Multinomial-Naive-Bayes-from-Scratch.ipynb

{
 "cells": [
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "1bd3a4e9-5e71-4d5f-8582-c2141e186370",
   "metadata": {},
   "source": [
    "# Multinomial Naïve Bayes\n"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "08fcb7b3-fc94-4ce9-b058-0b26a3b93f7a",
   "metadata": {},
   "source": [
    "This file will have you working and experimenting with the Multinomial Naïve Bayes classifier. Initially, you will transform the given data in csv file to count matrix, then calculate the priors. Use those priors to compute likelyhoods according to Multinomial Naive Bayes and then classify the test data. Please note that use of `sklearn` implementations is only for the final question of the assignment, for other doubts regarding libraries you can reach out to the TAs.\n",
    "\n",
    "The dataset is about `Spam SMS`. There is 1 attribute that is the `message`, and the class label which could be `spam` or `ham`. The data is present in `spam.csv`. It contains about 5-6000 samples.\n",
    "For your convinience the data is already pre-processed and loaded, but I suggest you to just take a look at the code for your own knowledge, and parts vectorization is left up to you which could be easily done with the help of the given example code."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "id": "aa8a96d4-58f3-4360-b6ae-a02405ffdddb",
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "630170e6-57f8-4dae-b644-ab275a3f53a2",
   "metadata": {
    "tags": []
   },
   "source": [
    "## Reading text-based data using pandas"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "id": "2430d4f2-e3ed-4e09-9e50-a446889d58cf",
   "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>label</th>\n",
       "      <th>message</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>ham</td>\n",
       "      <td>Go until jurong point, crazy.. Available only ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>ham</td>\n",
       "      <td>Ok lar... Joking wif u oni...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>spam</td>\n",
       "      <td>Free entry in 2 a wkly comp to win FA Cup fina...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>ham</td>\n",
       "      <td>U dun say so early hor... U c already then say...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>ham</td>\n",
       "      <td>Nah I don't think he goes to usf, he lives aro...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  label                                            message\n",
       "0   ham  Go until jurong point, crazy.. Available only ...\n",
       "1   ham                      Ok lar... Joking wif u oni...\n",
       "2  spam  Free entry in 2 a wkly comp to win FA Cup fina...\n",
       "3   ham  U dun say so early hor... U c already then say...\n",
       "4   ham  Nah I don't think he goes to usf, he lives aro..."
      ]
     },
     "execution_count": 84,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# read file into pandas using a relative path\n",
    "\n",
    "df = pd.read_csv(\"./spam.csv\", encoding='latin-1')\n",
    "df.dropna(how=\"any\", inplace=True, axis=1)\n",
    "df.columns = ['label', 'message']\n",
    "\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c5fbeec1-2b88-4af3-b8c6-906936938951",
   "metadata": {},
   "source": [
    "## Pre-processing\n",
    "\n",
    "- Our main issue with our data is that it is all in text format (strings). The classification algorithms that we usally use need some sort of numerical feature vector in order to perform the classification task. There are actually many methods to convert a corpus to a vector format. The simplest is the bag-of-words approach, where each unique word in a text will be represented by one number.\n",
    "\n",
    "- As a first step, let's write a function that will split a message into its individual words and return a list. We'll also remove very common words, ('the', 'a', etc..). To do this we will take advantage of the NLTK library. It's pretty much the standard library in Python for processing text and has a lot of useful features. We'll only use some of the basic ones here."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "id": "d736a35c-e849-4029-bee7-42a31f7d0b11",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[nltk_data] Downloading package stopwords to\n",
      "[nltk_data]     C:\\Users\\ACER\\AppData\\Roaming\\nltk_data...\n",
      "[nltk_data]   Package stopwords is already up-to-date!\n"
     ]
    }
   ],
   "source": [
    "import string\n",
    "import nltk\n",
    "nltk.download('stopwords')\n",
    "\n",
    "from nltk.corpus import stopwords\n",
    "\n",
    "def text_process(mess):\n",
    "    \"\"\"\n",
    "    Takes in a string of text, then performs the following:\n",
    "    1. Remove all punctuation\n",
    "    2. Remove all stopwords\n",
    "    3. Returns a list of the cleaned text\n",
    "    \"\"\"\n",
    "    STOPWORDS = stopwords.words('english') + ['u', 'ü', 'ur', '4', '2', 'im', 'dont', 'doin', 'ure']\n",
    "    # Check characters to see if they are in punctuation\n",
    "    nopunc = [char for char in mess if char not in string.punctuation]\n",
    "\n",
    "    # Join the characters again to form the string.\n",
    "    nopunc = ''.join(nopunc)\n",
    "    \n",
    "    # Now just remove any stopwords\n",
    "    return ' '.join([word for word in nopunc.split() if word.lower() not in STOPWORDS])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "id": "ed807b40-7762-4968-b546-2f72817beed3",
   "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>label</th>\n",
       "      <th>message</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>ham</td>\n",
       "      <td>Go jurong point crazy Available bugis n great ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>ham</td>\n",
       "      <td>Ok lar Joking wif oni</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>spam</td>\n",
       "      <td>Free entry wkly comp win FA Cup final tkts 21s...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>ham</td>\n",
       "      <td>dun say early hor c already say</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>ham</td>\n",
       "      <td>Nah think goes usf lives around though</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  label                                            message\n",
       "0   ham  Go jurong point crazy Available bugis n great ...\n",
       "1   ham                              Ok lar Joking wif oni\n",
       "2  spam  Free entry wkly comp win FA Cup final tkts 21s...\n",
       "3   ham                    dun say early hor c already say\n",
       "4   ham             Nah think goes usf lives around though"
      ]
     },
     "execution_count": 86,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['message'] = df.message.apply(text_process)\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "id": "ad88d89f-1452-4e7a-9f9c-cb1094f890d1",
   "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>label</th>\n",
       "      <th>message</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>Go jurong point crazy Available bugis n great ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>Ok lar Joking wif oni</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>Free entry wkly comp win FA Cup final tkts 21s...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "      <td>dun say early hor c already say</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>Nah think goes usf lives around though</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   label                                            message\n",
       "0      0  Go jurong point crazy Available bugis n great ...\n",
       "1      0                              Ok lar Joking wif oni\n",
       "2      1  Free entry wkly comp win FA Cup final tkts 21s...\n",
       "3      0                    dun say early hor c already say\n",
       "4      0             Nah think goes usf lives around though"
      ]
     },
     "execution_count": 87,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['label'] = df.label.map({'ham':0, 'spam':1})\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f22eddba-546a-4e97-b013-2906e90946b7",
   "metadata": {},
   "source": [
    "## Splitting the data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "id": "f859993a-2b29-4baf-a168-117151ed240b",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "X: (5572,)\n",
      "y: (5572,)\n",
      "\n",
      "X_train: (4179,)\n",
      "y_train: (4179,)\n",
      "\n",
      "X_test: (1393,)\n",
      "y_test: (1393,)\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# split X and y into training and testing sets \n",
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "X = df.message\n",
    "y = df.label\n",
    "\n",
    "print(f'X: {X.shape}')\n",
    "print(f'y: {y.shape}')\n",
    "print()\n",
    "\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.25, random_state=1)\n",
    "\n",
    "print(f'X_train: {X_train.shape}')\n",
    "print(f'y_train: {y_train.shape}')\n",
    "print()\n",
    "\n",
    "print(f'X_test: {X_test.shape}')\n",
    "print(f'y_test: {y_test.shape}')\n",
    "print()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ba6deffc-0424-4cb6-85d2-eb74f97e8cdd",
   "metadata": {
    "tags": []
   },
   "source": [
    "## Helper code / Example code for Representing text as Numerical data using Sci-kit learn\n",
    "\n",
    "📌 From the scikit-learn documentation:\n",
    "- Text Analysis is a major application field for machine learning algorithms. However the raw data, a sequence of symbols cannot be fed directly to the algorithms themselves as most of them expect numerical feature vectors with a fixed size rather than the raw text documents with variable length.\n",
    "- We will use CountVectorizer to \"convert text into a matrix of token counts\":"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "id": "1309604c-b9c1-4910-ab60-432f7e12824b",
   "metadata": {},
   "outputs": [],
   "source": [
    "# example text for model training (SMS messages)\n",
    "simple_train = ['call you tonight', 'Call me a cab', 'Please call me... PLEASE!']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 90,
   "id": "0fde24ff-b303-4526-899d-fb902eabc1d8",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['cab', 'call', 'me', 'please', 'tonight', 'you'], dtype=object)"
      ]
     },
     "execution_count": 90,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# import and instantiate CountVectorizer (with the default parameters)\n",
    "from sklearn.feature_extraction.text import CountVectorizer\n",
    "\n",
    "vect = CountVectorizer()\n",
    "simple_train = vect.fit_transform(simple_train)\n",
    "\n",
    "vect.get_feature_names_out()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 91,
   "id": "576f6e05-6e94-45a1-96cc-3e694cc971d6",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['cab', 'call', 'me', 'please', 'tonight', 'you'], dtype=object)"
      ]
     },
     "execution_count": 91,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "vect.get_feature_names_out()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "id": "0aef831c-50f0-4723-a04a-27af774f0542",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[0, 1, 0, 0, 1, 1],\n",
       "       [1, 1, 1, 0, 0, 0],\n",
       "       [0, 1, 1, 2, 0, 0]], dtype=int64)"
      ]
     },
     "execution_count": 92,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# convert sparse matrix to a dense matrix\n",
    "simple_train.toarray()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "aa523614-ff0d-4f22-b8f2-4505e33a2460",
   "metadata": {},
   "source": [
    "In this scheme, features and samples are defined as follows:\n",
    "\n",
    "- Each individual token occurrence frequency (normalized or not) is treated as a feature.\n",
    "- The vector of all the token frequencies for a given document is considered a multivariate sample.\n",
    "\n",
    "A corpus of documents can thus be represented by a matrix with one row per document and one column per token (e.g. word) occurring in the corpus."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 93,
   "id": "5d0a1561-6e5c-495e-b056-fc9fe376121d",
   "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>cab</th>\n",
       "      <th>call</th>\n",
       "      <th>me</th>\n",
       "      <th>please</th>\n",
       "      <th>tonight</th>\n",
       "      <th>you</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   cab  call  me  please  tonight  you\n",
       "0    0     1   0       0        1    1\n",
       "1    1     1   1       0        0    0\n",
       "2    0     1   1       2        0    0"
      ]
     },
     "execution_count": 93,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# examine the vocabulary and document-term matrix together\n",
    "pd.DataFrame(simple_train.toarray(), columns=vect.get_feature_names_out())"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "02042133-9bad-4aa3-b0be-a53158e7102e",
   "metadata": {},
   "source": [
    "### Transform Testing data into a document-term matrix (using existing / training vocabulary)\n",
    "\n",
    "- You are supposed to use the training vocabolary to make the count matrix for test data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 94,
   "id": "1a50c164-91ac-4861-993d-e14a2be7c6aa",
   "metadata": {},
   "outputs": [],
   "source": [
    "simple_test = [\"please don't call me\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "id": "30e0b5eb-d7ab-468a-86e2-89dcd0fa0184",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[0, 1, 1, 1, 0, 0]], dtype=int64)"
      ]
     },
     "execution_count": 95,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "simple_test_dtm = vect.transform(simple_test)\n",
    "simple_test_dtm.toarray()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 96,
   "id": "9df0574c-6088-4418-92b4-d6d69c9d9d41",
   "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>cab</th>\n",
       "      <th>call</th>\n",
       "      <th>me</th>\n",
       "      <th>please</th>\n",
       "      <th>tonight</th>\n",
       "      <th>you</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   cab  call  me  please  tonight  you\n",
       "0    0     1   1       1        0    0"
      ]
     },
     "execution_count": 96,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\n",
    "# examine the vocabulary and document-term matrix together\n",
    "pd.DataFrame(simple_test_dtm.toarray(), columns=vect.get_feature_names_out())"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ec45268a-bc57-4a34-b496-b30751119403",
   "metadata": {},
   "source": [
    "## Multinomial Naive Bayes Implementation\n",
    "\n",
    "- Your task is to implement Mutlinomial Naive Bayes from scratch, you can use numpy to vectorize your code and matplotlib  to show your analysis.\n",
    "- Below some information has given from the documentation about Multinomial Naive Bayes, this will give you some idea about using *Smoothing Priors*.\n",
    "- There is a sub-question for experimenting with $\\alpha > 0$, you don't have to implement it separetely, try to incomporate it in same Model Class / Function."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d3c9e325-5e71-4810-83c9-a2e3b5a547b2",
   "metadata": {},
   "source": [
    "📌 From the scikit-learn documentation:\n",
    "\n",
    "- Multinomial Naive Bayes implements the naive Bayes algorithm for multinomially distributed data, and is one of the two classic naive Bayes variants used in text classification (where the data are typically represented as word vector counts, although tf-idf vectors are also known to work well in practice).\n",
    "\n",
    "- The distribution $\\theta_y = (\\theta_{y1}, \\theta_{y2}, \\dots, \\theta_{yn})$ is parametrized by vectors for each class $y$, where $n$ is the number of features (in text classification, the size of the vocabulary) and $\\theta_{yi}$ is the probability $P(x_i|y)$ of feature appearing in a sample belonging to class.\n",
    "\n",
    "- The parameters $\\theta_y$ is estimated by a smoothed version of maximum likelihood, i.e. relative frequency counting:\n",
    "\n",
    "$$\n",
    "\\hat{\\theta}_{yi} = \\frac{N_{yi} + \\alpha}{N_{y} + \\alpha n}\n",
    "$$\n",
    "\n",
    " where $N_{yi} = \\sum_{x \\in T}{x_i}$ is the number of times feature $i $ appears in a sample of class in the training set $T$, and $N_{y} = \\sum^{n}_{i=1}{N_{yi}}$ is the total count of all features for class $y$.\n",
    "\n",
    "- The smoothing priors $\\alpha \\gt 0$ accounts for features not present in the learning samples and **prevents zero probabilities** in further computations. Setting $\\alpha = 1$ is called Laplace smoothing, while $\\alpha \\lt 1$ is called Lidstone smoothing.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 97,
   "id": "14862c11-1b3c-4314-a64e-f98a185ce394",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "37062\n",
      "37062\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[('call', 288),\n",
       " ('get', 256),\n",
       " ('ltgt', 203),\n",
       " ('know', 191),\n",
       " ('go', 191),\n",
       " ('like', 180),\n",
       " ('got', 153),\n",
       " ('come', 149),\n",
       " ('time', 142),\n",
       " ('day', 132),\n",
       " ('lor', 131),\n",
       " ('want', 131),\n",
       " ('Ill', 126),\n",
       " ('Call', 120),\n",
       " ('need', 118),\n",
       " ('back', 116),\n",
       " ('good', 115),\n",
       " ('one', 115),\n",
       " ('send', 115),\n",
       " ('going', 113),\n",
       " ('love', 111),\n",
       " ('home', 109),\n",
       " ('Ok', 108),\n",
       " ('still', 106),\n",
       " ('text', 99),\n",
       " ('see', 99),\n",
       " ('later', 92),\n",
       " ('ok', 92),\n",
       " ('Ì', 90),\n",
       " ('n', 88),\n",
       " ('free', 86),\n",
       " ('think', 85),\n",
       " ('mobile', 85),\n",
       " ('da', 84),\n",
       " ('cant', 84),\n",
       " ('much', 83),\n",
       " ('r', 82),\n",
       " ('take', 81),\n",
       " ('today', 81),\n",
       " ('way', 80),\n",
       " ('FREE', 79),\n",
       " ('week', 79),\n",
       " ('Sorry', 78),\n",
       " ('give', 77),\n",
       " ('night', 76),\n",
       " ('phone', 76),\n",
       " ('tell', 75),\n",
       " ('reply', 74),\n",
       " ('make', 73),\n",
       " ('work', 71),\n",
       " ('Hey', 70),\n",
       " ('amp', 65),\n",
       " ('new', 65),\n",
       " ('tomorrow', 65),\n",
       " ('Hi', 64),\n",
       " ('already', 63),\n",
       " ('right', 63),\n",
       " ('said', 62),\n",
       " ('Oh', 60),\n",
       " ('ask', 60),\n",
       " ('say', 60),\n",
       " ('well', 59),\n",
       " ('message', 59),\n",
       " ('number', 59),\n",
       " ('great', 57),\n",
       " ('really', 57),\n",
       " ('prize', 57),\n",
       " ('contact', 56),\n",
       " ('meet', 56),\n",
       " ('msg', 56),\n",
       " ('would', 55),\n",
       " ('ill', 55),\n",
       " ('Please', 55),\n",
       " ('didnt', 54),\n",
       " ('claim', 54),\n",
       " ('Txt', 54),\n",
       " ('e', 54),\n",
       " ('Good', 53),\n",
       " ('stop', 53),\n",
       " ('pick', 52),\n",
       " ('every', 52),\n",
       " ('could', 52),\n",
       " ('please', 50),\n",
       " ('find', 49),\n",
       " ('buy', 49),\n",
       " ('Pls', 49),\n",
       " ('sent', 48),\n",
       " ('sure', 48),\n",
       " ('txt', 48),\n",
       " ('thing', 48),\n",
       " ('around', 47),\n",
       " ('STOP', 47),\n",
       " ('per', 47),\n",
       " ('morning', 46),\n",
       " ('next', 46),\n",
       " ('last', 46),\n",
       " ('wat', 46),\n",
       " ('Lol', 46),\n",
       " ('many', 46),\n",
       " ('anything', 45),\n",
       " ('soon', 45),\n",
       " ('miss', 45),\n",
       " ('cash', 45),\n",
       " ('feel', 45),\n",
       " ('care', 44),\n",
       " ('tonight', 44),\n",
       " ('gonna', 44),\n",
       " ('something', 44),\n",
       " ('told', 43),\n",
       " ('place', 43),\n",
       " ('wont', 43),\n",
       " ('life', 43),\n",
       " ('also', 41),\n",
       " ('leave', 41),\n",
       " ('wait', 41),\n",
       " ('first', 41),\n",
       " ('us', 41),\n",
       " ('people', 40),\n",
       " ('service', 40),\n",
       " ('getting', 40),\n",
       " ('thats', 40),\n",
       " ('late', 40),\n",
       " ('dear', 40),\n",
       " ('sleep', 39),\n",
       " ('babe', 39),\n",
       " ('Thanks', 38),\n",
       " ('let', 38),\n",
       " ('b', 38),\n",
       " ('wan', 38),\n",
       " ('Happy', 38),\n",
       " ('help', 37),\n",
       " ('cos', 37),\n",
       " ('coming', 37),\n",
       " ('ÌÏ', 37),\n",
       " ('Yes', 37),\n",
       " ('hope', 37),\n",
       " ('money', 37),\n",
       " ('class', 36),\n",
       " ('Haha', 36),\n",
       " ('Yup', 36),\n",
       " ('Text', 36),\n",
       " ('lunch', 36),\n",
       " ('dun', 36),\n",
       " ('friends', 36),\n",
       " ('keep', 36),\n",
       " ('Well', 36),\n",
       " ('trying', 35),\n",
       " ('youre', 35),\n",
       " ('happy', 35),\n",
       " ('job', 35),\n",
       " ('someone', 35),\n",
       " ('Dear', 34),\n",
       " ('Hope', 34),\n",
       " ('yet', 34),\n",
       " ('Free', 34),\n",
       " ('even', 34),\n",
       " ('stuff', 34),\n",
       " ('c', 34),\n",
       " ('nice', 33),\n",
       " ('try', 33),\n",
       " ('always', 33),\n",
       " ('Yeah', 33),\n",
       " ('waiting', 33),\n",
       " ('K', 33),\n",
       " ('things', 32),\n",
       " ('talk', 32),\n",
       " ('best', 32),\n",
       " ('thk', 32),\n",
       " ('Nokia', 32),\n",
       " ('went', 32),\n",
       " ('lar', 32),\n",
       " ('use', 31),\n",
       " ('friend', 31),\n",
       " ('meeting', 31),\n",
       " ('R', 30),\n",
       " ('sorry', 30),\n",
       " ('URGENT', 30),\n",
       " ('Reply', 30),\n",
       " ('finish', 30),\n",
       " ('liao', 30),\n",
       " ('thought', 30),\n",
       " ('another', 30),\n",
       " ('Thats', 30),\n",
       " ('fine', 30),\n",
       " ('name', 30),\n",
       " ('done', 30),\n",
       " ('pls', 29),\n",
       " ('days', 29),\n",
       " ('havent', 29),\n",
       " ('line', 29),\n",
       " ('Hello', 29),\n",
       " ('eat', 29),\n",
       " ('Send', 29),\n",
       " ('ya', 29),\n",
       " ('ready', 28),\n",
       " ('Get', 28),\n",
       " ('draw', 28),\n",
       " ('min', 28),\n",
       " ('end', 28),\n",
       " ('bit', 28),\n",
       " ('person', 28),\n",
       " ('car', 28),\n",
       " ('Ive', 27),\n",
       " ('shows', 27),\n",
       " ('long', 27),\n",
       " ('awarded', 27),\n",
       " ('customer', 27),\n",
       " ('month', 27),\n",
       " ('dinner', 27),\n",
       " ('hi', 27),\n",
       " ('receive', 27),\n",
       " ('dat', 26),\n",
       " ('real', 26),\n",
       " ('year', 26),\n",
       " ('Gud', 26),\n",
       " ('pay', 26),\n",
       " ('special', 26),\n",
       " ('room', 26),\n",
       " ('better', 26),\n",
       " ('plan', 26),\n",
       " ('big', 26),\n",
       " ('PO', 26),\n",
       " ('Claim', 26),\n",
       " ('watch', 26),\n",
       " ('wish', 26),\n",
       " ('guys', 26),\n",
       " ('problem', 26),\n",
       " ('man', 25),\n",
       " ('C', 25),\n",
       " ('hour', 25),\n",
       " ('ever', 25),\n",
       " ('Aight', 25),\n",
       " ('enough', 25),\n",
       " ('mins', 25),\n",
       " ('word', 24),\n",
       " ('See', 24),\n",
       " ('live', 24),\n",
       " ('might', 24),\n",
       " ('cool', 24),\n",
       " ('birthday', 24),\n",
       " ('little', 23),\n",
       " ('shit', 23),\n",
       " ('guess', 23),\n",
       " ('able', 23),\n",
       " ('quite', 23),\n",
       " ('early', 23),\n",
       " ('Mobile', 23),\n",
       " ('never', 23),\n",
       " ('latest', 23),\n",
       " ('Wat', 23),\n",
       " ('put', 23),\n",
       " ('two', 23),\n",
       " ('chat', 23),\n",
       " ('guaranteed', 23),\n",
       " ('half', 23),\n",
       " ('house', 23),\n",
       " ('bed', 23),\n",
       " ('award', 23),\n",
       " ('bad', 22),\n",
       " ('weekend', 22),\n",
       " ('smile', 22),\n",
       " ('V', 22),\n",
       " ('evening', 22),\n",
       " ('landline', 22),\n",
       " ('abt', 22),\n",
       " ('minutes', 22),\n",
       " ('lot', 22),\n",
       " ('selected', 22),\n",
       " ('ah', 22),\n",
       " ('den', 22),\n",
       " ('goes', 22),\n",
       " ('hes', 21),\n",
       " ('CALL', 21),\n",
       " ('TCs', 21),\n",
       " ('else', 21),\n",
       " ('guy', 21),\n",
       " ('wanted', 21),\n",
       " ('yeah', 21),\n",
       " ('face', 21),\n",
       " ('sweet', 21),\n",
       " ('win', 21),\n",
       " ('bus', 21),\n",
       " ('heart', 21),\n",
       " ('hear', 21),\n",
       " ('start', 21),\n",
       " ('dis', 21),\n",
       " ('Love', 21),\n",
       " ('reach', 21),\n",
       " ('collection', 21),\n",
       " ('looking', 21),\n",
       " ('k', 21),\n",
       " ('account', 21),\n",
       " ('Hows', 20),\n",
       " ('ive', 20),\n",
       " ('apply', 20),\n",
       " ('mail', 20),\n",
       " ('forgot', 20),\n",
       " ('working', 20),\n",
       " ('left', 20),\n",
       " ('watching', 20),\n",
       " ('New', 20),\n",
       " ('kiss', 20),\n",
       " ('princess', 20),\n",
       " ('thanks', 20),\n",
       " ('leh', 20),\n",
       " ('girl', 20),\n",
       " ('till', 20),\n",
       " ('oh', 20),\n",
       " ('Box', 20),\n",
       " ('says', 20),\n",
       " ('called', 19),\n",
       " ('Great', 19),\n",
       " ('Orange', 19),\n",
       " ('easy', 19),\n",
       " ('tmr', 19),\n",
       " ('mind', 19),\n",
       " ('bring', 19),\n",
       " ('fun', 19),\n",
       " ('tv', 19),\n",
       " ('v', 19),\n",
       " ('haf', 19),\n",
       " ('wif', 19),\n",
       " ('play', 19),\n",
       " ('attempt', 19),\n",
       " ('show', 19),\n",
       " ('shopping', 19),\n",
       " ('town', 19),\n",
       " ('tried', 19),\n",
       " ('x', 18),\n",
       " ('network', 18),\n",
       " ('hours', 18),\n",
       " ('Yo', 18),\n",
       " ('SMS', 18),\n",
       " ('til', 18),\n",
       " ('hav', 18),\n",
       " ('GUARANTEED', 18),\n",
       " ('Tell', 18),\n",
       " ('asked', 18),\n",
       " ('hair', 18),\n",
       " ('sexy', 18),\n",
       " ('means', 18),\n",
       " ('look', 18),\n",
       " ('may', 18),\n",
       " ('wants', 18),\n",
       " ('part', 18),\n",
       " ('check', 18),\n",
       " ('worry', 18),\n",
       " ('luv', 17),\n",
       " ('holiday', 17),\n",
       " ('chance', 17),\n",
       " ('WIN', 17),\n",
       " ('though', 17),\n",
       " ('words', 17),\n",
       " ('tone', 17),\n",
       " ('address', 17),\n",
       " ('test', 17),\n",
       " ('shall', 17),\n",
       " ('stay', 17),\n",
       " ('away', 17),\n",
       " ('school', 17),\n",
       " ('actually', 17),\n",
       " ('oso', 17),\n",
       " ('doesnt', 17),\n",
       " ('baby', 17),\n",
       " ('Got', 17),\n",
       " ('either', 17),\n",
       " ('years', 17),\n",
       " ('pa', 17),\n",
       " ('wanna', 17),\n",
       " ('saw', 17),\n",
       " ('Todays', 17),\n",
       " ('smoke', 17),\n",
       " ('Holiday', 17),\n",
       " ('SAE', 17),\n",
       " ('details', 17),\n",
       " ('remember', 16),\n",
       " ('lei', 16),\n",
       " ('comes', 16),\n",
       " ('Prize', 16),\n",
       " ('since', 16),\n",
       " ('afternoon', 16),\n",
       " ('times', 16),\n",
       " ('finished', 16),\n",
       " ('everything', 16),\n",
       " ('trip', 16),\n",
       " ('game', 16),\n",
       " ('lose', 16),\n",
       " ('join', 16),\n",
       " ('Cos', 16),\n",
       " ('Today', 16),\n",
       " ('online', 16),\n",
       " ('calling', 16),\n",
       " ('TC', 16),\n",
       " ('collect', 16),\n",
       " ('office', 16),\n",
       " ('ltDECIMALgt', 16),\n",
       " ('noe', 16),\n",
       " ('top', 16),\n",
       " ('Thanx', 16),\n",
       " ('LOVE', 16),\n",
       " ('made', 15),\n",
       " ('bored', 15),\n",
       " ('order', 15),\n",
       " ('food', 15),\n",
       " ('sch', 15),\n",
       " ('making', 15),\n",
       " ('camera', 15),\n",
       " ('nothing', 15),\n",
       " ('unsubscribe', 15),\n",
       " ('came', 15),\n",
       " ('tot', 15),\n",
       " ('yesterday', 15),\n",
       " ('must', 15),\n",
       " ('charge', 15),\n",
       " ('probably', 15),\n",
       " ('answer', 15),\n",
       " ('without', 15),\n",
       " ('set', 15),\n",
       " ('Valid', 15),\n",
       " ('rite', 15),\n",
       " ('boy', 15),\n",
       " ('Xmas', 15),\n",
       " ('rate', 15),\n",
       " ('minute', 15),\n",
       " ('together', 15),\n",
       " ('Enjoy', 15),\n",
       " ('happen', 15),\n",
       " ('hows', 15),\n",
       " ('await', 15),\n",
       " ('entry', 15),\n",
       " ('tomo', 15),\n",
       " ('email', 15),\n",
       " ('juz', 15),\n",
       " ('knw', 15),\n",
       " ('weeks', 15),\n",
       " ('pain', 15),\n",
       " ('boytoy', 15),\n",
       " ('maybe', 15),\n",
       " ('sat', 14),\n",
       " ('true', 14),\n",
       " ('saying', 14),\n",
       " ('dad', 14),\n",
       " ('Day', 14),\n",
       " ('full', 14),\n",
       " ('sis', 14),\n",
       " ('sms', 14),\n",
       " ('nite', 14),\n",
       " ('anyone', 14),\n",
       " ('xxx', 14),\n",
       " ('cost', 14),\n",
       " ('second', 14),\n",
       " ('driving', 14),\n",
       " ('smth', 14),\n",
       " ('pics', 14),\n",
       " ('taking', 14),\n",
       " ('Cant', 14),\n",
       " ('world', 14),\n",
       " ('drive', 14),\n",
       " ('busy', 14),\n",
       " ('GO', 14),\n",
       " ('open', 14),\n",
       " ('forget', 14),\n",
       " ('question', 14),\n",
       " ('Okie', 14),\n",
       " ('mean', 14),\n",
       " ('everyone', 14),\n",
       " ('video', 14),\n",
       " ('wake', 14),\n",
       " ('goin', 14),\n",
       " ('Ltd', 14),\n",
       " ('important', 14),\n",
       " ('colour', 14),\n",
       " ('Urgent', 14),\n",
       " ('offer', 14),\n",
       " ('jus', 14),\n",
       " ('Take', 14),\n",
       " ('company', 14),\n",
       " ('anyway', 14),\n",
       " ('angry', 13),\n",
       " ('wid', 13),\n",
       " ('story', 13),\n",
       " ('Sir', 13),\n",
       " ('Ya', 13),\n",
       " ('visit', 13),\n",
       " ('telling', 13),\n",
       " ('drop', 13),\n",
       " ('Going', 13),\n",
       " ('project', 13),\n",
       " ('XXX', 13),\n",
       " ('took', 13),\n",
       " ('prob', 13),\n",
       " ('missed', 13),\n",
       " ('thinking', 13),\n",
       " ('Whats', 13),\n",
       " ('speak', 13),\n",
       " ('started', 13),\n",
       " ('used', 13),\n",
       " ('sir', 13),\n",
       " ('change', 13),\n",
       " ('weekly', 13),\n",
       " ('Babe', 13),\n",
       " ('fuck', 13),\n",
       " ('texts', 13),\n",
       " ('worth', 13),\n",
       " ('land', 13),\n",
       " ('wit', 13),\n",
       " ('Speak', 13),\n",
       " ('hurt', 13),\n",
       " ('TsCs', 13),\n",
       " ('wonderful', 13),\n",
       " ('run', 13),\n",
       " ('far', 13),\n",
       " ('tones', 13),\n",
       " ('UK', 13),\n",
       " ('chennai', 13),\n",
       " ('hot', 13),\n",
       " ('mu', 13),\n",
       " ('found', 12),\n",
       " ('alone', 12),\n",
       " ('Nope', 12),\n",
       " ('Want', 12),\n",
       " ('mob', 12),\n",
       " ('NEW', 12),\n",
       " ('leaving', 12),\n",
       " ('shes', 12),\n",
       " ('mates', 12),\n",
       " ('movie', 12),\n",
       " ('gift', 12),\n",
       " ('snow', 12),\n",
       " ('ard', 12),\n",
       " ('Thank', 12),\n",
       " ('vouchers', 12),\n",
       " ('meant', 12),\n",
       " ('Last', 12),\n",
       " ('lesson', 12),\n",
       " ('YES', 12),\n",
       " ('wil', 12),\n",
       " ('father', 12),\n",
       " ('calls', 12),\n",
       " ('book', 12),\n",
       " ('CHAT', 12),\n",
       " ('theres', 12),\n",
       " ('g', 12),\n",
       " ('Congrats', 12),\n",
       " ('feeling', 12),\n",
       " ('type', 12),\n",
       " ('gone', 12),\n",
       " ('enjoy', 12),\n",
       " ('close', 12),\n",
       " ('college', 12),\n",
       " ('family', 12),\n",
       " ('card', 12),\n",
       " ('God', 12),\n",
       " ('outside', 12),\n",
       " ('light', 12),\n",
       " ('needs', 12),\n",
       " ('whole', 11),\n",
       " ('almost', 11),\n",
       " ('sound', 11),\n",
       " ('phones', 11),\n",
       " ('Dunno', 11),\n",
       " ('points', 11),\n",
       " ('post', 11),\n",
       " ('sad', 11),\n",
       " ('Bonus', 11),\n",
       " ('supposed', 11),\n",
       " ('Let', 11),\n",
       " ('Customer', 11),\n",
       " ('song', 11),\n",
       " ('head', 11),\n",
       " ('listen', 11),\n",
       " ('cause', 11),\n",
       " ('wot', 11),\n",
       " ('months', 11),\n",
       " ('price', 11),\n",
       " ('pounds', 11),\n",
       " ('treat', 11),\n",
       " ('download', 11),\n",
       " ('Win', 11),\n",
       " ('point', 11),\n",
       " ('happened', 11),\n",
       " ('earlier', 11),\n",
       " ('content', 11),\n",
       " ('sea', 11),\n",
       " ('pretty', 11),\n",
       " ('Double', 11),\n",
       " ('SavaMob', 11),\n",
       " ('search', 11),\n",
       " ('ringtone', 11),\n",
       " ('Update', 11),\n",
       " ('whats', 11),\n",
       " ('whatever', 11),\n",
       " ('de', 11),\n",
       " ('yr', 11),\n",
       " ('id', 11),\n",
       " ('dreams', 11),\n",
       " ('thinks', 11),\n",
       " ('plz', 11),\n",
       " ('dude', 11),\n",
       " ('Anything', 11),\n",
       " ('wk', 11),\n",
       " ('takes', 11),\n",
       " ('sun', 11),\n",
       " ('believe', 11),\n",
       " ('Even', 11),\n",
       " ('messages', 11),\n",
       " ('E', 11),\n",
       " ('smiling', 11),\n",
       " ('dating', 11),\n",
       " ('walk', 11),\n",
       " ('makes', 11),\n",
       " ('player', 11),\n",
       " ('Congratulations', 11),\n",
       " ('tho', 11),\n",
       " ('couple', 11),\n",
       " ('Cost', 11),\n",
       " ('rock', 10),\n",
       " ('Stop', 10),\n",
       " ('Miss', 10),\n",
       " ('paper', 10),\n",
       " ('bout', 10),\n",
       " ('wasnt', 10),\n",
       " ('drink', 10),\n",
       " ('friday', 10),\n",
       " ('Nothing', 10),\n",
       " ('operator', 10),\n",
       " ('picking', 10),\n",
       " ('hold', 10),\n",
       " ('isnt', 10),\n",
       " ('final', 10),\n",
       " ('Ugh', 10),\n",
       " ('Dude', 10),\n",
       " ('cum', 10),\n",
       " ('fucking', 10),\n",
       " ('gas', 10),\n",
       " ('party', 10),\n",
       " ('update', 10),\n",
       " ('Cool', 10),\n",
       " ('area', 10),\n",
       " ('Code', 10),\n",
       " ('w', 10),\n",
       " ('grins', 10),\n",
       " ('less', 10),\n",
       " ('gets', 10),\n",
       " ('currently', 10),\n",
       " ('Okay', 10),\n",
       " ('Yar', 10),\n",
       " ('BSLVYL', 10),\n",
       " ('offers', 10),\n",
       " ('within', 10),\n",
       " ('Go', 10),\n",
       " ('choose', 10),\n",
       " ('LIVE', 10),\n",
       " ('Also', 10),\n",
       " ('lucky', 10),\n",
       " ('Life', 10),\n",
       " ('TEXT', 10),\n",
       " ('confirm', 10),\n",
       " ('Shes', 10),\n",
       " ('pub', 10),\n",
       " ('lots', 10),\n",
       " ('plus', 10),\n",
       " ('secret', 10),\n",
       " ('via', 10),\n",
       " ('hard', 10),\n",
       " ('NTT', 10),\n",
       " ('decided', 10),\n",
       " ('Smile', 10),\n",
       " ('gd', 10),\n",
       " ('ex', 10),\n",
       " ('course', 10),\n",
       " ('numbers', 10),\n",
       " ('Still', 10),\n",
       " ('May', 10),\n",
       " ('simple', 10),\n",
       " ('girls', 10),\n",
       " ('brother', 10),\n",
       " ('friendship', 10),\n",
       " ('complimentary', 10),\n",
       " ('okay', 10),\n",
       " ('extra', 10),\n",
       " ('news', 10),\n",
       " ('figure', 10),\n",
       " ('hungry', 10),\n",
       " ('fast', 10),\n",
       " ('whenever', 10),\n",
       " ('Come', 10),\n",
       " ('games', 10),\n",
       " ('code', 10),\n",
       " ('pic', 10),\n",
       " ('move', 9),\n",
       " ('Wen', 9),\n",
       " ('crazy', 9),\n",
       " ('Cs', 9),\n",
       " ('lovely', 9),\n",
       " ('computer', 9),\n",
       " ('gave', 9),\n",
       " ('Caller', 9),\n",
       " ('sending', 9),\n",
       " ('Hee', 9),\n",
       " ('Jus', 9),\n",
       " ('GOOD', 9),\n",
       " ('yes', 9),\n",
       " ('representative', 9),\n",
       " ('questions', 9),\n",
       " ('weekends', 9),\n",
       " ('Alright', 9),\n",
       " ('aft', 9),\n",
       " ('mine', 9),\n",
       " ('Motorola', 9),\n",
       " ('brings', 9),\n",
       " ('ones', 9),\n",
       " ('cut', 9),\n",
       " ('Camera', 9),\n",
       " ('eg', 9),\n",
       " ('semester', 9),\n",
       " ('Lets', 9),\n",
       " ('seeing', 9),\n",
       " ('etc', 9),\n",
       " ('difficult', 9),\n",
       " ('One', 9),\n",
       " ('eve', 9),\n",
       " ('sister', 9),\n",
       " ('mah', 9),\n",
       " ('Quiz', 9),\n",
       " ('seen', 9),\n",
       " ('B', 9),\n",
       " ('Msg', 9),\n",
       " ('mate', 9),\n",
       " ('Watching', 9),\n",
       " ('rates', 9),\n",
       " ('least', 9),\n",
       " ('Ard', 9),\n",
       " ('Dun', 9),\n",
       " ('touch', 9),\n",
       " ('Nah', 9),\n",
       " ('Club', 9),\n",
       " ('park', 9),\n",
       " ('gud', 9),\n",
       " ('Year', 9),\n",
       " ('side', 9),\n",
       " ('PRIVATE', 9),\n",
       " ('Account', 9),\n",
       " ('Statement', 9),\n",
       " ('unredeemed', 9),\n",
       " ('Identifier', 9),\n",
       " ('Expires', 9),\n",
       " ('music', 9),\n",
       " ('gotta', 9),\n",
       " ('CASH', 9),\n",
       " ('Awesome', 9),\n",
       " ('missing', 9),\n",
       " ('beautiful', 9),\n",
       " ('happiness', 9),\n",
       " ('rply', 9),\n",
       " ('pm', 9),\n",
       " ('X', 9),\n",
       " ('delivery', 9),\n",
       " ('dnt', 8),\n",
       " ('deep', 8),\n",
       " ('voucher', 8),\n",
       " ('loving', 8),\n",
       " ('hoping', 8),\n",
       " ('Thk', 8),\n",
       " ('Need', 8),\n",
       " ('fri', 8),\n",
       " ('TIME', 8),\n",
       " ('direct', 8),\n",
       " ('talking', 8),\n",
       " ('Offer', 8),\n",
       " ('mum', 8),\n",
       " ('neva', 8),\n",
       " ('reason', 8),\n",
       " ('NICE', 8),\n",
       " ('coz', 8),\n",
       " ('wonder', 8),\n",
       " ('Morning', 8),\n",
       " ('knew', 8),\n",
       " ('Think', 8),\n",
       " ('date', 8),\n",
       " ('old', 8),\n",
       " ('opt', 8),\n",
       " ('ring', 8),\n",
       " ('valued', 8),\n",
       " ('Omg', 8),\n",
       " ('felt', 8),\n",
       " ('Make', 8),\n",
       " ('comin', 8),\n",
       " ('wondering', 8),\n",
       " ('luck', 8),\n",
       " ('hit', 8),\n",
       " ('Huh', 8),\n",
       " ('Merry', 8),\n",
       " ('kind', 8),\n",
       " ('reaching', 8),\n",
       " ('ten', 8),\n",
       " ('carlos', 8),\n",
       " ('drugs', 8),\n",
       " ('break', 8),\n",
       " ('Wish', 8),\n",
       " ('wishes', 8),\n",
       " ('Actually', 8),\n",
       " ('Joys', 8),\n",
       " ('John', 8),\n",
       " ('crave', 8),\n",
       " ('txts', 8),\n",
       " ('seems', 8),\n",
       " ('ends', 8),\n",
       " ('Would', 8),\n",
       " ('credits', 8),\n",
       " ('god', 8),\n",
       " ('hell', 8),\n",
       " ('Youre', 8),\n",
       " ('Ha', 8),\n",
       " ('mrng', 8),\n",
       " ('safe', 8),\n",
       " ('decide', 8),\n",
       " ('loads', 8),\n",
       " ('wen', 8),\n",
       " ('disturb', 8),\n",
       " ('wrong', 8),\n",
       " ('awesome', 8),\n",
       " ('admirer', 8),\n",
       " ('Ufind', 8),\n",
       " ('Rreveal', 8),\n",
       " ('specialcall', 8),\n",
       " ('frnds', 8),\n",
       " ('link', 8),\n",
       " ('sick', 8),\n",
       " ('TONE', 8),\n",
       " ('chikku', 8),\n",
       " ('dunno', 8),\n",
       " ('case', 8),\n",
       " ('access', 8),\n",
       " ('write', 8),\n",
       " ('leaves', 8),\n",
       " ('blue', 8),\n",
       " ('laugh', 8),\n",
       " ('aha', 8),\n",
       " ('woke', 8),\n",
       " ('swing', 8),\n",
       " ('Theres', 8),\n",
       " ('kinda', 8),\n",
       " ('ass', 8),\n",
       " ('anytime', 8),\n",
       " ('Phone', 8),\n",
       " ('Plz', 8),\n",
       " ('feels', 8),\n",
       " ('thank', 8),\n",
       " ('Anyway', 8),\n",
       " ('Someone', 8),\n",
       " ('yo', 8),\n",
       " ('Service', 8),\n",
       " ('club', 8),\n",
       " ('self', 8),\n",
       " ('gym', 8),\n",
       " ('charged', 8),\n",
       " ('usf', 8),\n",
       " ('add', 8),\n",
       " ('ending', 8),\n",
       " ('match', 8),\n",
       " ('frm', 8),\n",
       " ('Maybe', 8),\n",
       " ('march', 8),\n",
       " ('Video', 8),\n",
       " ('film', 8),\n",
       " ('askd', 7),\n",
       " ('dream', 7),\n",
       " ('available', 7),\n",
       " ('mid', 7),\n",
       " ('wana', 7),\n",
       " ('Bluetooth', 7),\n",
       " ('bt', 7),\n",
       " ('fact', 7),\n",
       " ('water', 7),\n",
       " ('Mins', 7),\n",
       " ('somebody', 7),\n",
       " ('exam', 7),\n",
       " ('MUCH', 7),\n",
       " ('th', 7),\n",
       " ('NOKIA', 7),\n",
       " ('Check', 7),\n",
       " ('TODAY', 7),\n",
       " ('discount', 7),\n",
       " ('credit', 7),\n",
       " ('train', 7),\n",
       " ('support', 7),\n",
       " ('wine', 7),\n",
       " ('sight', 7),\n",
       " ('heard', 7),\n",
       " ('Wanna', 7),\n",
       " ('notice', 7),\n",
       " ('rent', 7),\n",
       " ('longer', 7),\n",
       " ('mayb', 7),\n",
       " ('Den', 7),\n",
       " ('FREEPHONE', 7),\n",
       " ('HMV', 7),\n",
       " ('wednesday', 7),\n",
       " ('BOX', 7),\n",
       " ('Probably', 7),\n",
       " ('Sounds', 7),\n",
       " ('across', 7),\n",
       " ('studying', 7),\n",
       " ('malaria', 7),\n",
       " ('players', 7),\n",
       " ('poor', 7),\n",
       " ('Never', 7),\n",
       " ('Latest', 7),\n",
       " ('huh', 7),\n",
       " ('uncle', 7),\n",
       " ('spend', 7),\n",
       " ('login', 7),\n",
       " ('changed', 7),\n",
       " ('frnd', 7),\n",
       " ('AGE', 7),\n",
       " ('fr', 7),\n",
       " ('request', 7),\n",
       " ('lol', 7),\n",
       " ('MUSIC', 7),\n",
       " ('sleeping', 7),\n",
       " ('truth', 7),\n",
       " ('slow', 7),\n",
       " ('Croydon', 7),\n",
       " ('library', 7),\n",
       " ('trouble', 7),\n",
       " ('weed', 7),\n",
       " ('cell', 7),\n",
       " ('reward', 7),\n",
       " ('TXT', 7),\n",
       " ('Co', 7),\n",
       " ('meh', 7),\n",
       " ('bb', 7),\n",
       " ('voice', 7),\n",
       " ('urself', 7),\n",
       " ('Plus', 7),\n",
       " ('Oops', 7),\n",
       " ('bank', 7),\n",
       " ('Really', 7),\n",
       " ('Tenerife', 7),\n",
       " ('quick', 7),\n",
       " ('frens', 7),\n",
       " ('shower', 7),\n",
       " ('orchard', 7),\n",
       " ('Wait', 7),\n",
       " ('Guess', 7),\n",
       " ('Wow', 7),\n",
       " ('mrt', 7),\n",
       " ('plans', 7),\n",
       " ('na', 7),\n",
       " ('met', 7),\n",
       " ('study', 7),\n",
       " ('reading', 7),\n",
       " ('FreeMsg', 7),\n",
       " ('pass', 7),\n",
       " ('nt', 7),\n",
       " ('World', 7),\n",
       " ('matches', 7),\n",
       " ('msgs', 7),\n",
       " ('HAPPY', 7),\n",
       " ('Shall', 7),\n",
       " ('mom', 7),\n",
       " ('tel', 7),\n",
       " ('hand', 7),\n",
       " ('darren', 7),\n",
       " ('slave', 7),\n",
       " ('finally', 7),\n",
       " ('Dad', 7),\n",
       " ('Sony', 7),\n",
       " ('worried', 7),\n",
       " ('youll', 7),\n",
       " ('Fuck', 7),\n",
       " ('net', 7),\n",
       " ('picked', 7),\n",
       " ('Gods', 7),\n",
       " ('alright', 7),\n",
       " ('comp', 7),\n",
       " ('small', 7),\n",
       " ('HELLO', 7),\n",
       " ('forever', 7),\n",
       " ('bid', 7),\n",
       " ('seriously', 6),\n",
       " ('using', 6),\n",
       " ('however', 6),\n",
       " ('rest', 6),\n",
       " ('short', 6),\n",
       " ('lessons', 6),\n",
       " ('Dogging', 6),\n",
       " ('noon', 6),\n",
       " ('poly', 6),\n",
       " ('tariffs', 6),\n",
       " ('lazy', 6),\n",
       " ('saturday', 6),\n",
       " ('GOT', 6),\n",
       " ('txting', 6),\n",
       " ('POBox', 6),\n",
       " ('heavy', 6),\n",
       " ('usual', 6),\n",
       " ('registered', 6),\n",
       " ('enter', 6),\n",
       " ...]"
      ]
     },
     "execution_count": 97,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\"\"\"\n",
    "Your code here\n",
    "\"\"\"\n",
    "from collections import Counter\n",
    "words=[]\n",
    "for email in X_train:\n",
    "    words+=email.split(\" \")\n",
    "print(len(words))\n",
    "# print(words)\n",
    "for i in range(len(words)):\n",
    "    if not words[i].isalpha():\n",
    "        words[i]=\"\"\n",
    "print(len(words))\n",
    "word_dict=Counter(words)\n",
    "len(word_dict)\n",
    "del word_dict[\"\"]\n",
    "word_dict=word_dict.most_common(4000)\n",
    "word_dict\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0dbcd1df-1c3d-477d-b0f6-c8c17868c797",
   "metadata": {},
   "source": [
    "## Vectorizing Training Sample\n",
    "\n",
    "- Use the Helper code above to vectorize for training samples\n",
    "- Don't overthink it, its very easy to do"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 98,
   "id": "c2561b73-fc1b-47aa-8ae9-6d1ba8ad3015",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(4179, 4000)\n"
     ]
    }
   ],
   "source": [
    "\"\"\"\n",
    "Your code here\n",
    "\"\"\"\n",
    "mat=[]\n",
    "res=[]\n",
    "for email,lable in zip(X_train,y_train):\n",
    "    data=[]\n",
    "    wrd=email.split(\" \")\n",
    "    for j in word_dict:\n",
    "        data.append(wrd.count(j[0]))\n",
    "    \n",
    "    mat.append(data)\n",
    "    res.append(lable)\n",
    "\n",
    "mat=np.array(mat)\n",
    "res=np.array(res)\n",
    "print(mat.shape)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ec22e087-4e03-4428-8f47-4cb4f2ec1645",
   "metadata": {},
   "source": [
    "## Calculate Priors and Estimate Model's performance on Training Sample\n",
    "\n",
    "- Calculate priors based on Training Sample using your NB implementation\n",
    "- Evaluate your model's performance on Training Data ($\\alpha = 0$)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 99,
   "id": "1aa85aed-f38c-4acc-9c76-339d5914118d",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy on train sample  0.9916247906197655\n"
     ]
    }
   ],
   "source": [
    "\"\"\"\n",
    "Your code here\n",
    "\"\"\"\n",
    "class Multi_NB:\n",
    "    def __init__(self,alpha=0.5):\n",
    "        self.alpha = alpha\n",
    "    def fit(self, x, y):\n",
    "        self.y = y\n",
    "        \n",
    "        self.class_ = np.unique(y)\n",
    "\n",
    "        self.samp_train, self.column = x.shape\n",
    "        self.n_class_ = self.class_.shape[0]\n",
    "        temp1, self.dist = np.unique(self.y,return_counts=True)\n",
    "        P = np.zeros((self.n_class_))\n",
    "        \n",
    "        \n",
    "        for i in range(self.class_.shape[0]):\n",
    "            pirpob= self.dist[i] / self.samp_train\n",
    "            P[i]=pirpob\n",
    "            \n",
    "        \n",
    "\n",
    "        self.N_yi = np.zeros((self.n_class_, self.column))\n",
    "        self.class_priors_ = P\n",
    "        self.N_y = np.zeros((self.n_class_))\n",
    "        for i in self.class_:\n",
    "            columnwise_sum = []\n",
    "            num=self.column\n",
    "            indices = np.argwhere(self.y==i).flatten()\n",
    "            \n",
    "            for j in range(0,num):\n",
    "                temp21=np.sum(x[indices,j])\n",
    "                columnwise_sum.append(temp21)\n",
    "            \n",
    "            self.N_yi[i] = columnwise_sum \n",
    "            self.N_y[i] = np.sum(columnwise_sum) \n",
    "    \n",
    "    def _likelyhood(self, x, h):\n",
    "       \n",
    "        tmp = []\n",
    "        t3=x.shape[0]\n",
    "        for i in range(t3):\n",
    "            Nyi=self.N_yi[h,i]\n",
    "            numerator = Nyi + self.alpha\n",
    "            extra=self.alpha * self.column\n",
    "            Ny  = self.N_y[h]\n",
    "            denominator = Ny + extra\n",
    "            res_ans=(numerator / denominator)**x[i]\n",
    "            \n",
    "            tmp.append(res_ans)\n",
    "        \n",
    "        return np.prod(tmp)\n",
    "    \n",
    "    def predict(self, X):\n",
    "        samples, features = X.shape\n",
    "        n3=X.shape[0]\n",
    "        self.predict_proba = np.zeros((samples,self.n_class_))\n",
    "        \n",
    "        for i in range(n3):\n",
    "            joint_likelyhood = np.zeros((self.n_class_))\n",
    "\n",
    "            for h in range(self.n_class_):\n",
    "                pob= self.class_priors_[h] * self._likelyhood(X[i],h)\n",
    "                joint_likelyhood[h] =pob\n",
    "                \n",
    "            denominator = np.sum(joint_likelyhood)\n",
    "            n4=self.n_class_\n",
    "            for h in range(n4):\n",
    "                pot= (joint_likelyhood[h]/ denominator)\n",
    "                self.predict_proba[i,h] =pot\n",
    "            \n",
    "        indices = np.argmax(self.predict_proba,axis=1)\n",
    "        res1=self.class_[indices]\n",
    "        return res1\n",
    "\n",
    "nb_train = Multi_NB(alpha=0)\n",
    "nb_train.fit(mat,res)\n",
    "sk_y_train = nb_train.predict(mat)\n",
    "\n",
    "\n",
    "\n",
    "print(\"Accuracy on train sample \",accuracy_score(sk_y_train,res))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d8a6ec92-2ba4-40bf-9b5d-01825338c034",
   "metadata": {},
   "source": [
    "## Vectorizing Test Sample\n",
    "\n",
    "- Use the Training Sample vocabulary to create word count matrix for test samples\n",
    "- This is also shown in the Helper code"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 100,
   "id": "a16cc470-b033-444f-a366-5b3213eba877",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(1393, 4000)\n"
     ]
    }
   ],
   "source": [
    "\"\"\"\n",
    "Your code here\n",
    "\"\"\"\n",
    "mat1=[]\n",
    "res1=[]\n",
    "for email1,lable1 in zip(X_test,y_test):\n",
    "    data1=[]\n",
    "    wrd1=email1.split(\" \")\n",
    "    for j in word_dict:\n",
    "        data1.append(wrd1.count(j[0]))\n",
    "    \n",
    "    mat1.append(data1)\n",
    "    res1.append(lable1)\n",
    "\n",
    "mat1=np.array(mat1)\n",
    "res1=np.array(res1)\n",
    "print(mat1.shape)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7c02b84a-3116-4260-846d-50e6849efe48",
   "metadata": {},
   "source": [
    "## Estimate Model's performance on Test Sample\n",
    "\n",
    "- Evaluate your model's performance on Test Sample, using the Training Priors ($\\alpha = 0$)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 101,
   "id": "cb763cb8-5a31-4d7d-b538-b741c8cfaae8",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\ACER\\AppData\\Local\\Temp\\ipykernel_8324\\1400903059.py:70: RuntimeWarning: invalid value encountered in double_scalars\n",
      "  pot= (joint_likelyhood[h]/ denominator)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy on test sample 0.95908111988514\n"
     ]
    }
   ],
   "source": [
    "\"\"\"\n",
    "Your code here\n",
    "\"\"\"\n",
    "from sklearn.metrics import accuracy_score\n",
    "nb = Multi_NB(alpha=0)\n",
    "nb.fit(mat,res)\n",
    "sk_y = nb.predict(mat1)\n",
    "\n",
    "\n",
    "\n",
    "print(\"Accuracy on test sample\",accuracy_score(sk_y,res1))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b415868e-4b61-4a13-abb5-22124e5312bc",
   "metadata": {},
   "source": [
    "## Select Smoothing Priors\n",
    "\n",
    "- Refactor your code to incorporate smoothing priors, select $\\alpha = 0$ for the previous estimates / sub-questions\n",
    "- Compare the performance with different values of $\\alpha \\gt 0$ as smoothing priors to take care of zero probabilities\n",
    "- You can display a Plot or Table to show the comparison."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 105,
   "id": "a707cc0d-1f16-40a5-93ec-0366165b077c",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\ACER\\AppData\\Local\\Temp\\ipykernel_8324\\1400903059.py:70: RuntimeWarning: invalid value encountered in double_scalars\n",
      "  pot= (joint_likelyhood[h]/ denominator)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "At 𝛼 =  0  Accuracy =  0.95908111988514\n",
      "At 𝛼 =  0.4  Accuracy =  0.9813352476669059\n",
      "At 𝛼 =  0.6  Accuracy =  0.9820531227566404\n",
      "At 𝛼 =  0.8  Accuracy =  0.9834888729361091\n",
      "At 𝛼 =  1.0  Accuracy =  0.9798994974874372\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "\"\"\"\n",
    "Your code here\n",
    "\"\"\"\n",
    "from sklearn.metrics import accuracy_score\n",
    "acc=[]\n",
    "for i in [0,0.4,0.6,0.8,1.0]:\n",
    "    nb = Multi_NB(alpha=i)\n",
    "    nb.fit(mat,res)\n",
    "    sk_y = nb.predict(mat1)\n",
    "    acrc=accuracy_score(sk_y,res1)\n",
    "    acc.append(acrc)\n",
    "    print(\"At 𝛼 = \",i,\" Accuracy = \",acrc)\n",
    "\n",
    "plt.plot([0.0,0.4,0.6,0.8,1.0],acc)\n",
    "plt.xlabel('Values of alpha') \n",
    "plt.ylabel('accuracy') \n",
    "plt.title('Smoothing Priors')\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 103,
   "id": "f0958ec9",
   "metadata": {},
   "outputs": [],
   "source": [
    "# At 𝛼=0.8 the Accuracy is maximum Accuracy=0.983"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d58051f8-67fb-4632-884b-23c719d4cae7",
   "metadata": {},
   "source": [
    "## Comparison with Sci-kit Learn Implementation\n",
    "\n",
    "- Use sci-kit learn's `sklearn.naive_bayes.MultinomialNB` model to compare your implementation's performance\n",
    "- (Optional) try other classifiers from `sklearn.naive_bayes` and see if you can make them work`"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "id": "192a26af-2b0a-4165-9d36-b5ed774c11fb",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy when using sklearn MultinomialNB\n",
      "At 𝛼 =  0  Accuracy =  0.9784637473079684\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\ProgramData\\Anaconda3\\lib\\site-packages\\sklearn\\naive_bayes.py:555: UserWarning: alpha too small will result in numeric errors, setting alpha = 1.0e-10\n",
      "  warnings.warn(\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "At 𝛼 =  0.4  Accuracy =  0.9813352476669059\n",
      "At 𝛼 =  0.6  Accuracy =  0.9820531227566404\n",
      "At 𝛼 =  0.8  Accuracy =  0.9834888729361091\n",
      "At 𝛼 =  1.0  Accuracy =  0.9798994974874372\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAkkAAAHFCAYAAADmGm0KAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8qNh9FAAAACXBIWXMAAA9hAAAPYQGoP6dpAABpiklEQVR4nO3deVwVhf7/8dfhsK8qICKQmiZq7qAo2lXLNHMtLVs019tqqXnLLM2sW3rtq61qi6JmpqbZbqW30twCQcwFl9wFwQUVEGQ98/uD5HdJ3BAYlvfz8eDxuAxz5rzPXOO8mZnzGYthGAYiIiIiUoid2QFEREREyiOVJBEREZEiqCSJiIiIFEElSURERKQIKkkiIiIiRVBJEhERESmCSpKIiIhIEVSSRERERIqgkiQiIiJSBJUkESEyMpJ77rmHm266CScnJ/z8/Gjfvj3jxo0zO9oVxcXF8corr3D48OFLfta5c2eaNm161W0cPnwYi8XCggULSj7gNT73xS87Ozu8vb25++672bx58zVtY8GCBVgsliL3gYjcGJUkkSru+++/Jzw8nNTUVKZPn87q1at555136NChA8uWLTM73hXFxcUxZcqUGyoI/v7+bN68mZ49e5ZcsOv09NNPs3nzZtavX8/UqVP5448/6NKlC7GxsVd9bM+ePdm8eTP+/v5lkFSkarE3O4CImGv69OnUq1ePn376CXv7//8r4YEHHmD69OkmJisbTk5OtGvXztQMN910U0GGDh060KBBA+644w5mz57Nxx9/XORjLly4gLOzM76+vvj6+pZYloyMDFxdXUtseyIVmY4kiVRxycnJ+Pj4FCpIF9nZFf4VUbduXXr16sV3331Hq1atcHFxoXHjxnz33XdA/qmfxo0b4+bmRtu2bYmOjr5km9988w3t27fH1dUVDw8P7rzzziJPLW3YsIE77rgDDw8PXF1dCQ8P5/vvvy/4+YIFC7jvvvsA6NKlS8Epq7+fNtuyZQu33XYbrq6u3HzzzUybNg2bzVbw86JOt73yyitYLBZ27drFgw8+iJeXF35+fgwfPpyUlJRC2z937hwjRoygRo0auLu707NnTw4ePIjFYuGVV14peqdfxcXCdOTIkYLXarFYWL16NcOHD8fX1xdXV1eysrIue7otIiKCFi1a4OzsTI0aNbjnnnvYvXt3oXWGDh2Ku7s7O3bsoFu3bnh4eHDHHXcAEBsbS69evahZsyZOTk7Url2bnj17Eh8fX6zXJFIRqSSJVHHt27cnMjKSZ555hsjISHJycq64/h9//MGECRMYP348K1euxMvLi3vvvZfJkyczd+5c3njjDRYvXkxKSgq9evXiwoULBY/97LPP6Nu3L56enixZsoR58+Zx9uxZOnfuzIYNGwrWW7duHbfffjspKSnMmzePJUuW4OHhQe/evQtOAfbs2ZM33ngDgFmzZrF58+ZLTpslJSXx8MMPM2jQIL755ht69OjBhAkT+PTTT69p3/Tv35+GDRvyxRdf8MILL/DZZ58xduzYgp/bbDZ69+7NZ599xvjx4/nyyy8JCwvjrrvuuqbtX87+/fsBLjlCNHz4cBwcHFi0aBErVqzAwcGhyMdPnTqVESNGcOutt7Jy5Ureeecdtm/fTvv27fnzzz8LrZudnU2fPn24/fbb+frrr5kyZQrp6enceeednDhxglmzZrFmzRrefvttbrrpJtLS0m7otYlUKIaIVGmnT582OnbsaAAGYDg4OBjh4eHG1KlTjbS0tELr1qlTx3BxcTHi4+MLlm3bts0ADH9/fyM9Pb1g+VdffWUAxjfffGMYhmHk5eUZtWvXNpo1a2bk5eUVrJeWlmbUrFnTCA8PL1jWrl07o2bNmoWePzc312jatKkRGBho2Gw2wzAMY/ny5QZg/Prrr5e8rk6dOhmAERkZWWh5kyZNjO7duxd8f+jQIQMw5s+fX7Bs8uTJBmBMnz690GOffPJJw9nZueD5v//+ewMw5syZU2i9qVOnGoAxefLkS3L9r4vP/Z///MfIyckxMjMzjZiYGKNNmzYGYHz//feGYRjG/PnzDcB45JFHLtnGxZ8dOnTIMAzDOHv2rOHi4mLcfffdhdY7evSo4eTkZDz00EMFy4YMGWIARkRERKF1o6OjDcD46quvrphfpLLTkSSRKs7b25v169ezZcsWpk2bRt++fdm3bx8TJkygWbNmnD59utD6LVu2JCAgoOD7xo0bA/mfJvvfa1kuLr94ymjv3r0cP36cwYMHFzqN5+7uTv/+/fn999/JyMggPT2dyMhIBgwYgLu7e8F6VquVwYMHEx8fz969e6/ptdWqVYu2bdsWWta8efOCTFfTp0+fSx6bmZnJyZMngfwjXgD3339/ofUefPDBa9r+RePHj8fBwQFnZ2dCQkI4evQoH374IXfffXeh9fr373/VbW3evJkLFy4wdOjQQsuDgoK4/fbb+fnnny95zN+326BBA6pXr8748eP54IMPiIuLu67XI1JZqCSJCAChoaGMHz+e5cuXc/z4ccaOHcvhw4cvuXi7Ro0ahb53dHS84vLMzEwg/9onoMhPYdWuXRubzcbZs2c5e/YshmFcdr3/3dbVeHt7X7LMycmp0CnA63m8k5MTQMHjk5OTsbe3v+S1+/n5XdP2Lxo9ejRbtmwhJiaGAwcOkJiYyKOPPnrJetfyCbar7ee/7ztXV1c8PT0LLfPy8mLdunW0bNmSF198kVtvvZXatWszefLkq56OFalMVJJE5BIODg5MnjwZgJ07d5bINi8WjsTExEt+dvz4cezs7KhevTrVq1fHzs7ususB+Pj4lEimG+Xt7U1ubi5nzpwptDwpKem6thMYGEhoaCitW7fm5ptvxmKxFLne5Zb/PRNcfj//fd9dbpvNmjVj6dKlJCcns23bNgYOHMirr77KjBkzrppBpLJQSRKp4op6MwUKPgl18ejNjQoODiYgIIDPPvsMwzAKlqenp/PFF18UfOLNzc2NsLAwVq5cWeiIj81m49NPPyUwMJCGDRsClx7ZKWudOnUCuGSe1NKlS82IA+RfiO/i4nLJxenx8fH88ssvBZ9eu1YWi4UWLVrw1ltvUa1aNbZu3VqScUXKNc1JEqniunfvTmBgIL1796ZRo0bYbDa2bdvGjBkzcHd3Z/To0SXyPHZ2dkyfPp2HH36YXr168dhjj5GVlcWbb77JuXPnmDZtWsG6U6dO5c4776RLly7861//wtHRkdmzZ7Nz506WLFlScPTj4kTtjz76CA8PD5ydnalXr16Rp9lKw1133UWHDh0YN24cqamphISEsHnzZj755JOC11zWqlWrxqRJk3jxxRd55JFHePDBB0lOTmbKlCk4OzsXHCG8ku+++47Zs2fTr18/br75ZgzDYOXKlZw7d44777yzDF6FSPmgkiRSxU2cOJGvv/6at956i8TERLKysvD396dr165MmDCh4ALskvDQQw/h5ubG1KlTGThwIFarlXbt2vHrr78SHh5esF6nTp345ZdfmDx5MkOHDsVms9GiRQu++eYbevXqVbBevXr1ePvtt3nnnXfo3LkzeXl5zJ8//5KLlkuLnZ0d3377LePGjWPatGlkZ2fToUMHPv30U9q1a0e1atXKJMffTZgwgZo1a/Luu++ybNkyXFxc6Ny5M2+88Qa33HLLVR9/yy23UK1aNaZPn87x48dxdHQkODiYBQsWMGTIkDJ4BSLlg8X43+PeIiJywz777DMefvhhNm7cWKj8iUjFopIkInIDlixZQkJCAs2aNcPOzo7ff/+dN998k1atWhWMCBCRikmn20REboCHhwdLly7l3//+N+np6fj7+zN06FD+/e9/mx1NRG6QjiSJiIiIFEEjAERERESKoJIkIiIiUgSVJBEREZEi6MLtYrLZbBw/fhwPD49rulWAiIiImM8wDNLS0qhdu/ZVB76qJBXT8ePHCQoKMjuGiIiIFMOxY8cIDAy84joqScXk4eEB5O/kv99BW0RERMqn1NRUgoKCCt7Hr0QlqZgunmLz9PRUSRIREalgruVSGV24LSIiIlIElSQRERGRIqgkiYiIiBRBJUlERESkCCpJIiIiIkVQSRIREREpgkqSiIiISBFUkkRERESKoJIkIiIiUgSVJBEREZEiqCSJiIiIFEElSURERKQIKkkiIiLXISM71+wIUkZUkkRERK6BYRhM/GoHTV7+iZ92JZkdR8qASpKIiMhVGIbBG6t28+nvRwGYt/6QyYmkLKgkiYiIXMX7v+zn47+KkcUCUYfPcPh0usmppLSpJImIiFzB/I2HmLFmHwAv92pCxwY+AKzcGm9mLCkDKkkiIiKXsSImninfxgEwtmtDhnesx32hQQB8sTUBm80wM56UMpUkERGRIvy4M5HnV/wBwIiO9XjmjgYAdGvih4ezPQnnLvD7wWQzI0opU0kSERH5m9/2neLpJbHYDBgYGsTEno2xWCwAODtY6d2iNpB/pEkqL5UkERGR/xF9+AyPLYohJ8+gZzN/3ri3WUFBumhASCAAq3YmkpaZY0ZMKQMqSSIiIn/ZdTyFYQu2cCEnj04NfXlrYEusdpZL1msVVI2bfd3IzLGxakeiCUmlLKgkiYiIAAdOneeReVGkZebStm4NPhgUgqN90W+TFouF+0LyL+DWKbfKSyVJRESqvIRzFxg8N5Lk9GyaBngyd2goLo7WKz7mnlYB2Flgy+GzmplUSakkiYhIlXYqLYtBcyM5npJJfV83Fg5ri6ezw1UfV8vLmdtu8QXgC81MqpRUkkREpMpKychh8LxIDp1OJ7C6C4tHtsPb3emaH3/xAu4vYuLJ08ykSkclSUREqqT0rFyGLohiT1Iavh5OfDoijFpezte1jTub+OHpbM/xlEw2H9DMpMpGJUlERKqczJw8Hl0UTezRc3i5OPDpiDDq+rhd93acHaz0aXlxZtKxko4pJlNJEhGRKiU3z8YzS2LZuD8ZN0crC4e3JbiWR7G3N+CvT7n9uCuJVM1MqlRUkkREpMqw2QyeX7Gd1XEncLS34+MhobQMqnZD22wR6EWDmu75M5O2a2ZSZaKSJCIiVYJhGLzy7S5WxiZgtbMw+6HWhNf3ueHtWiyWggu4l2tmUqWikiQiIlXCjNX7+GTzESwWmHl/C7o28SuxbV+cmRRz5CwHT50vse2KuVSSRESk0vtw3QHe/3U/AK/1bUrflgElun0/T2c6NdTMpMpGJUlERCq1zyKPMvWHPQCMv6sRg9rVKZXnuXgB98qtCZqZVEmoJImISKX19bYEXvpqBwBPdq7PE53rl9pz3dG4Jl4uDiSmZLLpwOlSex4pOypJIiJSKf28+wTjPv8Dw4DB7erwXPfgUn0+ZwcrfVrkz0xaHq1TbpWBSpKIiFQ6mw8k8+TireTaDO5pFcCUPrdisVhK/Xkvfsrtp11JpFzQzKSKTiVJREQqlT+OnWPkwi1k5dq4s4kf0wc0x86u9AsSQPNAL26p6U5Wro3vNTOpwlNJEhGRSmNvUhpD5keRnp1HeH1v3nuwFQ7Wsnurs1gs3BeafzRJtymp+FSSRESkUjiSnM7geZGcy8ihZVA1Pn4kFGcHa5nn6NcyAKudha1Hz3FAM5MqNJUkERGp8JJSMnl4biQn07JoVMuDBcPa4OZkb0qWmv8zM2mFJnBXaCpJIiJSoZ1Jz2bQvEjiz16grrcrn4xoSzVXR1MzXbyAe+XWeM1MqsBML0mzZ8+mXr16ODs7ExISwvr166+4/qxZs2jcuDEuLi4EBwfzySefXLLO22+/TXBwMC4uLgQFBTF27FgyMzMLfj5nzhyaN2+Op6cnnp6etG/fnh9++KHEX5uIiJSu1MwchkREsf/kefy9nPl0ZBg1PZzNjsUdjWtSzdWBE6lZbNivmUkVlakladmyZYwZM4aXXnqJ2NhYbrvtNnr06MHRo0eLXH/OnDlMmDCBV155hV27djFlyhSeeuopvv3224J1Fi9ezAsvvMDkyZPZvXs38+bNY9myZUyYMKFgncDAQKZNm0Z0dDTR0dHcfvvt9O3bl127dpX6axYRkZJxITuPkQui2ZGQgrebI4tGhBFY3dXsWAA42Vvp+9fMJJ1yq7gshmGYdhwwLCyM1q1bM2fOnIJljRs3pl+/fkydOvWS9cPDw+nQoQNvvvlmwbIxY8YQHR3Nhg0bABg1ahS7d+/m559/Llhn3LhxREVFXfEoVY0aNXjzzTcZMWLENWVPTU3Fy8uLlJQUPD09r+kxIiJSMrJzbTy6KJq1e0/h4WTPkkfb0TTAy+xYheyIT6H3+xtwtLdjy0td8XJxMDuScH3v36YdScrOziYmJoZu3boVWt6tWzc2bdpU5GOysrJwdi58GNXFxYWoqChycvKHdnXs2JGYmBiioqIAOHjwIKtWraJnz55FbjMvL4+lS5eSnp5O+/btb/RliYhIKcuzGYz9fBtr957C2cGOiGFtyl1BAmga4EmwnwfZuTa+237c7DhSDKaVpNOnT5OXl4efn1+h5X5+fiQlJRX5mO7duzN37lxiYmIwDIPo6GgiIiLIycnh9On8c74PPPAAr732Gh07dsTBwYH69evTpUsXXnjhhULb2rFjB+7u7jg5OfH444/z5Zdf0qRJk8vmzcrKIjU1tdCXiIiULcMweOnLHXy/PREHq4UPB4fSpm4Ns2MVyWKxFFzArduUVEymX7j99zHxhmFcdnT8pEmT6NGjB+3atcPBwYG+ffsydOhQAKzW/FkYa9eu5fXXX2f27Nls3bqVlStX8t133/Haa68V2lZwcDDbtm3j999/54knnmDIkCHExcVdNufUqVPx8vIq+AoKCrqBVy0iItfLMAxe/343S7ccw84C7z7QquCj9uVV31a1sdpZ2HbsHPtPppkdR66TaSXJx8cHq9V6yVGjkydPXnJ06SIXFxciIiLIyMjg8OHDHD16lLp16+Lh4YGPjw+QX6QGDx7MyJEjadasGffccw9vvPEGU6dOxWazFWzL0dGRBg0aEBoaytSpU2nRogXvvPPOZfNOmDCBlJSUgq9jxzRJVUSkLL33y37mbjgEwLT+zenRzN/kRFdX08OZLsEXZyYlmJxGrpdpJcnR0ZGQkBDWrFlTaPmaNWsIDw+/4mMdHBwIDAzEarWydOlSevXqhZ1d/kvJyMgo+N8XWa1WDMPgSteoG4ZBVlbWZX/u5ORUMDLg4peIiJSN+RsPMXPNPgBe7tWE+0MrztH8i6fcvozVzKSKxpxxpH959tlnGTx4MKGhobRv356PPvqIo0eP8vjjjwP5R28SEhIKZiHt27ePqKgowsLCOHv2LDNnzmTnzp0sXLiwYJu9e/dm5syZtGrVirCwMPbv38+kSZPo06dPwSm5F198kR49ehAUFERaWhpLly5l7dq1/Pjjj2W/E0RE5IqWRx9jyrf5l0OM7dqQ4R3rmZzo+tzeyI/qf81MWv/nKToH1zQ7klwjU0vSwIEDSU5O5tVXXyUxMZGmTZuyatUq6tSpA0BiYmKhmUl5eXnMmDGDvXv34uDgQJcuXdi0aRN169YtWGfixIlYLBYmTpxIQkICvr6+9O7dm9dff71gnRMnTjB48GASExPx8vKiefPm/Pjjj9x5551l9tpFROTqftyZyPgvtgMwomM9nrmjgcmJrp+jvR19WwawYNNhlsfEqyRVIKbOSarINCdJRKR0/bbvFCMWbiEnz2BgaBDT+je77Ad7yrudCSn0em8Djta/Zia5amaSWSrEnCQREZHLiT58hkcXRZOTZ9CzmT9v3FtxCxLArbU9aVTLg+w8G99oZlKFoZIkIiLlys6EFIYt2EJmjo1ODX15a2BLrHYVtyBB4ZlJuk1JxaGSJCIi5caBU+cZEhFFWmYubepW54NBITjaV463qn6tArC3s/DHsXP8eUIzkyqCyvEvT0REKrz4sxkMmhtJcno2TQM8mTe0DS6OVrNjlRgfd6eCi7Z1NKliUEkSERHTnUrLYtDcSBJTMqnv68bCYW3xdK58FzdfPOW2MjaB3DzbVdYWs6kkiYiIqVIychg8L5LDyRkEVHPh05FheLs7mR2rVNzeqCY13Bw5lZbF+j9Pmx1HrkIlSURETJOelcvQBVHsSUrD18OJxSPD8PdyMTtWqcmfmVQb0Cm3ikAlSURETJGZk8eji6KJPXoOLxcHFo1oS10fN7NjlbqLp9zWxJ3gXEa2yWnkSlSSRESkzOXm2Xh6SSwb9yfj5mhl4fC2NKpVNQbz3lrbi8b+nvkzk/7QzKTyTCVJRETKlM1m8PyK7ayJO4GjvR0fDwmlZVA1s2OVKc1MqhhUkkREpMwYhsEr3+5iZWwCVjsLsx9qTXh9H7Njlbl+LWtjb2dhe3wKe5M0M6m8UkkSEZEy83+r9/LJ5iNYLDDz/hZ0beJndiRTeLs7cXuj/JlJX2zV0aTySiVJRETKxAfrDjDr1wMAvNa3KX1bBpicyFwFM5O2amZSeaWSJCIipW5x5BGm/bAHgPF3NWJQuzomJzJfl0Y18XZz5PT5LNbtO2V2HCmCSpKIiJSqr7clMPGrnQA82bk+T3Sub3Ki8sHBaldwNE0XcJdPKkkiIlJqft59gnGf/4FhwOB2dXiue7DZkcqV+0LzT7n9d/cJzqZrZlJ5o5IkIiKlYvOBZJ5YvJVcm8E9rQKY0udWLBaL2bHKlcb+ntxa25OcPEMzk8ohlSQRESlx246dY+TCLWTn2riziR/TBzTHzk4FqSiamVR+qSSJiEiJ2puUxtD5UaRn5xFe35v3HmyFg1VvN5fTt2UADlYLOxJS2JOUanYc+R/6VysiIiXmSHI6g+ZFci4jh5ZB1fj4kVCcHaxmxyrXarg5FsxMWhGto0nliUqSiIiUiKSUTB6eG8mptCwa1fJgwbA2uDnZmx2rQrgvJAiAr7YlkKOZSeWGSpKIiNyw5PNZDJoXSfzZC9T1duWTEW2p5upodqwKo1OwLz7ujpw+n826vZqZVF6oJImIyA1JzcxhyPwo9p88j7+XM5+ODKOmh7PZsSoUB6sd/TQzqdxRSRIRkWK7kJ3HyAXR7ExIxdvNkUUjwgis7mp2rAqp/1+fcvt5zwnOaGZSuaCSJCIixZKda+OJxTFEHT6Dh5M9C4e3pUFNd7NjVViN/T1pGpA/M+nrbQlmxxFUkkREpBjybAZjl21j7d5TODvYETGsDU0DvMyOVeENaK2ZSeWJSpKIiFwXwzB4ceUOvt+RiIPVwoeDQ2lTt4bZsSqFizOTdh1PJe64ZiaZTSVJRESumWEYvP79bpZFH8POAu880IpODX3NjlVpVHdzpGtjPwC+2KqjSWZTSRIRkWv23i/7mbvhEADT+jfn7mb+JieqfC7epuSrWM1MMptKkoiIXJOIDYeYuWYfAC/3asL9oUEmJ6qc/tHQFx93J5LTs/l1z0mz41RpKkkiInJVy6OP8ep3cQCM6XoLwzvWMzlR5eVgteOeVrUBXcBtNpUkERG5oh92JDL+i+0AjOhYj9F33GJyospvwF+3Kfllz0mSz2eZnKbqUkkSEZHLWrfvFM8sjcVmwP2hgUzs2RiLxWJ2rEovuJYHzQO9yLUZfL3tuNlxqiyVJBERKVL04TM8tiianDyDns38mXpvcxWkMnTxAm6dcjOPSpKIiFxiZ0IKw+ZvITPHRqeGvrw1sCVWOxWkstS7eW0crXbEJaay63iK2XGqJJUkEREpZP/J8wyJiCItK5c2davzwaAQHO31dlHWqrs50rVJTUBHk8yif/UiIlIg/mwGg+dFkpyeTdMAT+YNbYOLo9XsWFXWfX9dwP31tuNk52pmUllTSRIREQBOpmUyaG4kiSmZ1Pd1Y+Gwtng6O5gdq0q77RYffD2cOJOeza97NTOprKkkiYgIKRk5PDIvisPJGQRUc+HTkWF4uzuZHavKs7facW+rAECn3MygkiQiUsWlZ+UydEEUe5LS8PVwYvHIMPy9XMyOJX/p/9en3H7dc5LTmplUplSSRESqsMycPB5dFE3s0XN4uTiwaERb6vq4mR1L/kdDPw9a/DUz6avYBLPjVCkqSSIiVVROno2nl8SycX8ybo5WFg5vS6NanmbHkiIM+Os+eSti4jEMw+Q0VYdKkohIFWSzGTy/Yjtr4k7gaG/Hx0NCaRlUzexYchl9/pqZtCcpjV3HU82OU2WoJImIVDGGYTD5m118GZuA1c7C7IdaE17fx+xYcgVerg7ceasfoAu4y5JKkohIFfN/q/ey6PcjWCww8/4WdG3iZ3YkuQYXb1Py9bYEzUwqIypJIiJVyAfrDjDr1wMAvNa3KX1bBpicSK7VbQ18qOnhxNmMHH7Zc8LsOFWCSpKISBWxOPII037YA8D4uxoxqF0dkxPJ9bC32nFva930tiypJImIVAFfb0tg4lc7AXiyc32e6Fzf5ERSHANC8o/8/br3FKfSNDOptKkkiYhUcv+NO8Gzn/+BYcDgdnV4rnuw2ZGkmBrU9KBlUDXybAZfb9PMpNKmkiQiUoltOnCaJz/bSp7N4J5WAUzpcysWi8XsWHIDLl7AvTxaM5NKm0qSiEglte3YOf65MJrsXBt3NvFj+oDm2NmpIFV0vZvXxtHejr0n0tiZoJlJpUklSUSkEtqblMaQiCjSs/MIr+/New+2wsGqX/mVgZerA91vrQXAiphjJqep3PRfjIhIJXMkOZ1B8yJJuZBDy6BqfPRIKM4OVrNjSQkqmJn0x3GycvNMTlN5qSSJiFQiiSkXeHhuJKfSsmhUy4MFw9rg7mRvdiwpYR0b+ODn6cS5jBx+2X3S7DiVlkqSiEglkXw+i0FzI4k/e4E63q58MqIt1VwdzY4lpcBqZ9HMpDKgkiQiUgmkZuYwZH4UB06l4+/lzKcjwqjp4Wx2LClFF0+5rd13ipNpmSanqZxML0mzZ8+mXr16ODs7ExISwvr166+4/qxZs2jcuDEuLi4EBwfzySefXLLO22+/TXBwMC4uLgQFBTF27FgyM///P6CpU6fSpk0bPDw8qFmzJv369WPv3r0l/tpEREqbzWbw3fbj9HlvAzsTUqnh5siiEWEE1XA1O5qUsvq+7rS+KX9m0lexmplUGkwtScuWLWPMmDG89NJLxMbGctttt9GjRw+OHj1a5Ppz5sxhwoQJvPLKK+zatYspU6bw1FNP8e233xass3jxYl544QUmT57M7t27mTdvHsuWLWPChAkF66xbt46nnnqK33//nTVr1pCbm0u3bt1IT08v9dcsIlISDMNg/Z+n6DtrI6M+i+VwcgY+7k58MrwtDWq6mx1PysiAkCAg/5SbZiaVPIth4l4NCwujdevWzJkzp2BZ48aN6devH1OnTr1k/fDwcDp06MCbb75ZsGzMmDFER0ezYcMGAEaNGsXu3bv5+eefC9YZN24cUVFRlz1KderUKWrWrMm6dev4xz/+cU3ZU1NT8fLyIiUlBU9Pz2t6jIhISdh27BzTf9zDpgPJALg5WvnnP25m5G036yLtKiblQg5tX/8vWbk2vhnVgeaB1cyOVO5dz/u3aUeSsrOziYmJoVu3boWWd+vWjU2bNhX5mKysLJydC59jd3FxISoqipycHAA6duxITEwMUVFRABw8eJBVq1bRs2fPy2ZJSUkBoEaNGpddJysri9TU1EJfIiJlaf/J8zy+KIZ+szay6UAyjlY7hneox2/Pd2FM14YqSFWQl8v/zkzSBdwlzbSSdPr0afLy8vDz8yu03M/Pj6SkpCIf0717d+bOnUtMTAyGYRAdHU1ERAQ5OTmcPn0agAceeIDXXnuNjh074uDgQP369enSpQsvvPBCkds0DINnn32Wjh070rRp08vmnTp1Kl5eXgVfQUFBxXzlIiLX5/i5C4xfsZ1ub63jx11JWCzQv3UgP4/rxMu9m+Dt7mR2RDFRwcykbcfJzNHMpJJk+p8df7+HkGEYl72v0KRJk0hKSqJdu3YYhoGfnx9Dhw5l+vTpWK35g9LWrl3L66+/zuzZswkLC2P//v2MHj0af39/Jk2adMk2R40axfbt2wtO113OhAkTePbZZwu+T01NVVESkVJ1Nj2b2Wv3s3DzEbJzbQDc2cSPf3ULJriWh8nppLzo0MAHfy9nElMy+Xn3SXo29zc7UqVhWkny8fHBarVectTo5MmTlxxdusjFxYWIiAg+/PBDTpw4gb+/Px999BEeHh74+PgA+UVq8ODBjBw5EoBmzZqRnp7Oo48+yksvvYSd3f8/ePb000/zzTff8NtvvxEYGHjFvE5OTjg56a81ESl9Gdm5RGw4xIfrDpKWlQtA27o1GN8jmJA6l78sQKqm/JlJAcz69QArYo6pJJUg0063OTo6EhISwpo1awotX7NmDeHh4Vd8rIODA4GBgVitVpYuXUqvXr0Kyk9GRkahIgRgtVoxDKPgyn/DMBg1ahQrV67kl19+oV69eiX4ykREiic718aizYf5x/S1/N/qfaRl5dLY35P5w9qw7LF2KkhyWf3/Giy5bt8pTqZqZlJJMfV027PPPsvgwYMJDQ2lffv2fPTRRxw9epTHH38cyD/FlZCQUDALad++fURFRREWFsbZs2eZOXMmO3fuZOHChQXb7N27NzNnzqRVq1YFp9smTZpEnz59Ck7JPfXUU3z22Wd8/fXXeHh4FBzN8vLywsXFpYz3gohUdTabwbfbjzNj9T6OnskA4KYarozr1pDezWtjZ1f0JQgiF93s605InerEHDnLl7EJPNapvtmRKgVTS9LAgQNJTk7m1VdfJTExkaZNm7Jq1Srq1KkDQGJiYqGZSXl5ecyYMYO9e/fi4OBAly5d2LRpE3Xr1i1YZ+LEiVgsFiZOnEhCQgK+vr707t2b119/vWCdiyMHOnfuXCjP/PnzGTp0aKm9XhGR/2UYBmv3nWL6j3vZnZj/iVkfdydG39GAgW1uwtHe9Hm/UoEMCAkk5shZVsTE8+g/br7s9b1y7Uydk1SRaU6SiNyImCNn+c+Pe4g6dAYADyd7Hut0M8M71sPV0fTP1EgFlJqZPzMpM8fGV091oGVQNbMjlUvX8/6t/xJFRMrQvhNpvPnTXtbEnQDA0d6OoeF1eaJTfaq76Wa0Unyezg7cdWstvtp2nBUxx1SSSoBKkohIGYg/m8Fba/5kZWw8hgF2FrgvJIjRXW+hdjVdCyklY0BIEF9tO843244zsWcTnB2sZkeq0FSSRERKUfL5LGb9eoBPfz9Cdl7+rKMeTWsxrluw7rEmJa59fW9qezlzPCWT/+4+Qa/mtc2OVKGpJImIlILzWbnMXX+Qj387SHp2/hTk8PrePH9XI50GkVKTPzMpkPd/3c+KmHiVpBukkiQiUoKycvP4LPIo7/+yn+T0bACaBngy/q5GdGzgo08cSanrH5Jfkn7bd4qklExqeTlf/UFSJJUkEZESkGcz+Co2gZlr9pFw7gIA9Xzc+Fe3YHo0raVZR1Jm6vm40aZudbYczp+Z9ERnzUwqLpUkEZEbYBgGP+8+yZs/7WXviTQA/DydGH1HQ+4LDcTBqllHUvYGhASy5fBZVsQc4/FOmplUXCpJIiLFFHXoDP/5cQ8xR84C4Olsz5NdGjCkfV1cHPWpIjHP3c38mfzNLg6cSmfbsXO0uqm62ZEqJJUkEZHrtDsxlek/7uHXvacAcHawY1iHejz+j/p4uTqYnE4EPJwd6NHUny9jE1gRE6+SVEwqSSIi1+hocgYz1+zl6z+OYxj5nyR6oE0Qz9xxC36eujhWypcBIYF8GZvAN38cZ1IvzUwqDpUkEZGrOJWWxfu//MlnUUfJycu/k1Ov5v6M6xZMPR83k9OJFK39zd4EVHMh4dwFVsedoE8LjQO4XipJIiKXkZqZw9zfDjJ3wyEy/pp1dNstPoy/qxFNA7xMTidyZXZ2Fvq3DuDdX/JnJqkkXT+VJBGRv8nMyePT348w69f9nM3IAaBFUDXGdw8mvIGPyelErl3/kEDe/WU/G/7UzKTiUEkSEflLbp6NlbEJvL1mH8dTMgGo7+vGc92D6X5rLX2MWiqcOt5utK1bg6jDZ1gZG8+TnRuYHalCUUkSkSrPMAx+2nWC/1u9l/0nzwPg7+XM2K4Nubd1APaadSQV2ICQQKIOn2FFdDxPdKqvsn8dVJJEpErbfCCZ//y4h23HzgFQzdWBpzo3YHD7Ovo0kFQKdzfPn5l08HQ6W4+eI6SOxgFcK5UkEamSdiakMP2nvfy2L3/WkYuDlZG31eOf/7gZT2fNOpLKw93Jnh7NarFya/7MJJWka6eSJCJVyuHT6fzf6r18tz0RAHs7Cw+F3cSo2xtQ00MXtUrlNCAkkJVbE/juj+NM7q2ZSddKJUlEqoQTqZm8+/OfLNtyjFybgcUCfVvU5tk7g7nJ29XseCKlql29/z8z6addSfRtGWB2pApBJUlEKrWUCzl8sO4A8zceIjPHBkCXYF+e696IJrU9TU4nUjbs7Cz54wB+/pMVMfEqSddIJUlEKqUL2Xks3HyYOWsPkHIhf9ZRSJ3qPN89mLCbvU1OJ1L2BrTOL0kb9p/m+LkL1K7mYnakck8lSUQqldw8G59Hx/POz/s4kZoFQEM/d57v3og7GtfUx5+lyrrJ25WwejWIPHSGL2MTeKqLZiZdjUqSiFQKNpvBDzuTmLF6LwdPpwMQUM2FZ+9sSL9WAVjtVI5EBoQEEnnoDCti4nmys2YmXY1KkohUeOv/PMX0H/eyIyEFAG83R0bd3oCHwm7CyV6f4hG56O5m+TOTDp1OZ+vRs4TUqWF2pHJNJUlEKqw/jp1j+k972Lg/GQA3Ryv//MfNjLztZtyd9OtN5O/cnOzp0dSfL7bG/zUzSSXpSvRbREQqnAOnzjNj9V5W7UgCwNFqx6B2dXiqS3283Z1MTidSvt0XGsgXW+P59o9EXu51Ky6OOtp6OSpJIlJhJKZc4J3//snymHjy/pp1dG+rQMZ0vYWgGpp1JHIt2tatQVANF46dyZ+Z1K+VxgFcjkqSiJR75zKymbP2AAs2HSYrN3/W0Z1N/PhXt2CCa3mYnE6kYrGzs9C/dSBv/zd/ZpJK0uWpJIlIuZWRncv8jYf5YN0B0jJzgfy/gsf3CNa1FCI34GJJ2njgNAnnLhCgmUlFUkkSkXInJ8/G0i3HePfnPzmVlj/rqLG/J8/fFUznhr762LLIDQqq4Uq7m2vw+8EzfLk1nlG332J2pHJJJUlEyg2bzeDb7ceZuWYfR5IzALiphivjujWkd/Pa2GnWkUiJGRASxO8H82cmPdWlgf74KIJKkoiYzjAM1u3Ln3UUl5gKgI+7E6PvaMDANjfhaG9nckKRyufuZrWY/PVODidnEH3kLG3q6hT236kkiYipYo6cZfqPe4g8dAYADyd7Hut0M8M61MNNs45ESo2roz13N/NneUw8K6LjVZKKUKw/z9auXVvCMUSkqtl3Io1/fhJN/zmbiDx0Bkd7O/55Wz1+e74Lo26/RQVJpAwMCAkE4PsdiWRk55qcpvwp1m+hu+66i4CAAIYNG8aQIUMICgoq6VwiUknFn83g7f/+ycqt8dgMsLPAfSFBjO56i+5KLlLG2tStwU01XDl6JoOfdiVxT6tAsyOVK8U6knT8+HFGjx7NypUrqVevHt27d+fzzz8nOzu7pPOJSCWRfD6LV7+N4/b/W8eKmPyCdNettVg99h/8Z0BzFSQRE1ycmQSwPDre5DTlj8UwDONGNrBt2zYiIiJYsmQJNpuNhx9+mBEjRtCiRYuSylgupaam4uXlRUpKCp6enmbHESm3zmflMm/9IT5ef5DzWfmH89vf7M34Ho1oGVTN3HAiQvzZDDr+51cANozvQmD1yj29/nrev2/4IyMtW7bkhRde4KmnniI9PZ2IiAhCQkK47bbb2LVr141uXkQqsKSUTO5+Zz1v/Xcf57NyaRrgyaIRbfnsn2EqSCLlRGB1V8LrewOwcmuCyWnKl2KXpJycHFasWMHdd99NnTp1+Omnn3j//fc5ceIEhw4dIigoiPvuu68ks4pIBZJ8PotB8yI5eiaDgGouvP9QK755qiO33aJhkCLlzcULuFfExHODJ5gqlWKVpKeffhp/f38ef/xxGjZsSGxsLJs3b2bkyJG4ubkRFBTEtGnT2LNnT0nnFZEKIDUzhyHzo9h/8jz+Xs4se6wdvTQMUqTcuqtpLdwcrRw9k8GWw2fNjlNuFOvTbXFxcbz33nv0798fR0fHItepXbs2v/766w2FE5GK50J2HiMXRLMzIRVvN0cWjQir9Nc4iFR0ro729Gzuz+fR8SyPPkbbepqZBMU8kvTzzz/z4IMPXrYgAdjb29OpU6diBxORiic718YTi2OIOnwGDyd7Fg5vS4Oa7mbHEpFrcF9o/jif73ckkp6lmUlQzJI0depUIiIiLlkeERHBf/7znxsOJSIVT57NYOyybazdewpnBzsihrWhaYCX2bFE5BqF1qlOHW9XMrLz+HFnktlxyoVilaQPP/yQRo0aXbL81ltv5YMPPrjhUCJSsRiGwYsrd/D9jkQcrBY+HByqWxyIVDAWi4UBrf//BdxSzJKUlJSEv7//Jct9fX1JTEy84VAiUnEYhsHr3+9mWfQx7Czw7gOt6NTQ1+xYIlIM94YEYrHA5oPJHDuTYXYc0xWrJAUFBbFx48ZLlm/cuJHatWvfcCgRqTje+2U/czccAmBa/+b0aHbpH1AiUjEEVHMpmJn0xVYdTSpWSRo5ciRjxoxh/vz5HDlyhCNHjhAREcHYsWP55z//WdIZRaScithwiJlr9gHwcq8m3B+q+ziKVHT3heT/d/zF1nhstqo9M6lYIwCef/55zpw5w5NPPllwvzZnZ2fGjx/PhAkTSjSgiJRPy6OP8ep3cQCM7dqQ4R3rmZxIREpC91tr4e5kz7EzF4g6fIZ2N3ubHck0xTqSZLFY+M9//sOpU6f4/fff+eOPPzhz5gwvv/xySecTkXLohx2JjP9iOwAjOtbjmTsamJxIREqKi6OVXs3zT5tX9Qu4b+jebe7u7rRp04amTZvi5ORUUplEpBxbt+8UzyyNxWbAwNAgJvZsrNuMiFQyF29TsqqKz0wq1uk2gC1btrB8+XKOHj1acMrtopUrV95wMBEpf6IPn+GxRdHk5Bn0bObPG/c2U0ESqYRC6lSnno8bh06n88POpILSVNUU60jS0qVL6dChA3FxcXz55Zfk5OQQFxfHL7/8gpeXhseJVEY7E1IYNn8LmTk2OjX05a2BLbHqXmwilZLFYikoRsujj5mcxjzFKklvvPEGb731Ft999x2Ojo6888477N69m/vvv5+bbrqppDOKiMn2nzzPkIgo0rJyaVu3Bh8MCsHR/obO1otIOXdPqwAsFog8dIajyVVzZlKxfssdOHCAnj17AuDk5ER6ejoWi4WxY8fy0UcflWhAETFX/NkMBs+LJDk9m6YBnswdGoqLo9XsWCJSympXc6FjAx+g6s5MKlZJqlGjBmlpaQAEBASwc+dOAM6dO0dGRtVsmyKV0am0LAbNjSQxJZP6vm4sHNYWT2cHs2OJSBm5eMqtqs5MKlZJuu2221izZg0A999/P6NHj+af//wnDz74IHfccUeJBhQRc6Rk5DB4XiSHkzMIrO7C4pHt8HbXp1hFqpJuTWrh4WRP/NkLRB46Y3acMleskvT+++/zwAMPADBhwgT+9a9/ceLECe69917mzZt3XduaPXs29erVw9nZmZCQENavX3/F9WfNmkXjxo1xcXEhODiYTz755JJ13n77bYKDg3FxcSEoKIixY8eSmZlZ8PPffvuN3r17U7t2bSwWC1999dV1ZRap7NKzchm6IIo9SWn4ejjx6Ygwank5mx1LRMqYi6OVXi3ybze2PKbqXcB93SUpNzeXb7/9Fju7/Ifa2dnx/PPP88033zBz5kyqV69+zdtatmwZY8aM4aWXXiI2NpbbbruNHj16cPTo0SLXnzNnDhMmTOCVV15h165dTJkyhaeeeopvv/22YJ3FixfzwgsvMHnyZHbv3s28efNYtmxZoUng6enptGjRgvfff/96X75IpZeZk8eji6KJPXoOLxcHPh0RRl0fN7NjiYhJLp5y+2FHEuer2Mwki2EY132S0dXVld27d1OnTp0bevKwsDBat27NnDlzCpY1btyYfv36MXXq1EvWDw8Pp0OHDrz55psFy8aMGUN0dDQbNmwAYNSoUezevZuff/65YJ1x48YRFRVV5FEqi8XCl19+Sb9+/a4re2pqKl5eXqSkpODp6XldjxUpr3LzbDyxeCtr4k7g5mhl8T/b0TKomtmxRMREhmFwx4x1HDydzvQBzSv8PRqv5/27WKfbwsLCiI2NLVa4i7Kzs4mJiaFbt26Flnfr1o1NmzYV+ZisrCycnQsf8ndxcSEqKoqcnBwAOnbsSExMDFFRUQAcPHiQVatWFXwar7iysrJITU0t9CVSmdhsBs+v2M6auBM42tvx8ZBQFSQRwWKx0P+vo0lV7TYlxZq4/eSTTzJu3Dji4+MJCQnBza3wofjmzZtfdRunT58mLy8PPz+/Qsv9/PxISkoq8jHdu3dn7ty59OvXj9atWxMTE0NERAQ5OTmcPn0af39/HnjgAU6dOkXHjh0xDIPc3FyeeOIJXnjhheK81AJTp05lypQpN7QNkfLKMAxe+XYXK2MTsNpZmP1Qa8Lr+5gdS0TKiXtbBzBj9V6iDp3hSHI6dbyrxin4YpWkgQMHAvDMM88ULLNYLBiGgcViIS8v75q39fdbGlzcRlEmTZpEUlIS7dq1wzAM/Pz8GDp0KNOnT8dqzZ/bsnbtWl5//XVmz55NWFgY+/fvZ/To0fj7+zNp0qTrfakFJkyYwLPPPlvwfWpqKkFBFfuQo8hF/7d6L59sPoLFAjPvb0HXJn5Xf5CIVBn+Xi50vMWX3/ad4ouYeJ7tFmx2pDJRrJJ06NChG35iHx8frFbrJUeNTp48ecnRpYtcXFyIiIjgww8/5MSJE/j7+/PRRx/h4eGBj0/+X72TJk1i8ODBjBw5EoBmzZqRnp7Oo48+yksvvVRwwfn1cnJy0k18pVL6YN0BZv16AIDX+jalb8sAkxOJSHk0ICQwvyRtTWBM14bYVYHbEhWrJN3oBdsAjo6OhISEsGbNGu65556C5WvWrKFv375XfKyDgwOBgfnnR5cuXUqvXr0Kyk9GRsYlRchqtWIYBsW4Rl2kUlsceYRpP+wBYPxdjRjU7sb/2xaRyqlbEz88nO1JOHeB3w8mE96g8p+SL1ZJKmo20f965JFHrmk7zz77LIMHDyY0NJT27dvz0UcfcfToUR5//HEg/xRXQkJCwfPt27ePqKgowsLCOHv2LDNnzmTnzp0sXLiwYJu9e/dm5syZtGrVquB026RJk+jTp0/BKbnz58+zf//+gsccOnSIbdu2UaNGDd17TqqMr7clMPGr/Gn5T3auzxOd65ucSETKM2cHK71b1OazyKOsiIlXSbqc0aNHF/o+JyeHjIwMHB0dcXV1veaSNHDgQJKTk3n11VdJTEykadOmrFq1quBIVWJiYqGZSXl5ecyYMYO9e/fi4OBAly5d2LRpE3Xr1i1YZ+LEiVgsFiZOnEhCQgK+vr707t2b119/vWCd6OhounTpUvD9xWuNhgwZwoIFC653d4hUOD/vPsG4z//AMGBwuzo8171qXF8gIjdmQEggn0UeZdXORKb0vRWPSn6bomLNSSrKn3/+yRNPPMFzzz1H9+7dS2KT5ZrmJElFtflAMkPmR5Gda+OeVgHMuK9Flbi2QERunGEYdJ25jgOn0vlP/2YMbFPxzr6U+pykotxyyy1MmzbtkqNMIlJ+bDt2jpELt5Cda+POJn5MH9BcBUlErpnFYmFASP4nu6vCzKQSK0mQf4H08ePHS3KTIlJC9ialMXR+FOnZeYTX9+a9B1vhYC3RXwEiUgXc0yoAOwtsOXyWw6fTzY5Tqop1TdI333xT6HvDMEhMTOT999+nQ4cOJRJMRErOkeR0Bs2L5FxGDi2DqvHxI6E4O1jNjiUiFVAtL2duu8WXdftO8cXWeMZV4plJxSpJf7/PmcViwdfXl9tvv50ZM2aURC4RKSFJKZk8PDeSU2lZNKrlwYJhbXBzKtZ/+iIiQP4F3Ov+Giw5thLPTCrWb0qbzVbSOUSkFCSfz2LQvEjiz16grrcrn4xoSzVXR7NjiUgFd2cTPzyd7Tmeksnmg8l0qKTjAHRBgkgllZqZw5D5Uew/eR5/L2c+HRlGTQ/nqz9QROQqnB2s9GlZG4Dl0cdMTlN6ilWSBgwYwLRp0y5Z/uabb3LffffdcCgRuTEXsvMYuSCanQmpeLs5smhEGIHVXc2OJSKVyMVPuf24K4nUzByT05SOYpWkdevW0bNnz0uW33XXXfz22283HEpEii8718YTi2OIOnwGDyd7Fg5vS4Oa7mbHEpFKpkWgFw1qupOZY2PV9kSz45SKYpWk8+fP4+h46XUNDg4OpKam3nAoESmePJvB2GXbWLv3FM4OdkQMa0PTAC+zY4lIJZQ/Myn/PqqVdWZSsUpS06ZNWbZs2SXLly5dSpMmTW44lIhcP8MweHHlDr7fkYiD1cKHg0NpU7eG2bFEpBK7ODMp+shZDlXCmUnF+nTbpEmT6N+/PwcOHOD2228H4Oeff2bJkiUsX768RAOKyNUZhsHr3+9mWfQx7Czw7gOt6NTQ1+xYIlLJ+Xk606mhL7/uPcWKmGM8172R2ZFKVLGOJPXp04evvvqK/fv38+STTzJu3Dji4+P573//e8kMJREpfe/9sp+5Gw4BMK1/c3o08zc5kYhUFRcv4F65NYE8W4ncDrbcKPZEuZ49exZ58baIlK2IDYeYuWYfAC/3asL9oUEmJxKRquSOxjXxcnEgMSWTTQdOc9stlecodrGOJG3ZsoXIyMhLlkdGRhIdHX3DoUTk2iyPPsar38UBMLZrQ4Z3rGdyIhGpapwdrPRpkT8zqbJdwF2skvTUU09x7Nilw6MSEhJ46qmnbjiUiFzdDzsSGf/FdgBGdKzHM3c0MDmRiFRVFz/l9uPOyjUzqVglKS4ujtatW1+yvFWrVsTFxd1wKBG5snX7TvHM0lhsBgwMDWJiz8ZYLJXz3kkiUv41D/SioZ87Wbk2vvuj8sxMKlZJcnJy4sSJE5csT0xMxN5eN84UKU3Rh8/w2KJocvIMejbz5417m6kgiYipCs9Mqjy3KSlWSbrzzjuZMGECKSkpBcvOnTvHiy++yJ133lli4USksJ0JKQxbsIXMHBudGvry1sCWWCvp3bdFpGLp1zIAq52FrUfPceDUebPjlIhilaQZM2Zw7Ngx6tSpQ5cuXejSpQv16tUjKSmJGTNmlHRGEQEOnDrPkIgo0jJzaVu3Bh8MCsHRXveoFpHyoeZfM5MAvqgkF3AX6zdsQEAA27dvZ/r06TRp0oSQkBDeeecdduzYQVCQPn4sUtLiz2YwaG4kyenZNA3wZO7QUFwcrWbHEhEp5OIpt8oyM6nYFxC5ubnRsWNHbrrpJrKzswH44YcfgPxhkyJSMk6lZTFobiSJKZnU93Vj4bC2eDo7mB1LROQSdzSuSTVXB5JSM9mw/3SFn/xfrJJ08OBB7rnnHnbs2IHFYsEwjEIXjubl5ZVYQJGqLCUjh8HzIjmcnEFANRc+HRmGt7uT2bFERIrkZG+lb4vaLNx8hBUx8RW+JBXrdNvo0aOpV68eJ06cwNXVlZ07d7Ju3TpCQ0NZu3ZtCUcUqZrSs3IZuiCKPUlp+Ho4sXhkGP5eLmbHEhG5oou3KflpVxIpFyr2zKRilaTNmzfz6quv4uvri52dHVarlY4dOzJ16lSeeeaZks4oUuVk5uTx6KJoYo+ew8vFgUUj2lLXx83sWCIiV9U0wJNgPw+yc218t/242XFuSLFKUl5eHu7u7gD4+Phw/Hj+TqhTpw579+4tuXQiVVBuno2nl8SycX8ybo5WFg5vS6NanmbHEhG5JoVnJlXsT7kVqyQ1bdqU7dvzb4cQFhbG9OnT2bhxI6+++io333xziQYUqUpsNoPnV2xnTdwJHO3t+HhIKC2DqpkdS0TkuvRrlT8zKfboOfafTDM7TrEVqyRNnDgRm80GwL///W+OHDnCbbfdxqpVq3j33XdLNKBIVWEYBq98u4uVsQlY7SzMfqg14fV9zI4lInLdfD2c6BKcf9H2ipgEk9MUn8UwjBIZZHDmzBmqV69eZW6PkJqaipeXFykpKXh66lSI3Lg3f9rDrF8PYLHA2wNb0rdlgNmRRESK7cediTz+6Vb8PJ3Y9MId5ebuANfz/l1i43pr1KhRZQqSSEn7YN0BZv16AIDX+jZVQRKRCu/2Rn5Ud3XgRGoW6/88ZXacYtE9DURMtjjyCNN+2APA+LsaMahdHZMTiYjcOEd7u4I/+CrqBdwqSSIm+npbAhO/2gnAk53r80Tn+iYnEhEpORc/5bY67gQpGRVvZpJKkohJft59gnGf/4FhwOB2dXiue7DZkUREStSttT1pVCt/ZtI3FXBmkkqSiAk2H0jmicVbybUZ3NMqgCl9btU1fSJS6VT0mUkqSSJlbNuxc4xcuIXsXBt3NvFj+oDm2JWTT32IiJS0fq0CsLez8Mexc/x5omLNTFJJEilDe5PSGDo/ivTsPMLre/Peg61wsOo/QxGpvHzcnegcXBOAFVsr1tEk/XYWKSNHktMZNC+Scxk5tAyqxsePhOLsYDU7lohIqbsvNP+U25dbE8jNs5mc5tqpJImUgaSUTB6eG8mptCwa1fJgwbA2uDnZmx1LRKRMdAmuSQ03R06mZbH+z9Nmx7lmKkkipSz5fBaD5kUSf/YCdb1d+WREW6q5OpodS0SkzOTPTKoNVKwLuFWSREpRamYOQ+ZHsf/kefy9nPl0ZBg1PZzNjiUiUuYufsptTdwJzmVkm5zm2qgkiZSSC9l5jFwQzc6EVLzdHFk0IozA6q5mxxIRMcWttb1o7O9Jdp6Nb/+oGDOTVJJESkF2ro0nFscQdfgMHk72LBzelgY13c2OJSJiqoo2M0klSaSE5dkMxi7bxtq9p3B2sCNiWBuaBniZHUtExHT9WtbOn5kUn8K+CjAzSSVJpAQZhsGLK3fw/Y5EHKwWPhwcSpu6NcyOJSJSLni7O3F7o79mJlWAo0kqSSIlxDAMXv9+N8uij2FngXcfaEWnhr5mxxIRKVcunnJbWQFmJqkkiZSQ937Zz9wNhwCY1r85PZr5m5xIRKT86dKoJt5ujpw+n8Vvf54yO84VqSSJlID5Gw8xc80+AF7u1YT7Q4NMTiQiUj45WO3o2zIAKP+n3FSSRG7Q8uhjTPk2DoCxXRsyvGM9kxOJiJRvF29T8t+4k5xNL78zk1SSRG7AjzsTGf/FdgBGdKzHM3c0MDmRiEj519jfk1tr589M+qYcz0xSSRIppt/2neLpJbHYDBgYGsTEno2xWCxmxxIRqRAqwswklSSRYog+fIZHF0WTk2fQs5k/b9zbTAVJROQ69G0ZgIPVwo6EFPYkpZodp0gqSSLXaWdCCsMWbCEzx0anhr68NbAlVjsVJBGR61HDzbFgZtIX5fRokkqSyHU4cOo8QyKiSMvMpU3d6nwwKARHe/1nJCJSHPeF5H8S+MvYBHLK4cwk/XYXuUbxZzMYNDeS5PRsmgZ4Mm9oG1wcrWbHEhGpsDoF++Lj7sjp89ms21v+ZiapJIlcg1NpWQyaG0liSib1fd1YOKwtns4OZscSEanQHKx29CvHM5NUkkSuIiUjh8HzIjmcnEFANRc+HRmGt7uT2bFERCqF/n99yu3nPSc4U85mJqkkiVxBelYuQxdEsScpDV8PJxaPDMPfy8XsWCIilUZjf0+aBniSk2fwzbYEs+MUYnpJmj17NvXq1cPZ2ZmQkBDWr19/xfVnzZpF48aNcXFxITg4mE8++eSSdd5++22Cg4NxcXEhKCiIsWPHkpmZeUPPK1VPZk4ejy6KJvboObxcHFg0oi11fdzMjiUiUulcvIB7xdbydcrN1JK0bNkyxowZw0svvURsbCy33XYbPXr04OjRo0WuP2fOHCZMmMArr7zCrl27mDJlCk899RTffvttwTqLFy/mhRdeYPLkyezevZt58+axbNkyJkyYUOznlaonN8/G00ti2bg/GTdHKwuHt6VRLU+zY4mIVEp9WtTGwWphZ0IquxPLz8wki2EYhllPHhYWRuvWrZkzZ07BssaNG9OvXz+mTp16yfrh4eF06NCBN998s2DZmDFjiI6OZsOGDQCMGjWK3bt38/PPPxesM27cOKKiogqOFl3v8xYlNTUVLy8vUlJS8PTUm2dlYrMZ/Gv5H6yMTcDR3o4Fw9oQXt/H7FgiIpXaE5/G8MPOJEZ0rMekXk1K7Xmu5/3btCNJ2dnZxMTE0K1bt0LLu3XrxqZNm4p8TFZWFs7OzoWWubi4EBUVRU5ODgAdO3YkJiaGqKgoAA4ePMiqVavo2bNnsZ/34nOnpqYW+pLKxzAMXvl2FytjE7DaWZj9UGsVJBGRMnDxNiVflaOZSaaVpNOnT5OXl4efn1+h5X5+fiQlJRX5mO7duzN37lxiYmIwDIPo6GgiIiLIycnh9OnTADzwwAO89tprdOzYEQcHB+rXr0+XLl144YUXiv28AFOnTsXLy6vgKygo6EZevpRT/7d6L59sPoLFAjPvb0HXJn5Xf5CIiNywTg198XF3Ijk9m7XlZGaS6Rdu//1+V4ZhXPYeWJMmTaJHjx60a9cOBwcH+vbty9ChQwGwWvOH+q1du5bXX3+d2bNns3XrVlauXMl3333Ha6+9VuznBZgwYQIpKSkFX8eOHbvelyrl3AfrDjDr1wMAvNa3KX3/mt0hIiKlz95qx72tL85MKh/vsaaVJB8fH6xW6yVHb06ePHnJUZ6LXFxciIiIICMjg8OHD3P06FHq1q2Lh4cHPj75p0QmTZrE4MGDGTlyJM2aNeOee+7hjTfeYOrUqdhstmI9L4CTkxOenp6FvqTyWBx5hGk/7AFg/F2NGNSujsmJRESqnv6t/5qZtPskyeezTE5jYklydHQkJCSENWvWFFq+Zs0awsPDr/hYBwcHAgMDsVqtLF26lF69emFnl/9SMjIyCv73RVarFcMwMAzjhp5XKqevtyUw8audADzZuT5PdK5vciIRkaopuJYHzQO9yLUZfL3tuNlxsDfzyZ999lkGDx5MaGgo7du356OPPuLo0aM8/vjjQP4proSEhIJZSPv27SMqKoqwsDDOnj3LzJkz2blzJwsXLizYZu/evZk5cyatWrUiLCyM/fv3M2nSJPr06VNwSu5qzytVx8+7TzDu8z8wDBjcrg7PdQ82O5KISJU2ICSQ7fEprIiJZ3jHeqZmMbUkDRw4kOTkZF599VUSExNp2rQpq1atok6d/FMdiYmJhWYX5eXlMWPGDPbu3YuDgwNdunRh06ZN1K1bt2CdiRMnYrFYmDhxIgkJCfj6+tK7d29ef/31a35eqRo2H0jmicVbybUZ3NMqgCl9br3idWkiIlL6+rSozb+/201cYiq7jqdwa20v07KYOiepItOcpIpt27FzPPzx76Rn53FnEz9mP9waB6vpn2MQERHgqcVb+X5HIsM71OPl3iU7M6lCzEkSMcvepDSGzo8iPTuP8PrevPdgKxUkEZFyZEBIIFY7C6mZOabmMPV0m0hZO5KczqB5kZzLyKFlUDU+fiQUZwer2bFEROR/3HaLD5sn3E5ND+err1yK9OezVBlJKZk8PDeSU2lZNKrlwYJhbXBz0t8JIiLljb3VzvSCBCpJUkUkn89i0LxI4s9eoK63K5+MaEs1V0ezY4mISDmmkiSVXmpmDkPmR7H/5Hn8vZz5dGRYufgLRUREyjeVJKnULmTnMXJBNDsTUvF2c2TRiDACq7uaHUtERCoAlSSptLJzbTyxOIaow2fwcLJn4fC2NKjpbnYsERGpIFSSpFLKsxmMXbaNtXtP4exgR8SwNjQNMG8gmYiIVDwqSVLpGIbBiyt38P2ORBysFj4cHEqbujXMjiUiIhWMSpJUKoZh8Pr3u1kWfQw7C7z7QCs6NfQ1O5aIiFRAKklSqbz3y37mbjgEwLT+zenRzN/kRCIiUlGpJEmlMX/jIWau2QfAy72acH9okMmJRESkIlNJkkphefQxpnwbB8DYrg0Z3rGeyYlERKSiU0mSCu/HnYmM/2I7ACM61uOZOxqYnEhERCoDlSSp0H7bd4qnl8RiM+D+0EAm9myMxWIxO5aIiFQCKklSYUUfPsOji6LJyTPo2cyfqfc2V0ESEZESo5IkFdLOhBSGLdhCZo6NTg19eWtgS6x2KkgiIlJyVJKkwjlw6jxDIqJIy8ylTd3qfDAoBEd7/VMWEZGSpXcWqVDiz2YwaG4kyenZNA3wZN7QNrg4Ws2OJSIilZBKklQYp9KyGDQ3ksSUTOr7urFwWFs8nR3MjiUiIpWUSpJUCCkZOQyeF8nh5AwCqrnw6cgwvN2dzI4lIiKVmEqSlHvpWbkMXRDFnqQ0fD2cWDwyDH8vF7NjiYhIJaeSJOVaZk4ejy6KJvboObxcHFg0oi11fdzMjiUiIlWASpKUW7l5Np5eEsvG/cm4OVpZOLwtjWp5mh1LRESqCJUkKZdsNoPnV2xnTdwJHO3t+HhIKC2DqpkdS0REqhCVJCl3DMPglW93sTI2AaudhdkPtSa8vo/ZsUREpIpRSZJy5/9W7+WTzUewWGDm/S3o2sTP7EgiIlIFqSRJufLBugPM+vUAAK/1bUrflgEmJxIRkapKJUnKjcWRR5j2wx4Axt/ViEHt6picSEREqjKVJCkXvt6WwMSvdgLwZOf6PNG5vsmJRESkqlNJEtP9vPsE4z7/A8OAwe3q8Fz3YLMjiYiIqCSJuTYfSOaJxVvJtRnc0yqAKX1uxWKxmB1LREREJUnMs+3YOUYu3EJ2ro07m/gxfUBz7OxUkEREpHxQSRJT7E1KY+j8KNKz8wiv7817D7bCwap/jiIiUn7oXUnK3JHkdAbNi+RcRg4tg6rx8SOhODtYzY4lIiJSiEqSlKmklEwenhvJqbQsGtXyYMGwNrg52ZsdS0RE5BIqSVJmks9nMWheJPFnL1DX25VPRrSlmquj2bFERESKpJIkZSI1M4ch86PYf/I8/l7OfDoyjJoezmbHEhERuSyVJCl1F7LzGLkgmp0JqXi7ObJoRBiB1V3NjiUiInJFKklSqrJzbTyxOIaow2fwcLJn4fC2NKjpbnYsERGRq1JJklKTZzMY+/k21u49hbODHRHD2tA0wMvsWCIiItdEJUlKhWEYvPTlDr7fnoiD1cKHg0NpU7eG2bFERESumUqSlDjDMHj9+90s3XIMOwu8+0ArOjX0NTuWiIjIdVFJkhL33i/7mbvhEADT+jenRzN/kxOJiIhcP5UkKVHzNx5i5pp9ALzcqwn3hwaZnEhERKR4VJKkxCyPPsaUb+MAGNP1FoZ3rGdyIhERkeJTSZIS8ePORMZ/sR2AER3rMfqOW0xOJCIicmNUkuSG/bbvFE8vicVmwP2hgUzs2RiLxWJ2LBERkRuikiQ3JPrwGR5dFE1OnkHPZv5Mvbe5CpKIiFQKKklSbDsTUhi2YAuZOTY6NfTlrYEtsdqpIImISOWgkiTFcuDUeYZERJGWmUubutX5YFAIjvb65yQiIpWH3tXkusWfzWDQ3EiS07NpGuDJvKFtcHG0mh1LRESkRKkkyXU5lZbFoLmRJKZkUt/XjYXD2uLp7GB2LBERkRKnkiTXLCUjh8HzIjmcnEFANRc+HRmGt7uT2bFERERKhUqSXJP0rFyGLohiT1Iavh5OLB4Zhr+Xi9mxRERESo1KklxVZk4ejy6KJvboObxcHFg0oi11fdzMjiUiIlKqTC9Js2fPpl69ejg7OxMSEsL69euvuP6sWbNo3LgxLi4uBAcH88knnxT6eefOnbFYLJd89ezZs2CdtLQ0xowZQ506dXBxcSE8PJwtW7aUyuur6HLzbDy9JJaN+5Nxc7SycHhbGtXyNDuWiIhIqbM388mXLVvGmDFjmD17Nh06dODDDz+kR48exMXFcdNNN12y/pw5c5gwYQIff/wxbdq0ISoqin/+859Ur16d3r17A7By5Uqys7MLHpOcnEyLFi247777CpaNHDmSnTt3smjRImrXrs2nn35K165diYuLIyAgoPRfeAVhsxk8v2I7a+JO4Ghvx8dDQmkZVM3sWCIiImXCYhiGYdaTh4WF0bp1a+bMmVOwrHHjxvTr14+pU6desn54eDgdOnTgzTffLFg2ZswYoqOj2bBhQ5HP8fbbb/Pyyy+TmJiIm5sbFy5cwMPDg6+//rrQ0aWWLVvSq1cv/v3vf19T9tTUVLy8vEhJScHTs/IdWTEMg8nf7OKTzUew2ln4cFAIXZv4mR1LRETkhlzP+7dpp9uys7OJiYmhW7duhZZ369aNTZs2FfmYrKwsnJ2dCy1zcXEhKiqKnJycIh8zb948HnjgAdzc8q+hyc3NJS8vr8jtXK5oXXzu1NTUQl+V2f+t3ssnm49gscDM+1uoIImISJVjWkk6ffo0eXl5+PkVfvP18/MjKSmpyMd0796duXPnEhMTg2EYREdHExERQU5ODqdPn75k/aioKHbu3MnIkSMLlnl4eNC+fXtee+01jh8/Tl5eHp9++imRkZEkJiZeNu/UqVPx8vIq+AoKCirmKy//Plh3gFm/HgDgtb5N6dtSpyBFRKTqMf3C7b/fDNUwjMveIHXSpEn06NGDdu3a4eDgQN++fRk6dCgAVuulE5/nzZtH06ZNadu2baHlixYtwjAMAgICcHJy4t133+Whhx4qchsXTZgwgZSUlIKvY8eOXecrrRgWRx5h2g97ABh/VyMGtatjciIRERFzmFaSfHx8sFqtlxw1Onny5CVHly5ycXEhIiKCjIwMDh8+zNGjR6lbty4eHh74+PgUWjcjI4OlS5cWOop0Uf369Vm3bh3nz5/n2LFjBafr6tWrd9m8Tk5OeHp6FvqqbL7elsDEr3YC8GTn+jzRub7JiURERMxjWklydHQkJCSENWvWFFq+Zs0awsPDr/hYBwcHAgMDsVqtLF26lF69emFnV/ilfP7552RlZTFo0KDLbsfNzQ1/f3/Onj3LTz/9RN++fYv/giq4n3efYNznf2AYMLhdHZ7rHmx2JBEREVOZOgLg2WefZfDgwYSGhtK+fXs++ugjjh49yuOPPw7kn+JKSEgomIW0b98+oqKiCAsL4+zZs8ycOZOdO3eycOHCS7Y9b948+vXrh7e39yU/++mnnzAMg+DgYPbv389zzz1HcHAww4YNK90XXE5tPpDME4u3kmszuKdVAFP63HrZU54iIiJVhaklaeDAgSQnJ/Pqq6+SmJhI06ZNWbVqFXXq5F8Hk5iYyNGjRwvWz8vLY8aMGezduxcHBwe6dOnCpk2bqFu3bqHt7tu3jw0bNrB69eoinzclJYUJEyYQHx9PjRo16N+/P6+//joODlXvRq3bjp1j5MItZOfauLOJH9MHNMfOTgVJRETE1DlJFVllmJO0NymNgR9t5lxGDuH1vYkY2gZnh8tfvC4iIlLRVYg5SWKuI8npDJ4XybmMHFoGVePjR0JVkERERP6HSlIVlJSSycNzIzmZlkWjWh4sGNYGNydTz7yKiIiUOypJVcyZ9GwGzYsk/uwF6nq78smItlRzdTQ7loiISLmjklSFpGbmMCQiiv0nz+Pv5cynI8Oo6eF89QeKiIhUQSpJVcSF7DxGLohmR0IK3m6OLBoRRmB1V7NjiYiIlFsqSVVAdq6NJxbHEHX4DB5O9iwc3pYGNd3NjiUiIlKuqSRVcnk2g7Gfb2Pt3lM4O9gRMawNTQO8zI4lIiJS7qkkVWKGYfDSlzv4fnsiDlYLHw4OpU3dGmbHEhERqRBUkiopwzB4/fvdLN1yDDsLvPNAKzo19DU7loiISIWhklRJvffLfuZuOATAtP7NubuZv8mJREREKhaVpEpo/sZDzFyzD4CXezXh/tAgkxOJiIhUPCpJlczy6GNM+TYOgDFdb2F4x3omJxIREamYVJIqkR93JjL+i+0AjOhYj9F33GJyIhERkYpLJamS+G3fKZ5eEovNgPtDA5nYszEWi8XsWCIiIhWWSlIlEH34DI8uiiYnz6BnM3+m3ttcBUlEROQGqSRVcDsTUhi2YAuZOTY6NfTlrYEtsdqpIImIiNwolaQK7MCp8wyJiCItM5c2davzwaAQHO31f6mIiEhJ0DtqBRV/NoNBcyNJTs+maYAn84a2wcXRanYsERGRSkMlqQI6lZbFoLmRJKZkUt/XjYXD2uLp7GB2LBERkUpFJamCScnIYfC8SA4nZxBQzYVPR4bh7e5kdiwREZFKRyWpAknPymXogij2JKXh6+HE4pFh+Hu5mB1LRESkUlJJqiAyc/J4dFE0sUfP4eXiwKIRbanr42Z2LBERkUpLJakCyM2z8fSSWDbuT8bN0crC4W1pVMvT7FgiIiKVmkpSOWezGTy/Yjtr4k7gaG/Hx0NCaRlUzexYIiIilZ5KUjlmGAavfLuLlbEJWO0szH6oNeH1fcyOJSIiUiWoJJVj/7d6L59sPoLFAjPvb0HXJn5mRxIREakyVJLKqQ/WHWDWrwcAeK1vU/q2DDA5kYiISNWiklQOLY48wrQf9gAw/q5GDGpXx+REIiIiVY9KUjnz9bYEJn61E4AnO9fnic71TU4kIiJSNdmbHUAK83F3wsXBSv/WgTzXPdjsOCIiIlWWSlI506GBD98+3ZF63m5YLBaz44iIiFRZKknlUH1fd7MjiIiIVHm6JklERESkCCpJIiIiIkVQSRIREREpgkqSiIiISBFUkkRERESKoJIkIiIiUgSVJBEREZEiqCSJiIiIFEElSURERKQIKkkiIiIiRVBJEhERESmCSpKIiIhIEVSSRERERIpgb3aAisowDABSU1NNTiIiIiLX6uL79sX38StRSSqmtLQ0AIKCgkxOIiIiItcrLS0NLy+vK65jMa6lSsklbDYbx48fx8PDA4vFUqLbTk1NJSgoiGPHjuHp6Vmi25b/T/u5bGg/lw3t57Kh/Vx2SmtfG4ZBWloatWvXxs7uylcd6UhSMdnZ2REYGFiqz+Hp6an/CMuA9nPZ0H4uG9rPZUP7ueyUxr6+2hGki3ThtoiIiEgRVJJEREREiqCSVA45OTkxefJknJyczI5SqWk/lw3t57Kh/Vw2tJ/LTnnY17pwW0RERKQIOpIkIiIiUgSVJBEREZEiqCSJiIiIFEElSURERKQIKkkmmT17NvXq1cPZ2ZmQkBDWr19/xfXXrVtHSEgIzs7O3HzzzXzwwQdllLRiu579vHLlSu688058fX3x9PSkffv2/PTTT2WYtuK63n/PF23cuBF7e3tatmxZugErievdz1lZWbz00kvUqVMHJycn6tevT0RERBmlrbiudz8vXryYFi1a4Orqir+/P8OGDSM5ObmM0lZMv/32G71796Z27dpYLBa++uqrqz7GlPdBQ8rc0qVLDQcHB+Pjjz824uLijNGjRxtubm7GkSNHilz/4MGDhqurqzF69GgjLi7O+Pjjjw0HBwdjxYoVZZy8Yrne/Tx69GjjP//5jxEVFWXs27fPmDBhguHg4GBs3bq1jJNXLNe7ny86d+6ccfPNNxvdunUzWrRoUTZhK7Di7Oc+ffoYYWFhxpo1a4xDhw4ZkZGRxsaNG8swdcVzvft5/fr1hp2dnfHOO+8YBw8eNNavX2/ceuutRr9+/co4ecWyatUq46WXXjK++OILAzC+/PLLK65v1vugSpIJ2rZtazz++OOFljVq1Mh44YUXilz/+eefNxo1alRo2WOPPWa0a9eu1DJWBte7n4vSpEkTY8qUKSUdrVIp7n4eOHCgMXHiRGPy5MkqSdfgevfzDz/8YHh5eRnJycllEa/SuN79/Oabbxo333xzoWXvvvuuERgYWGoZK5trKUlmvQ/qdFsZy87OJiYmhm7duhVa3q1bNzZt2lTkYzZv3nzJ+t27dyc6OpqcnJxSy1qRFWc//53NZiMtLY0aNWqURsRKobj7ef78+Rw4cIDJkyeXdsRKoTj7+ZtvviE0NJTp06cTEBBAw4YN+de//sWFCxfKInKFVJz9HB4eTnx8PKtWrcIwDE6cOMGKFSvo2bNnWUSuMsx6H9QNbsvY6dOnycvLw8/Pr9ByPz8/kpKSinxMUlJSkevn5uZy+vRp/P39Sy1vRVWc/fx3M2bMID09nfvvv780IlYKxdnPf/75Jy+88ALr16/H3l6/gq5FcfbzwYMH2bBhA87Oznz55ZecPn2aJ598kjNnzui6pMsozn4ODw9n8eLFDBw4kMzMTHJzc+nTpw/vvfdeWUSuMsx6H9SRJJNYLJZC3xuGccmyq61f1HIp7Hr380VLlizhlVdeYdmyZdSsWbO04lUa17qf8/LyeOihh5gyZQoNGzYsq3iVxvX8e7bZbFgsFhYvXkzbtm25++67mTlzJgsWLNDRpKu4nv0cFxfHM888w8svv0xMTAw//vgjhw4d4vHHHy+LqFWKGe+D+jOujPn4+GC1Wi/5q+TkyZOXtOSLatWqVeT69vb2eHt7l1rWiqw4+/miZcuWMWLECJYvX07Xrl1LM2aFd737OS0tjejoaGJjYxk1ahSQ/2ZuGAb29vasXr2a22+/vUyyVyTF+ffs7+9PQEAAXl5eBcsaN26MYRjEx8dzyy23lGrmiqg4+3nq1Kl06NCB5557DoDmzZvj5ubGbbfdxr///W8d6S8hZr0P6khSGXN0dCQkJIQ1a9YUWr5mzRrCw8OLfEz79u0vWX/16tWEhobi4OBQalkrsuLsZ8g/gjR06FA+++wzXVNwDa53P3t6erJjxw62bdtW8PX4448THBzMtm3bCAsLK6voFUpx/j136NCB48ePc/78+YJl+/btw87OjsDAwFLNW1EVZz9nZGRgZ1f4rdRqtQL//0iH3DjT3gdL9bJwKdLFj5jOmzfPiIuLM8aMGWO4ubkZhw8fNgzDMF544QVj8ODBBetf/Ojj2LFjjbi4OGPevHkaAXANrnc/f/bZZ4a9vb0xa9YsIzExseDr3LlzZr2ECuF69/Pf6dNt1+Z693NaWpoRGBhoDBgwwNi1a5exbt0645ZbbjFGjhxp1kuoEK53P8+fP9+wt7c3Zs+ebRw4cMDYsGGDERoaarRt29asl1AhpKWlGbGxsUZsbKwBGDNnzjRiY2MLRi2Ul/dBlSSTzJo1y6hTp47h6OhotG7d2li3bl3Bz4YMGWJ06tSp0Ppr1641WrVqZTg6Ohp169Y15syZU8aJK6br2c+dOnUygEu+hgwZUvbBK5jr/ff8v1SSrt317ufdu3cbXbt2NVxcXIzAwEDj2WefNTIyMso4dcVzvfv53XffNZo0aWK4uLgY/v7+xsMPP2zEx8eXceqK5ddff73i79vy8j5oMQwdDxQRERH5O12TJCIiIlIElSQRERGRIqgkiYiIiBRBJUlERESkCCpJIiIiIkVQSRIREREpgkqSiIiISBFUkkSk3OjcuTNjxowxO8ZVvfLKK/j5+WGxWPjqq6+KtY2hQ4fSr1+/63pM3bp1efvtt4v1fCJy/VSSROSG9e7d+7I3A968eTMWi4WtW7eWcarSsXv3bqZMmcKHH35IYmIiPXr0MDuSiJQSlSQRuWEjRozgl19+4ciRI5f8LCIigpYtW9K6dWsTkpW8AwcOANC3b19q1aqFk5OTyYlEpLSoJInIDevVqxc1a9ZkwYIFhZZnZGSwbNkyRowYQXJyMg8++CCBgYG4urrSrFkzlixZcsXtFnU6q1q1aoWeJyEhgYEDB1K9enW8vb3p27cvhw8fLvj52rVradu2LW5ublSrVo0OHToUWeYu2rFjB7fffjsuLi54e3vz6KOPcv78eSD/NFvv3r0BsLOzw2KxFLmNvLw8RowYQb169XBxcSE4OJh33nnniq+1c+fOjBo1ilGjRlGtWjW8vb2ZOHHiJXeSz8jIYPjw4Xh4eHDTTTfx0UcfFfr5+PHjadiwIa6urtx8881MmjSJnJycKz63iBRNJUlEbpi9vT2PPPIICxYsKPSmvnz5crKzs3n44YfJzMwkJCSE7777jp07d/Loo48yePBgIiMji/28GRkZdOnSBXd3d3777Tc2bNiAu7s7d911F9nZ2eTm5tKvXz86derE9u3b2bx5M48++uhly01GRgZ33XUX1atXZ8uWLSxfvpz//ve/jBo1CoB//etfzJ8/H4DExEQSExOL3I7NZiMwMJDPP/+cuLg4Xn75ZV588UU+//zzK76ehQsXYm9vT2RkJO+++y5vvfUWc+fOLbTOjBkzCA0NJTY2lieffJInnniCPXv2FPzcw8ODBQsWEBcXxzvvvMPHH3/MW2+9dc37VET+R6nfQldEqoTdu3cbgPHLL78ULPvHP/5hPPjgg5d9zN13322MGzeu4PtOnToZo0ePLvgeML788stCj/Hy8jLmz59vGIZhzJs3zwgODjZsNlvBz7OysgwXFxfjp59+MpKTkw3AWLt27TW9ho8++sioXr26cf78+YJl33//vWFnZ2ckJSUZhmEYX375pVGcX51PPvmk0b9//4LvhwwZYvTt27fg+06dOhmNGzcu9FrGjx9vNG7cuOD7OnXqGIMGDSr43mazGTVr1rzi3dCnT59uhISEXHdeETEMHUkSkRLRqFEjwsPDiYiIAPKv3Vm/fj3Dhw8H8k9Bvf766zRv3hxvb2/c3d1ZvXo1R48eLfZzxsTEsH//fjw8PHB3d8fd3Z0aNWqQmZnJgQMHqFGjBkOHDqV79+707t2bd95557JHfyD/ouwWLVrg5uZWsKxDhw7YbDb27t17Xdk++OADQkND8fX1xd3dnY8//viqr7Vdu3aFjnK1b9+eP//8k7y8vIJlzZs3L/jfFouFWrVqcfLkyYJlK1asoGPHjtSqVQt3d3cmTZp0Q/tYpCpTSRKREjNixAi++OILUlNTmT9/PnXq1OGOO+4A8k8TvfXWWzz//PP88ssvbNu2je7du5OdnX3Z7Vkslkuuyfnf62tsNhshISFs27at0Ne+fft46KGHAJg/fz6bN28mPDycZcuW0bBhQ37//fcin88wjMueirvc8qJ8/vnnjB07luHDh7N69Wq2bdvGsGHDrvhar5WDg8MluWw2GwC///47DzzwAD169OC7774jNjaWl156qUSeV6Qqsjc7gIhUHvfffz+jR4/ms88+Y+HChfzzn/8sKBfr16+nb9++DBo0CMgvOH/++SeNGze+7PZ8fX0LHfn5888/ycjIKPi+devWLFu2jJo1a+Lp6XnZ7bRq1YpWrVoxYcIE2rdvz2effUa7du0uWa9JkyYsXLiQ9PT0gqNJGzduxM7OjoYNG17zfli/fj3h4eE8+eSTBcsufiruSv5e3n7//XduueUWrFbrNT3vxo0bqVOnDi+99FLBsitdpC4iV6YjSSJSYtzd3Rk4cCAvvvgix48fZ+jQoQU/a9CgAWvWrGHTpk3s3r2bxx57jKSkpCtu7/bbb+f9999n69atREdH8/jjjxc6kvLwww/j4+ND3759Wb9+PYcOHWLdunWMHj2a+Ph4Dh06xIQJE9i8eTNHjhxh9erV7Nu377LF7OGHH8bZ2ZkhQ4awc+dOfv31V55++mkGDx6Mn5/fNe+HBg0aEB0dzU8//cS+ffuYNGkSW7Zsuerjjh07xrPPPsvevXtZsmQJ7733HqNHj76u5z169ChLly7lwIEDvPvuu3z55ZfX/HgRKUwlSURK1IgRIzh79ixdu3blpptuKlg+adIkWrduTffu3encuTO1atW66sTpGTNmEBQUxD/+8Q8eeugh/vWvf+Hq6lrwc1dXV3777Tduuukm7r33Xho3bszw4cO5cOECnp6euLq6smfPHvr370/Dhg159NFHGTVqFI899liRz+fq6spPP/3EmTNnaNOmDQMGDOCOO+7g/fffv6598Pjjj3PvvfcycOBAwsLCSE5OLnRU6XIeeeQRLly4QNu2bXnqqad4+umnefTRR6/5efv27cvYsWMZNWoULVu2ZNOmTUyaNOm6sovI/2cx/n7CX0REylznzp1p2bKlbjsiUo7oSJKIiIhIEVSSRERERIqg020iIiIiRdCRJBEREZEiqCSJiIiIFEElSURERKQIKkkiIiIiRVBJEhERESmCSpKIiIhIEVSSRERERIqgkiQiIiJSBJUkERERkSL8P4flGWwqfx3lAAAAAElFTkSuQmCC",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "\"\"\"\n",
    "Your code here\n",
    "\"\"\"\n",
    "from sklearn.naive_bayes import MultinomialNB\n",
    "acc1=[]\n",
    "print(\"Accuracy when using sklearn MultinomialNB\")\n",
    "for i in [0,0.4,0.6,0.8,1.0]:\n",
    "    nb1 = MultinomialNB(alpha=i)\n",
    "    nb1.fit(mat,res)\n",
    "    sk_y1 = nb1.predict(mat1)\n",
    "    acrc1=accuracy_score(sk_y1,res1)\n",
    "    acc1.append(acrc1)\n",
    "    print(\"At 𝛼 = \",i,\" Accuracy = \",acrc1)\n",
    "\n",
    "plt.plot([0.0,0.4,0.6,0.8,1.0],acc1)\n",
    "plt.xlabel('Values of alpha') \n",
    "plt.ylabel('accuracy') \n",
    "plt.title('Smoothing Priors')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "33097981",
   "metadata": {},
   "source": [
    "The max accuracy in my model and sklearn model is same at 𝛼=0.8 but the diffrence is when 𝛼=0,0.4,0.6 my model show less accuracy as compair to sklearn model"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.9.13"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}