: ConvergenceWarning: The max_iter was reached which means the coef_ did not converge\n",
      "  \"the coef_ did not converge\", ConvergenceWarning)\n",
      "/usr/local/lib/python3.6/dist-packages/sklearn/linear_model/_sag.py:330: ConvergenceWarning: The max_iter was reached which means the coef_ did not converge\n",
      "  \"the coef_ did not converge\", ConvergenceWarning)\n",
      "/usr/local/lib/python3.6/dist-packages/sklearn/linear_model/_sag.py:330: ConvergenceWarning: The max_iter was reached which means the coef_ did not converge\n",
      "  \"the coef_ did not converge\", ConvergenceWarning)\n",
      "/usr/local/lib/python3.6/dist-packages/sklearn/linear_model/_sag.py:330: ConvergenceWarning: The max_iter was reached which means the coef_ did not converge\n",
      "  \"the coef_ did not converge\", ConvergenceWarning)\n",
      "/usr/local/lib/python3.6/dist-packages/sklearn/linear_model/_sag.py:330: ConvergenceWarning: The max_iter was reached which means the coef_ did not converge\n",
      "  \"the coef_ did not converge\", ConvergenceWarning)\n",
      "/usr/local/lib/python3.6/dist-packages/sklearn/linear_model/_sag.py:330: ConvergenceWarning: The max_iter was reached which means the coef_ did not converge\n",
      "  \"the coef_ did not converge\", ConvergenceWarning)\n",
      "/usr/local/lib/python3.6/dist-packages/sklearn/linear_model/_sag.py:330: ConvergenceWarning: The max_iter was reached which means the coef_ did not converge\n",
      "  \"the coef_ did not converge\", ConvergenceWarning)\n",
      "/usr/local/lib/python3.6/dist-packages/sklearn/linear_model/_sag.py:330: ConvergenceWarning: The max_iter was reached which means the coef_ did not converge\n",
      "  \"the coef_ did not converge\", ConvergenceWarning)\n",
      "/usr/local/lib/python3.6/dist-packages/sklearn/linear_model/_sag.py:330: ConvergenceWarning: The max_iter was reached which means the coef_ did not converge\n",
      "  \"the coef_ did not converge\", ConvergenceWarning)\n",
      "/usr/local/lib/python3.6/dist-packages/sklearn/linear_model/_sag.py:330: ConvergenceWarning: The max_iter was reached which means the coef_ did not converge\n",
      "  \"the coef_ did not converge\", ConvergenceWarning)\n",
      "/usr/local/lib/python3.6/dist-packages/sklearn/linear_model/_sag.py:330: ConvergenceWarning: The max_iter was reached which means the coef_ did not converge\n",
      "  \"the coef_ did not converge\", ConvergenceWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best hyperparameters:\n",
      "{'C': 10, 'penalty': 'l2', 'solver': 'liblinear'}\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.6/dist-packages/sklearn/linear_model/_sag.py:330: ConvergenceWarning: The max_iter was reached which means the coef_ did not converge\n",
      "  \"the coef_ did not converge\", ConvergenceWarning)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "LogisticRegression(C=10, class_weight=None, dual=False, fit_intercept=True,\n",
       "                   intercept_scaling=1, l1_ratio=None, max_iter=1000,\n",
       "                   multi_class='auto', n_jobs=None, penalty='l2',\n",
       "                   random_state=None, solver='liblinear', tol=0.0001, verbose=0,\n",
       "                   warm_start=False)"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "param_grid={\n",
    "    'C': [0.01, 0.1, 1,10,100],\n",
    "    'penalty': ['l1', 'l2'],\n",
    "    'solver': ['liblinear', 'saga']\n",
    "}\n",
    "grid=GridSearchCV(LogisticRegression(), param_grid, cv=5)\n",
    "grid.fit(X_train_tfidf, y_train)\n",
    "\n",
    "print(\"Best hyperparameters:\\n{}\".format(grid.best_params_))\n",
    "\n",
    "optimal_params={'C':10, 'penalty': 'l2', 'solver': 'liblinear'}\n",
    "LR_best = LogisticRegression(**optimal_params, max_iter=1000)\n",
    "LR_best.fit(X_train_tfidf, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Logistic Regression Accuracy: 0.8096514745308311\n",
      "[[143  39]\n",
      " [ 32 159]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.82      0.79      0.80       182\n",
      "           1       0.80      0.83      0.82       191\n",
      "\n",
      "    accuracy                           0.81       373\n",
      "   macro avg       0.81      0.81      0.81       373\n",
      "weighted avg       0.81      0.81      0.81       373\n",
      "\n"
     ]
    }
   ],
   "source": [
    "'''Logistic Regression Model Evaluation'''\n",
    "y_pred_lr=LR_best.predict(X_test_tfidf)\n",
    "print(\"Logistic Regression Accuracy:\", accuracy_score(y_test, y_pred_lr))\n",
    "print(confusion_matrix(y_test, y_pred_lr))\n",
    "print(classification_report(y_test, y_pred_lr))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "coef=LR_best.coef_.ravel()\n",
    "top_post_coeff=np.argsort(coef)[-20:]\n",
    "top_neg_coeff=np.argsort(coef)[:20]\n",
    "top_coeff=np.hstack([top_neg_coeff, top_pos_coeff])\n",
    "\n",
    "feature_names=np.array(tfidf_vectorizer.get_feature_names())\n",
    "top_features=feature_names[top_coeff]\n",
    "top_coeff_values=coef[top_coeff]\n",
    "\n",
    "plt.figure(figsize=(12,6))\n",
    "plt.bar(top_features, top_coeff_values)\n",
    "plt.xticks(rotation=45, ha='right')\n",
    "plt.xlabel('Top Features')\n",
    "plt.ylabel(\"Coefficient Value\")\n",
    "plt.title('Top Positive and Negative Features')\n",
    "plt.tight_layout()\n",
    "plt.show()\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.6/dist-packages/sklearn/linear_model/_sag.py:330: ConvergenceWarning: The max_iter was reached which means the coef_ did not converge\n",
      "  \"the coef_ did not converge\", ConvergenceWarning)\n",
      "/usr/local/lib/python3.6/dist-packages/sklearn/linear_model/_sag.py:330: ConvergenceWarning: The max_iter was reached which means the coef_ did not converge\n",
      "  \"the coef_ did not converge\", ConvergenceWarning)\n",
      "/usr/local/lib/python3.6/dist-packages/sklearn/linear_model/_sag.py:330: ConvergenceWarning: The max_iter was reached which means the coef_ did not converge\n",
      "  \"the coef_ did not converge\", ConvergenceWarning)\n",
      "/usr/local/lib/python3.6/dist-packages/sklearn/linear_model/_sag.py:330: ConvergenceWarning: The max_iter was reached which means the coef_ did not converge\n",
      "  \"the coef_ did not converge\", ConvergenceWarning)\n",
      "/usr/local/lib/python3.6/dist-packages/sklearn/linear_model/_sag.py:330: ConvergenceWarning: The max_iter was reached which means the coef_ did not converge\n",
      "  \"the coef_ did not converge\", ConvergenceWarning)\n",
      "/usr/local/lib/python3.6/dist-packages/sklearn/linear_model/_sag.py:330: ConvergenceWarning: The max_iter was reached which means the coef_ did not converge\n",
      "  \"the coef_ did not converge\", ConvergenceWarning)\n",
      "/usr/local/lib/python3.6/dist-packages/sklearn/linear_model/_sag.py:330: ConvergenceWarning: The max_iter was reached which means the coef_ did not converge\n",
      "  \"the coef_ did not converge\", ConvergenceWarning)\n",
      "/usr/local/lib/python3.6/dist-packages/sklearn/linear_model/_sag.py:330: ConvergenceWarning: The max_iter was reached which means the coef_ did not converge\n",
      "  \"the coef_ did not converge\", ConvergenceWarning)\n",
      "/usr/local/lib/python3.6/dist-packages/sklearn/linear_model/_sag.py:330: ConvergenceWarning: The max_iter was reached which means the coef_ did not converge\n",
      "  \"the coef_ did not converge\", ConvergenceWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best hyperparameters: {'C': 10, 'penalty': 'l2', 'solver': 'liblinear'}\n"
     ]
    }
   ],
   "source": [
    "param_grid={\n",
    "    'C': [0.01, 0.1, 1,10,100],\n",
    "    'penalty': ['l1', 'l2'],\n",
    "    'solver': ['liblinear', 'saga']\n",
    "}\n",
    "\n",
    "grid=GridSearchCV(LogisticRegression(max_iter=1000), param_grid, cv=5)\n",
    "grid.fit(X_train_tfidf, y_train)\n",
    "best_params=grid.best_params_\n",
    "print(\"Best hyperparameters:\", best_params)\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Tuned Logistic Regression Accuracy: 0.8096514745308311\n",
      "[[143  39]\n",
      " [ 32 159]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.82      0.79      0.80       182\n",
      "           1       0.80      0.83      0.82       191\n",
      "\n",
      "    accuracy                           0.81       373\n",
      "   macro avg       0.81      0.81      0.81       373\n",
      "weighted avg       0.81      0.81      0.81       373\n",
      "\n"
     ]
    }
   ],
   "source": [
    "best_params={'C': 10, 'penalty': 'l2', 'solver': 'liblinear'}\n",
    "best_LR_model=LogisticRegression(**best_params, max_iter=1000)\n",
    "\n",
    "best_LR_model.fit(X_train_tfidf, y_train)\n",
    "\n",
    "y_pred_best=best_LR_model.predict(X_test_tfidf)\n",
    "\n",
    "print(\"Tuned Logistic Regression Accuracy:\", accuracy_score(y_test, y_pred_best))\n",
    "print(confusion_matrix(y_test, y_pred_best))\n",
    "print(classification_report(y_test, y_pred_best))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Support Vector Machine Accuracy: 0.7962\n",
      "[[138  44]\n",
      " [ 32 159]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.81      0.76      0.78       182\n",
      "           1       0.78      0.83      0.81       191\n",
      "\n",
      "    accuracy                           0.80       373\n",
      "   macro avg       0.80      0.80      0.80       373\n",
      "weighted avg       0.80      0.80      0.80       373\n",
      "\n",
      "Random Forest Accuracy: 0.7614\n",
      "[[137  45]\n",
      " [ 44 147]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.76      0.75      0.75       182\n",
      "           1       0.77      0.77      0.77       191\n",
      "\n",
      "    accuracy                           0.76       373\n",
      "   macro avg       0.76      0.76      0.76       373\n",
      "weighted avg       0.76      0.76      0.76       373\n",
      "\n",
      "Gradient Boosting Accuracy: 0.7641\n",
      "[[130  52]\n",
      " [ 36 155]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.78      0.71      0.75       182\n",
      "           1       0.75      0.81      0.78       191\n",
      "\n",
      "    accuracy                           0.76       373\n",
      "   macro avg       0.77      0.76      0.76       373\n",
      "weighted avg       0.77      0.76      0.76       373\n",
      "\n",
      "Naive Bayes Accuracy: 0.8204\n",
      "[[138  44]\n",
      " [ 23 168]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.86      0.76      0.80       182\n",
      "           1       0.79      0.88      0.83       191\n",
      "\n",
      "    accuracy                           0.82       373\n",
      "   macro avg       0.82      0.82      0.82       373\n",
      "weighted avg       0.82      0.82      0.82       373\n",
      "\n",
      "K-Nearest Neighbors Accuracy: 0.6300\n",
      "[[150  32]\n",
      " [106  85]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.59      0.82      0.68       182\n",
      "           1       0.73      0.45      0.55       191\n",
      "\n",
      "    accuracy                           0.63       373\n",
      "   macro avg       0.66      0.63      0.62       373\n",
      "weighted avg       0.66      0.63      0.62       373\n",
      "\n"
     ]
    }
   ],
   "source": [
    "'''Model Selection'''\n",
    "\n",
    "models={\n",
    "    'Support Vector Machine': SVC(),\n",
    "    \"Random Forest\": RandomForestClassifier(n_estimators=100, random_state=42),\n",
    "    \"Gradient Boosting\": GradientBoostingClassifier(n_estimators=100, random_state=42),\n",
    "    'Naive Bayes': MultinomialNB(),\n",
    "    \"K-Nearest Neighbors\": KNeighborsClassifier()\n",
    "}\n",
    "\n",
    "\n",
    "for model_name, model in models.items():\n",
    "    model.fit(X_train_tfidf, y_train)\n",
    "    y_pred=model.predict(X_test_tfidf)\n",
    "    accuracy=accuracy_score(y_test, y_pred)\n",
    "    print(f'{model_name} Accuracy: {accuracy:.4f}')\n",
    "    print(confusion_matrix(y_test, y_pred))\n",
    "    print(classification_report(y_test, y_pred))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Features: [k], Tuned Logistic Regression Accuracy:0.6515\n",
      "[[108  74]\n",
      " [ 56 135]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.66      0.59      0.62       182\n",
      "           1       0.65      0.71      0.68       191\n",
      "\n",
      "    accuracy                           0.65       373\n",
      "   macro avg       0.65      0.65      0.65       373\n",
      "weighted avg       0.65      0.65      0.65       373\n",
      "\n",
      "Features: [k], Tuned Logistic Regression Accuracy:0.7346\n",
      "[[133  49]\n",
      " [ 50 141]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.73      0.73      0.73       182\n",
      "           1       0.74      0.74      0.74       191\n",
      "\n",
      "    accuracy                           0.73       373\n",
      "   macro avg       0.73      0.73      0.73       373\n",
      "weighted avg       0.73      0.73      0.73       373\n",
      "\n",
      "Features: [k], Tuned Logistic Regression Accuracy:0.7453\n",
      "[[133  49]\n",
      " [ 46 145]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.74      0.73      0.74       182\n",
      "           1       0.75      0.76      0.75       191\n",
      "\n",
      "    accuracy                           0.75       373\n",
      "   macro avg       0.75      0.74      0.75       373\n",
      "weighted avg       0.75      0.75      0.75       373\n",
      "\n",
      "Features: [k], Tuned Logistic Regression Accuracy:0.7453\n",
      "[[137  45]\n",
      " [ 50 141]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.73      0.75      0.74       182\n",
      "           1       0.76      0.74      0.75       191\n",
      "\n",
      "    accuracy                           0.75       373\n",
      "   macro avg       0.75      0.75      0.75       373\n",
      "weighted avg       0.75      0.75      0.75       373\n",
      "\n",
      "Features: [k], Tuned Logistic Regression Accuracy:0.7507\n",
      "[[136  46]\n",
      " [ 47 144]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.74      0.75      0.75       182\n",
      "           1       0.76      0.75      0.76       191\n",
      "\n",
      "    accuracy                           0.75       373\n",
      "   macro avg       0.75      0.75      0.75       373\n",
      "weighted avg       0.75      0.75      0.75       373\n",
      "\n"
     ]
    }
   ],
   "source": [
    "'''Feature Selection'''\n",
    "\n",
    "k_values=[10,20,30,50,100]\n",
    "for k in k_values:\n",
    "    selector=SelectKBest(score_func=chi2, k=k)\n",
    "    X_train_selected =selector.fit_transform(X_train_tfidf, y_train)\n",
    "    X_test_selected=selector.transform(X_test_tfidf)\n",
    "    \n",
    "    best_LR_model.fit(X_train_selected, y_train)\n",
    "    y_pred_selected=best_LR_model.predict(X_test_selected)\n",
    "    \n",
    "    accuracy=accuracy_score(y_test, y_pred_selected)\n",
    "    print(f'Features: [k], Tuned Logistic Regression Accuracy:{accuracy:.4f}')\n",
    "    print(confusion_matrix(y_test, y_pred_selected))\n",
    "    print(classification_report(y_test, y_pred_selected))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.6.9"
  },
  "toc": {
   "base_numbering": 1,
   "nav_menu": {},
   "number_sections": false,
   "sideBar": true,
   "skip_h1_title": false,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": false,
   "toc_position": {},
   "toc_section_display": true,
   "toc_window_display": false
  },
  "varInspector": {
   "cols": {
    "lenName": 16,
    "lenType": 16,
    "lenVar": 40
   },
   "kernels_config": {
    "python": {
     "delete_cmd_postfix": "",
     "delete_cmd_prefix": "del ",
     "library": "var_list.py",
     "varRefreshCmd": "print(var_dic_list())"
    },
    "r": {
     "delete_cmd_postfix": ") ",
     "delete_cmd_prefix": "rm(",
     "library": "var_list.r",
     "varRefreshCmd": "cat(var_dic_list()) "
    }
   },
   "types_to_exclude": [
    "module",
    "function",
    "builtin_function_or_method",
    "instance",
    "_Feature"
   ],
   "window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}