{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "\n",
    "import nltk\n",
    "import spacy\n",
    "\n",
    "import warnings\n",
    "warnings.filterwarnings('ignore')\n",
    "\n",
    "import seaborn as sns\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "\n",
    "from IPython.display import Image"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### This notebook covers the basics of starting with text data and implementing machine learning algorithms that take that data as input and output predictions. \n",
    "\n",
    "#### Using scikitlearn, we will walk through:\n",
    "- Naive Bayes Classifier\n",
    "- Logistic Regression "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Naive Bayes Classifier\n",
    "\n",
    "Generative model: Naive Bayes models the joint distribution of the feature X and target Y, and then predicts the posterior probability given as P(y|x)\n",
    "\n",
    "\n",
    "The naive bayes classifier uses Baye's rule to assign labels to input data, with the output as the label with the highest probability for a number of input features.\n",
    "\n",
    "\n",
    "### Steps:\n",
    "- Loading dataset\n",
    "- Visualization\n",
    "- Preprocessing\n",
    "- Model creation\n",
    "- Evaluation"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Loading the dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "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>sms</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                                                sms\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": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_sms = pd.read_csv('data/spam.csv',encoding='latin-1')\n",
    "df_sms = df_sms.drop([\"Unnamed: 2\", \"Unnamed: 3\", \"Unnamed: 4\"], axis=1)\n",
    "df_sms = df_sms.rename(columns={\"v1\":\"label\", \"v2\":\"sms\"})\n",
    "df_sms.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "ham     4825\n",
       "spam     747\n",
       "Name: label, dtype: int64"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_sms['label'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_sms['length'] = df_sms['sms'].apply(len)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Visualization"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.distplot(df_sms['length'])\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "df_sms.hist(column='length', by='label', bins=100,figsize=(10,4))\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "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>sms</th>\n",
       "      <th>length</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>Go until jurong point, crazy.. Available only ...</td>\n",
       "      <td>111</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>Ok lar... Joking wif u oni...</td>\n",
       "      <td>29</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>Free entry in 2 a wkly comp to win FA Cup fina...</td>\n",
       "      <td>155</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "      <td>U dun say so early hor... U c already then say...</td>\n",
       "      <td>49</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>Nah I don't think he goes to usf, he lives aro...</td>\n",
       "      <td>61</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   label                                                sms  length\n",
       "0      0  Go until jurong point, crazy.. Available only ...     111\n",
       "1      0                      Ok lar... Joking wif u oni...      29\n",
       "2      1  Free entry in 2 a wkly comp to win FA Cup fina...     155\n",
       "3      0  U dun say so early hor... U c already then say...      49\n",
       "4      0  Nah I don't think he goes to usf, he lives aro...      61"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_sms.loc[:,'label'] = df_sms.label.map({'ham':0, 'spam':1})\n",
    "df_sms.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Scikitlearn Count Vectorizer for Preprocessing\n",
    "\n",
    "Convert a collection of text documents to a matrix of token counts\n",
    "\n",
    "This implementation produces a sparse representation of the counts using scipy.sparse.csr_matrix.\n",
    "\n",
    "\n",
    "#### Steps:\n",
    "- Normalize to lower case\n",
    "- Remove punctuation\n",
    "- Tokenize\n",
    "- Stop word removal\n",
    "- Bag-of-words representation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "from nltk.corpus import stopwords\n",
    "stopWords = set(stopwords.words('english'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.feature_extraction.text import CountVectorizer\n",
    "from sklearn.model_selection import train_test_split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "documents = [\"Hi Mom, happy Thanksgiving, I know I'm your favorite.\",\n",
    "            \"However, I never read rubrics when I do peer reviews.\",\n",
    "            \"Epstein did not kill himself!\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['epstein',\n",
       " 'favorite',\n",
       " 'happy',\n",
       " 'hi',\n",
       " 'however',\n",
       " 'kill',\n",
       " 'know',\n",
       " 'mom',\n",
       " 'never',\n",
       " 'peer',\n",
       " 'read',\n",
       " 'reviews',\n",
       " 'rubrics',\n",
       " 'thanksgiving']"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "count_vector = CountVectorizer(stop_words=stopWords)\n",
    "count_vector.fit(documents)\n",
    "count_vector.get_feature_names()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>epstein</th>\n",
       "      <th>favorite</th>\n",
       "      <th>happy</th>\n",
       "      <th>hi</th>\n",
       "      <th>however</th>\n",
       "      <th>kill</th>\n",
       "      <th>know</th>\n",
       "      <th>mom</th>\n",
       "      <th>never</th>\n",
       "      <th>peer</th>\n",
       "      <th>read</th>\n",
       "      <th>reviews</th>\n",
       "      <th>rubrics</th>\n",
       "      <th>thanksgiving</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",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   epstein  favorite  happy  hi  however  kill  know  mom  never  peer  read  \\\n",
       "0        0         1      1   1        0     0     1    1      0     0     0   \n",
       "1        0         0      0   0        1     0     0    0      1     1     1   \n",
       "2        1         0      0   0        0     1     0    0      0     0     0   \n",
       "\n",
       "   reviews  rubrics  thanksgiving  \n",
       "0        0        0             1  \n",
       "1        1        1             0  \n",
       "2        0        0             0  "
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#convert the set of documents into an array using count_vector.transform\n",
    "doc_array = count_vector.transform(documents).toarray()\n",
    "\n",
    "#convert doc_array to dataframe for visualization\n",
    "frequency_matrix = pd.DataFrame(doc_array, columns = count_vector.get_feature_names())\n",
    "frequency_matrix"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Preprocessing the dataset\n",
    "Now lets apply the count vectorizer to our dataset as part of our pre-processing\n",
    "- Split the dataset into train and test subsets \n",
    "- Apply count vectorizer to both subset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "#train_test_split is used to split up our data into subsets for training and testing\n",
    "X_train, X_test, y_train, y_test = train_test_split(df_sms['sms'], df_sms['label'],test_size=0.20)\n",
    "\n",
    "count_vector = CountVectorizer()\n",
    "\n",
    "# bag of words matrix for training data\n",
    "training_data = count_vector.fit_transform(X_train)\n",
    "\n",
    "# bag of words matrix for training data\n",
    "testing_data = count_vector.transform(X_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Creating the model - Naive Bayes Classifier\n",
    "\n",
    "Note that there are different Naive Bayes algorithms available in sklearn. Another commonly used one is Gaussian Naive Bayes which is useful for continuous attributes."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "MultinomialNB(alpha=1.0, class_prior=None, fit_prior=True)"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.naive_bayes import MultinomialNB \n",
    "naive_bayes = MultinomialNB()\n",
    "naive_bayes.fit(training_data,y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "predictions = naive_bayes.predict(testing_data)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Accuracy: the ratio of correctly predicted observation to the total observations\n",
    "<br><br> $$\\text{Accuracy}=\\frac{\\text{TP+TN}}{\\text{TP+FP+FN+TN}}$$<br>\n",
    "\n",
    "Precision: the ratio of correctly predicted positive observations to the total predicted positive observations\n",
    "<br><br> $$\\text{Precision}=\\frac{\\text{TP}}{\\text{TP+FP}}$$<br>\n",
    "\n",
    "Recall: the ratio of correctly predicted positive observations to the all observations in actual class  \n",
    "<br><br> $$\\text{Recall}=\\frac{\\text{TP}}{\\text{TP+FN}}$$<br>\n",
    "\n",
    "F1: the weighted average of Precision and Recall\n",
    "- takes both false positives and false negatives into account\n",
    "- useful for uneven class distribution\n",
    "\n",
    "<br><br> $$\\text{F1}=\\frac{\\text{2*(Recall * Precision)}}{\\text{(Recall + Precision)}}$$<br>\n",
    "\n",
    "\n",
    "#### Accuracy is used when the True Positives and True negatives are more important \n",
    "\n",
    "#### F1-score is used when the False Negatives and False Positives are more important"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy score: 0.9847533632286996\n",
      "Precision score: 0.9523809523809523\n",
      "Recall score: 0.9333333333333333\n",
      "F1 score: 0.9427609427609427\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import accuracy_score, precision_score, recall_score, f1_score\n",
    "print('Accuracy score: {}'.format(accuracy_score(y_test, predictions)))\n",
    "print('Precision score: {}'.format(precision_score(y_test, predictions)))\n",
    "print('Recall score: {}'.format(recall_score(y_test, predictions)))\n",
    "print('F1 score: {}'.format(f1_score(y_test, predictions)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### A confusion matrix is useful for visualizing your classifier.\n",
    "\n",
    "\n",
    "- true positives (TP): These are cases in which we predicted yes (they have the disease), and they do have the disease.\n",
    "\n",
    "- true negatives (TN): We predicted no, and they don't have the disease.\n",
    "\n",
    "- false positives (FP): We predicted yes, but they don't actually have the disease. (Also known as a \"Type I error.\")\n",
    "\n",
    "- false negatives (FN): We predicted no, but they actually do have the disease. (Also known as a \"Type II error.\")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAdgAAAIDCAYAAACnyA/oAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvOIA7rQAAIABJREFUeJzt3Xm8HFWZ8PHfk0QISyAsAYSgIKCICojIwLixyL664yiLogFnXHEUZhSN7zAjiusMyhAIm4BsigTZxLAMIPuOghDZDIuEQAICgQDP+0edGzqXm5tOuHUrqfv78qnP7T51uupUp+mnn3NOVUVmIkmSBtawphsgSVIbGWAlSaqBAVaSpBoYYCVJqoEBVpKkGhhgJUmqgQG2IRGxRUScHhEPRcTzETE9Ii6KiH0iYniN+901Im6LiFkRkRExegC3vWXZ5pYDtc0u9zu+7PfZiFi+j/X7lPUZEesu5Pa3XsDX3BcRxy/ovl6tiLi0HOeJfaz7TFm31kJsd9CPp+NYsuPf986IOCQilhzMtkgLwwDbgIj4MnAlsCJwEPB+4NPAXcCRwC417XcEcDLwILAdsAXw1ADu4sayzRsHcJsLYjbw4T7K9+HVHee3gQUKsMAHgP94Fft8tT4RERsM4PaaOp5bqT5TWwA7A6dS/Xv8oIG2SAtkRNMNGGoi4r3Aj4AjMvOLvVafHRE/ApapafdrAKOA0zPz/wZ645n5JHD1QG93Afwa2AuY2FMQEWsCWwInAPvW3YCIWDIzn8vMm+reVz9uBlanCogfGogNNng8T2Vm52fq4tIL8VHgCw21SeqKGezgOwh4HPh6Xysz8y+ZeWvP84jYLCJ+HxF/j4inI2JyRGzW+ZqIOD4ipkbE2yPi8oh4JiLujogDOuqMB+4rTyeWLrdLy7o+u/9KnfEdz98YEWdFxKOli/mBiDijZMZ9dhFH5SsR8efSFf5wRBwREcv1sa9DI+KLEXFvRDwVEZdFxFu6eVOLE4H3RsTrO8r2Au4HXvGDIiK2i4jzSpueiYjbI+KrnV30EdFzqbNvdHRVji/ret73LSLiDxHxLPD93u9pRAwr3Z33dXZhR8TbSrfn4QtwjN14Gvgv4IMR8Y7+KnbzHvRxPO8s78NufWzv5xExLSJe01E2LiJuKZ+ZxyJiYkSs+CqO70ngNZ0FEbFnRFxc9v33iLgpIvbpVee2iDirjzb3fG536CjbKCImRcQT5d/oyoh4T6/XvTOqYZ3ppc49EfHzV3FcahkD7CAqX1pbAb/LzFld1N8QuAxYgSr72htYDrgsIjbqVX054BTgJGB34DrgyIjYqqw/BvhIeXwoVZfbPy/gIZxLlQV/DtgeOBh4jv4/R/9JlbFfBOxKFYD2Bc6NiN6v+yRVN+CXgE8Br6PK6rvtabmc6kfEJzrK9qJ6T/q6JugbgMlU3fM7U2W540ube2xR/h7Py12Vx3SsX56q2/KXwI5U/wZzycyXyrGNAo4CiIilyuv+CHyjy+NbEP8LPMDcx9KXbt6DuWTmdcCfqY5pjohYAvgYcGpmzi5lhwE/A34P7AZ8DdgBOL93EJ+XiBhRluUiYheqf9/T+jiOM8u6PYBzgGM6f2RShl8iYvVer90fuBe4sOxvE+APVEM4n6XqBZgO/L7nB0tELFvqv0j1ed4R+H/YK6hOmekySAuwKtUX/Xe7rH8mMAMY3VG2HFUG/OuOsuPLdrfqKFuS6kthQkfZuqXevr32cx9wfB/7T2B8ebxyeb5bP+3dstTZsjxfkSoAH9+r3id7b6s8vxt4TUfZh0v5P87nfRpf6o2g+pK7o5RvVsrXo/oSTGDdeWwjyuu/ATwBDOvVtkP7eE3P+757H+te8Z5SjWMm1Y+HCVTjwusN8GfsUuCK8ni/sr/3lOefKc/XWoj3YK7jKXWeBZbvKNujbH+z8nwtqgD0rV77eVept0cXx5J9LJOAkf28blg5jqOBWzrKR1Flv4d0lI0pn9GDO8omA3cAS3SUDS9lvynPNy1t2XAg//1c2rWYwS7a3gv8NjNn9BRkNc45CXhfr7rPZOYlHfWeo5o09boBast04B7gsIj4bESs18VrNgeWoMogO50KvMArj+GiLJlPcVv5uyDHcCKwfkS8kyrjvzoz7+6rYkS8NiKOioj7geepJkkdCowGVulyf7OB33ZTMTPPospgj6TKjL44r7Z1tHFYRwY3otusrzie6jPwX/1sf2Hfg5OofsR9pKNsL+DPmXlteb4tVbA7ufMYgGuofly8t4tjuAV4Z1neTdV7shlwRkREx3GsFxG/jIgHyzHMpvpB8aaeOpn5VGn3Zzp6T/al+mFxbNnOUlSfyzOAlzraHFRZeE+b76b68XtURHwyqrF+aS4G2ME1nepX/+vnV7FYEXi4j/JHqLqNOz3RR73ngJFdt64fmZlUX5jXA98F7ipjTp/r52U942xzHUNmvkD1XvQeh3u81/Pnyt+ujyEzpwBXUWVve1IF3FcoX7CTqGZsH0o1S/idvNw12u0+p2Xmi922j6oLdkngUfroTu7Dt3g5YMymyq66Utr1LeDdEbFj7/Wv5j3IzJ5x7b3KtkZTdTH/oqNaT4Ce0usYZlNlkyt1cRh/z8zry3JlZv4v1eSmXai6mnu6ay8CNqIatnhPOY5jqd7rTj+n+sG2UwnQ44CzMvPRsn5Fqmz1kD7a/HlghYgYlpkzqYZ7HirbfKCMXw/IpDK1g+MFgygzX4hqYtG2UWabzucljwOr9VG+Gn0H1IU1iyrTnCMiXvHll5n3AHuXL6aNqL5wfh4R92Xm+X1stydgrkY11tiz7RFUX669A+pAOZFq3O8Fqmy5L+tQdfPtlZlzMuyI2HUB99X1/R4jYmmqL/3bqbqtDwO+Mp+XTWDuDHlBTzc6nSroHEo1Ltvp1b4HvwCOjmpS2fa8srdievm7HX1/Xqf3UdaNns/ShsD5VOPir6fqCr+ip1JfY/eZeXtEXE417jqLathk/44qM4CXqD4/ff44y2pMncy8GfhQ2c+mwL8Bp0fERpl5+0Iem1rEADv4DqMaW/o+1WSeuUTE2sCorGYSX0b1S3tU6d4iIkZRTRa6dADbdD/w1l5lO8+rcslmb46IA6kyxbdSfdH1djVVt+OezJ15fYzqs3fpwje5X6dRfeHfmpnz+iGydPk7p0u6zHz9RB91nweWGoB2/ZRqktjGVBnYTyLigsy8cF4vyMyHqLKkhZKZGRHfpArSvbOrBXkP+nIGcESpvyNweclse1xEFaxel5kXLUTz52XD8nda+dvXcaxANdmvLz+n+iGwAnBXZl7csyIzny4BeCPgxp5g2p/SI3N1RBxCNZHrzVQ/ojTEGWAHWWb+XwlMP4rqQgDHU832XAHYhmrc6J+oTrD/D6ov4skR8T2qbOkgqi+U/zeAzToVODYifkz1RbwRvc4ZLTOaf0oVvKZQdaPtS5UlXkwfMvPxiPgh8G8R8TRwHtWXz6HAFVSzkgdcCaofmE+1O6h+WPxnRLxI9eU8r2zyT8DOEXEBVSb2UAl8XStdh5+hyhbvAf47IrYDToiIDTu6KAdcZp4bEVdS/ejotCDvQV/bfTIizgb+BXgt1bhy5/q/lM/tERHxJqofjLOANamGG47pnDcwD6MiYvPyeAmq4HoI1cVSfl3K/0A1eelnEfFtqvPIvwk8RjXLu7dfAT+hmmz11T7WH0jV/X1hREykGuJYGdgEGJ6ZB5fZzOOA31DNQF4G+CJVD8NV8zkmDRVNz7Iaqgvwj1QZwMNUX2yPA7+jmmHbOXvzH6gmV/yd6vzGyZRZmh11jgem9rGPS4FLO57PaxbxMKqxuvuBZ6hOP1iHuWcRr0I1fnhXqfM41Rfm9h3b2ZKOWcSlLKi+tP9MlQk+TNX9tlyvNrxipi7VLNRXtLeP4xxf6o3op86+9JpFTJVJXlGOZyrVj5ZXzLSl+iK+gSo4dL4nfb7vZd19lFm3VAHlceCkXnXGlPfjPCAG6HN1KWUWca/y9/HyLNzOY+v2PZhzPL22u3OpO9eM4l519qLqzXi6fI7voMp8x3ZxLJ2zh58H/kI1UWxsr7pbAzeVdvyFKtiNp3S49LHto0rdleax/s1UPzwfpZoLMJVqvHqnsv5NVD827y2fi2nl3/EfBvq7wmXxXSKz6yEkSVrslTHTKVRd2ns13R61l13EkoaEqK4e9laqIZg1gR822yK1nQFW0lCxCXAJVbfvl7KaBSzVxi5iSZJq4IUmJEmqgQFWkqQaGGAlSaqBAVaSpBoYYCVJqoEBVpKkGhhgJUmqgQFWkqQaGGAlSaqBAVaSpBoYYCVJqoEBVpKkGhhgJUmqgQFWkqQaGGAlSaqBAVaSpBoYYCVJqoEBVpKkGhhgJUmqwYimGzAvse3YbLoN0qv17AV3Nd0EaUCMHL501LXtOr7v86KptbW3W2awkiTVYJHNYCVJQ0Q0nmzWwgxWkqQamMFKkprV0lSvpYclSVKzzGAlSc1q6RisAVaS1Kx2xle7iCVJqoMZrCSpWS3tIjaDlSSpBmawkqRmtTTVM8BKkpplF7EkSeqWGawkqVntTGDNYCVJqoMZrCSpWcPamcIaYCVJzWpnfLWLWJKkOpjBSpKa5Wk6kiSpW2awkqRmtTOBNYOVJKkOZrCSpGZ5mo4kSTVoZ3y1i1iSpDqYwUqSmuVpOpIkqVtmsJKkZjnJSZKkGrQzvtpFLElSHcxgJUnNcpKTJEnqlhmsJKlZ7UxgDbCSpIa1dBaxXcSSJNXADFaS1Kx2JrBmsJIk1cEMVpLUrJaepmOAlSQ1q6V9qS09LEmSmmUGK0lqVku7iM1gJUmqgRmsJKlZ7UxgzWAlSaqDGawkqVktHYM1wEqSmtXSvtSWHpYkSc0yg5UkNaulXcRmsJIk1cAMVpLUrHYmsAZYSVLDvOG6JEnqlhmsJKlZTnKSJEndMoOVJDWrnQmsAVaS1Kywi1iSJHXLDFaS1CgzWEmS1DUzWElSo1qawJrBSpJUBzNYSVKjhrU0hTXASpIa5SQnSZLUNTNYSVKjzGAlSVLXzGAlSY1qawZrgJUkNaql8dUuYknS0BQRX4mIP0bE7RHxy4gYGRFrR8Q1ETElIk6LiCVK3SXL8yll/Vrz274BVpLUqIgY8KWLfa4BfBHYNDPfCgwH9gS+B/w4M9cFngD2Ky/ZD3iilP+41OuXAVaSNFSNAJaKiBHA0sDDwNbAmWX9CcAe5fHu5Tll/TYxn0hugJUkNaqODDYixkXE9R3LuM59ZuaDwA+AB6gC60zgBmBGZr5Qqk0F1iiP1wD+Wl77Qqm/Un/H5SQnSVKjgoGf5ZSZE4AJ89xnxApUWenawAzgDGCHgWyDGawkaSh6P3BvZk7LzNnAr4F3AaNLlzHAWODB8vhBYE2Asn55YHp/OzDASpIa1cQkJ6qu4c0jYukylroN8CfgEuDDpc4+wNnl8aTynLL+4szM/nZggJUkDTmZeQ3VZKUbgduo4uEE4CDgwIiYQjXGOrG8ZCKwUik/EDh4fvtwDFaS1KimLjSRmd8Gvt2r+B5gsz7qzgI+siDbN4OVJKkGZrCSpEZ5w3VJkmrQ1ov920UsSVINzGAlSY0yg5UkSV0zg5UkNaqlCawBVpLULLuIJUlS18xgJUmNMoOVJEldM4OVJDWqrRmsAVaS1Ki2Bli7iCVJqoEZrCSpUS1NYM1gJUmqgxmsJKlRjsFKkqSumcFKkhrV1gzWACtJatSwlgZYu4glSaqBGawkqVEtTWDNYCVJqoMZrCSpUU5ykiSpBkE7A6xdxJIk1cAMtuW++IH9+OyOHyciOPq8U/jpWRP59l4H8tmd/olpM6cD8O/Hfo/zr72YEcNHcMyBh7PJem9jxPDhnHjRmRx26s8aPgJp3u679z6+fuBBc55Pnfog//yFz/HJvT/RYKu0oOwi1mLnLWu9ic/u+HE2+8IuPD97Nhd89yR+e81kAH78q6P54ZlHzVX/I+/dhSVfswQbjns/Sy05kj8dcwm/vORs7v/b1CaaL83XWmuvxelnnQbAiy++yLZbbs/W22zVbKOkwgDbYm9+3bpcc+fNPPvcLAAuu/VqPvjuHedZP0mWGbk0w4cNZ6klRvL8C7N58pm/D1ZzpVflmquvZc3XjWX1NVZvuilaQG3NYGsfg42IDSNit4j4YM9S9z5Vuf2+P/Oet23GiqNGs9SSI9lps61Zc0z15fP53ffllqMuYuJXf8DoZZcH4Mz/O5enZz3Dw6fdyAMnX8sPzjiKJ56a0eQhSF274LwL2WGnHZpuhhZCxMAvi4JaA2xEHAscC3wI2LUsu9S5T73szgem8L3Tfs7vDjuFC/7rJG7+yx958aUXOfKcE1lnn3ex8QHb8fDjj/LD/Q8BYLP1N+bFl15i9T3fwdp7b8FXPzyOtVd7XcNHIc3f7Odnc9kll7Hd9ts23RRpjrq7iDfPzA26rRwR44BxAKw/GsYuU1e7hoxjLziVYy84FYD//PRBTJ32MI/OeGzO+qPPO4Xf/sfxAPzT1ntwwfWX8sKLLzBtxnSu/ON1bPrGDbn3kQeaaLrUtSsuv4L1N1iflVZeqemmaCHYRbxwroqIrgNsZk7IzE0zc1OD68AYM7r6wllzzOp88F07csrFv2G1FVeZs/4D79qB2+/7MwAPPPoQW2/8jwAsPXIpNn/zJtz5178MfqOlBXT+eRewo93DWsTUncGeSBVkHwGeAwLIzNyw5v2q+NW3JrDScisw+4UX+JcjvsHMp5/kfz7/H2y8zlvITO7721/Z/ycHA/Czs4/nuK/9iNuPnkxEcNyFp3PbvXc0fARS/5555lmu/sM1HDL+m003RQuprRlsZGZ9G4+YAhwI3Aa81FOemffP97Xbjq2vYdIgefaCu5pugjQgRg5furYouN4Ptx/w7/u7v3ph41G77gx2WmZOqnkfkqTFWFsz2LoD7E0RcQpwDlUXMQCZ+eua9ytJWky0NL7WHmCXogqs23WUJWCAlSS1Wq0BNjM/Vef2JUmLP7uIF0JEjAT2A94CjOwpz8xP17lfSZKaVvd5sL8AVgO2By4DxgJP1bxPSdJiJCIGfFkU1D0Gu25mfiQids/ME8qEp8tr3qckaTGyqATEgVZ3Bju7/J0REW8FlgdW6ae+JEmtUHcGOyEiVgC+CUwClgUOqXmfkqTFSEsT2NoD7C+o7qSzFnBCKVu15n1KktS4ugPs2cBM4AY6LjQhSVKPto7B1h1gx2amt7iQJM1TWwNs3ZOc/hARb6t5H5IkLXJqyWAj4jaqSyKOAD4VEffg7eokSX1oawZbVxfxLjVtV5KkxUItAbab+71KkgTtPU2n7jFYSZKGpLpnEUuS1C/HYCVJqkNLA6xdxJIk1cAMVpLUqLZ2EZvBSpJUAzNYSVKjWprAGmAlSc2yi1iSJHXNDFaS1CgzWEmS1DUzWElSo9qawRpgJUmNaml8tYtYkqQ6mMFKkhrV1i5iM1hJkmpgBitJapQZrCRJ6poZrCSpUW3NYA2wkqRGtTXA2kUsSVINzGAlSY1qaQJrBitJUh3MYCVJjWrrGKwBVpLUqLYGWLuIJUmqgRmsJKlRZrCSJKlrZrCSpEa1NIE1wEqSmmUXsSRJ6poBVpLUrIiBX7rabYyOiDMj4s6IuCMitoiIFSPiooi4u/xdodSNiPjviJgSEbdGxCbz274BVpI0VP0UuCAz1wc2Au4ADgYmZ+Z6wOTyHGBHYL2yjAOOnN/GDbCSpEZFxIAvXexzeeC9wESAzHw+M2cAuwMnlGonAHuUx7sDJ2blamB0RLy2v30YYCVJjRoWA790YW1gGnBcRNwUEcdExDLAqpn5cKnzCLBqebwG8NeO108tZfM+rgV5EyRJWhxExLiIuL5jGderyghgE+DIzHw78DQvdwcDkJkJ5MK2wdN0JEmNquM0ncycAEzop8pUYGpmXlOen0kVYP8WEa/NzIdLF/CjZf2DwJodrx9byubJDFaSNORk5iPAXyPiTaVoG+BPwCRgn1K2D3B2eTwJ2LvMJt4cmNnRldwnM1hJUqOGNXehiS8AJ0fEEsA9wKeoEs/TI2I/4H7go6XuecBOwBTgmVK3XwZYSdKQlJk3A5v2sWqbPuom8C8Lsn0DrCSpUW29VKIBVpLUqLZOBmrrcUmS1CgzWElSoxqc5FQrM1hJkmpgBitJapSTnCRJqoFdxJIkqWtmsJKkRrW1i9gMVpKkGpjBSpIa1dZMzwArSWqUk5wkSVLXzGAlSY1ykpMkSeqaGawkqVGOwUqSpK6ZwUqSGtXO/NUAK0lqmF3EkiSpa2awkqRGmcFKkqSumcFKkhrV1gtNGGAlSY2yi1iSJHXNDFaS1Kh25q9msJIk1cIMVpLUqLaOwRpgJUmNamuAtYtYkqQamMFKkhrV1vNgzWAlSaqBGawkqVGOwUqSpK6ZwUqSGtXO/NUAK0lqWFu7iOcZYCPiHCDntT4zd6ulRZIktUB/GewPBq0VkqQha8hlsJl52WA2RJKkNpnvGGxErAd8F9gAGNlTnplvqLFdkqQhYihfaOI44EjgBWAr4ETgpDobJUkaOobVsCwKumnHUpk5GYjMvD8zxwM719ssSZIWb92cpvNcRAwD7o6IzwMPAsvW2yxJ0lAxlLuIvwQsDXwReAewF7BPnY2SJGlxN98MNjOvKw//Dnyq3uZIkoaaIXeaTo+IuIQ+LjiRmVvX0iJJ0pAyZAMs8K8dj0cCH6KaUSxJkuahmy7iG3oVXRkR19bUHknSENPWSU7ddBGv2PF0GNVEp+Vra1Hx7AV31b0LqXbTZj3SdBOkAbHmMl5baEF100V8A9UYbFB1Dd8L7FdnoyRJQ8ewlt6wrpsA++bMnNVZEBFL1tQeSZJaoZvzYP/QR9lVA90QSdLQFBEDviwK+rsf7GrAGsBSEfF2Xr7p/HJUF56QJOlVG4qn6WwP7AuMBX7IywH2SeDf622WJEmLt/7uB3sCcEJEfCgzfzWIbZIkDSHR0klO3YzBviMiRvc8iYgVIuLQGtskSdJir5sAu2Nmzuh5kplPADvV1yRJ0lAy5CY5dRgeEUtm5nMAEbEU4Gk6kqQBMRQnOfU4GZgcEcdRTXTaFzihzkZJkrS46+ZaxN+LiFuA91Nd0elC4PV1N0ySNDREV6OVi59uj+pvVMH1I8DWwB21tUiSpBbo70ITbwQ+XpbHgNOAyMytBqltkqQhYCiOwd4JXA7skplTACLiK4PSKknSkLGozPodaP11EX8QeBi4JCKOjohtoKVnA0uSNMDmGWAz8zeZuSewPnAJ8GVglYg4MiK2G6wGSpLaLWr4b1Ew30lOmfl0Zp6SmbtSXZf4JuCg2lsmSdJirJvzYOcoV3GaUBZJkl61tk5yaufJR5IkNWyBMlhJkgZaW2cRG2AlSY0a1tLO1HYelSRJDTODlSQ1qq1dxGawkiTVwAxWktSotmawBlhJUqOGLSJXXhpodhFLklQDM1hJUqPa2kVsBitJUg3MYCVJjWrrtYgNsJKkRi0qt5cbaHYRS5JUAzNYSVKjhkU7c712HpUkSQ0zwEqSGhURA74swL6HR8RNEfHb8nztiLgmIqZExGkRsUQpX7I8n1LWrzW/bRtgJUlD2ZeAOzqefw/4cWauCzwB7FfK9wOeKOU/LvX6ZYCVJDUqavivq/1GjAV2Bo4pzwPYGjizVDkB2KM83r08p6zfJuaTKjvJSZLUqDrOg42IccC4jqIJmTmhV7WfAF8HRpXnKwEzMvOF8nwqsEZ5vAbwV4DMfCEiZpb6j82rDQZYSVLrlGDaO6DOERG7AI9m5g0RsWUdbTDASpIa1dCFJt4F7BYROwEjgeWAnwKjI2JEyWLHAg+W+g8CawJTI2IEsDwwvb8dOAYrSRpyMvPfMnNsZq4F7AlcnJmfAC4BPlyq7QOcXR5PKs8p6y/OzOxvH2awkqRGLWLXIj4IODUiDgVuAiaW8onALyJiCvA4VVDulwFWktSoaPhKTpl5KXBpeXwPsFkfdWYBH1mQ7dpFLElSDcxgJUmN8m46kiSpa2awkqRGLWKTnAaMAVaS1KgFuTj/4sQuYkmSamAGK0lq1DAnOUmSpG6ZwUqSGuUYrCRJ6poZrCSpUU1fKrEuBlhJUqOc5CRJkrpmBitJapSTnCRJUtfMYCVJjWrr3XQMsJKkRtlFLEmSumYGK0lqlKfpSJKkrpnBSpIa5ZWcJEmqQVtnEbfzZ4MkSQ0zg5UkNcrTdCRJUtfMYCVJjXIMVpIkdc0MVpLUqLaOwRpgJUmN8kpOkiSpa2awkqRGtbWL2AxWkqQamMFKkhoVLc31DLCSpEbZRSxJkrpmBitJapRXcpIkSV0zg5UkNWpYS8dgDbCSpEbZRSxJkrpmBitJapSn6UiSpK6ZwUqSGuWVnCRJqoFdxJIkqWtmsJKkRnnDdUmS1DUzWElSoxyDlSRJXTODlSQ1qq2XSjTASpIaZRexJEnqmhmsJKlRbb2SUzuPSpKkhpnBSpIa5Q3XJUmqQVtnEdtFLElSDcxgJUmN8jQdSZLUNTNYSVKj2joGa4CVJDXKLmIt1r71jfFs+e6t+eBuH55TNnPGTPbf7wB23WE39t/vAJ6c+WSDLZTm7fDxP+LD2+zJZz5ywCvWnfGLX/H+TXZk5hMzAchMjvj+key926f57Ec/x913TBns5kqAAXbI2P0Du3LkhJ/NVXbsMcex2eabcc4Fk9hs882YeMxxDbVO6t/2u27Ld4849BXljz4yjeuvupFVVltlTtm1V17Hgw88xAlnT+Qr3/wiP/3uEYPZVC2EYTX8tyhYNFqh2r1j03ew3PLLz1V2ycWXstseuwKw2x67csnkS5pomjRfG77jbYxaftQryo/84VGM+/J+dPYw/uHSq9l2l22ICDbY8M38/am/M33a44PYWqlS6xhsRAwHdgbW6txXZv6ozv2qO49Pn86YMWMAWHnllXl8+vSGWyR178pLr2LlVVZmnTe+Ya7yxx6dzphVV57zfMwqK/PYtMdYacyKg91Edckx2IVzDrAvsBIwqmPpU0SMi4jrI+L6iUcfW3PT1CkioKUfcrXPrGdn8ctjT2OfA/ZquinSPNU9i3hsZm7YbeXMnABMAJj14jPpOyZcAAALPklEQVRZW6sEwIorrcS0adMYM2YM06ZNY8UV/YWvxcNDUx/mkQcfYf89/xmAaY8+xgGf+AI/O/EnrLzKSkz722Nz6k579DFWHrPyvDalRUBbT9OpO4M9PyK2q3kfWkhbbvU+Jv3mHAAm/eYcttp6y0bbI3XrDeutzZmTT+Xkc0/g5HNPYMwqK/O/J/8PK668Ilu8b3Mu+u1kMpM/3XoHyyy7jN3Di7iIGPBlUVB3Bns1cFZEDANmAwFkZi5X837Vy0H/ejDXX3sDM2bMYNuttudznz+AT3/2U3ztKwfxm1/9hteu/loO/9H3m26m1Kf//LfDuOWGW5k540n23OGT7HPAXuy4x/Z91v2Hd7+Ta6+4jr13/zRLjhzJ18Z/ZZBbK1Uis76e2Ii4F9gduC0XcEd2EasNps16pOkmSANizWXeUFtaeN20Kwb8+/6dY97deBpbdxfxX4HbFzS4SpK0uKu7i/ge4NKIOB94rqfQ03QkST3aOsmp7gB7b1mWKIskSXNbRCYlDbRaA2xmfqfO7UuStKiq+0pOY4CvA28BRvaUZ+bWde5XkrT4aGsXcd2TnE4G7gTWBr4D3AdcV/M+JUlqXN0BdqXMnAjMzszLMvPTgNmrJGkOLzSxcGaXvw9HxM7AQ4CXVJEkzdHWLuK6A+yhEbE88FXgf4DlAC+rIklqvbpnEf+2PJwJbFXnviRJi6e2ZrC1jsFGxBsi4pyIeCwiHo2IsyPiDfN/pSRJi7e6JzmdApwOrAasDpwB/LLmfUqSFiNtneRUd4BdOjN/kZkvlOUkOs6HlSSpCRGxZkRcEhF/iog/RsSXSvmKEXFRRNxd/q5QyiMi/jsipkTErRGxyfz2MRj3gz04ItaKiNdHxNeB88oBOJtYkkTU8F8XXgC+mpkbAJsD/xIRGwAHA5Mzcz1gcnkOsCOwXlnGAUfObwd1zyL+aPm7P9BzR50A9izPHY+VpCGuiUlOmfkw8HB5/FRE3AGsQXWL1S1LtROAS4GDSvmJ5e5wV0fE6Ih4bdlOn+rOYA8CNsrMtYHjgFuAD2Xm2plpcJUkNS4i1gLeDlwDrNoRNB8BVi2P16C6BWuPqaVsnuoOsN/MzCcj4t1UV3A6hi7SaknS0FHHJKeIGBcR13cs4+ax72WBXwFfzswnO9eVbHWh72dedxfxi+XvzsDRmXluRBxa8z4lSUNcZk4AJvRXJyJeQxVcT87MX5fiv/V0/UbEa4FHS/mDwJodLx9byuap7gz2wYg4CvgY1eSmJQdhn5KkxUgTk5yiOpdnInBHZv6oY9UkYJ/yeB/g7I7yvcts4s2Bmf2Nv8LgTHLaAfhBZs4ovwa+VvM+JUmLkYbOW30XsBdwW0TcXMr+HTgMOD0i9gPu5+XJuucBOwFTgGeAT81vB1F1MS96Zr34zKLZMGkBTJv1SNNNkAbEmsu8obYo+KcZNw/49/0Gozdu/GoTdWewkiT1y2sRS5KkrpnBSpIa1dYM1gArSWrUonJx/oFmF7EkSTUwg5UkNaqtXcRmsJIk1cAMVpLUKDNYSZLUNTNYSVKj2jqL2AArSWpYOwOsXcSSJNXADFaS1Ki2dhGbwUqSVAMzWElSo9p6mo4BVpLUqLYGWLuIJUmqgRmsJKlRTnKSJEldM4OVJDWqrWOwBlhJUqPaGmDtIpYkqQZmsJKkRjnJSZIkdc0MVpLUKMdgJUlS18xgJUmNausYrAFWktQou4glSVLXzGAlSQ0zg5UkSV0yg5UkNaqd+asBVpLUsLbOIraLWJKkGpjBSpIaZgYrSZK6ZAYrSWpUO/NXA6wkqXHtDLF2EUuSVAMzWElSozxNR5Ikdc0AK0lSDQywkiTVwDFYSVKj2no/WAOsJKlRbQ2wdhFLklQDA6wkSTUwwEqSVAPHYCVJjfJCE5IkqWsGWEmSamAXsSSpUZ6mI0mSumYGK0lqWDszWAOsJKlR7QyvdhFLklQLM1hJUqM8D1aSJHXNDFaS1DAzWEmS1CUzWElSo9qZvxpgJUmNa2eItYtYkqQamMFKkhrlaTqSJKlrBlhJkmpgF7EkqVHerk6SJHXNDFaS1DAzWEmS1CUzWElSo9qZvxpgJUkN8zxYSZLUNTNYSVLDzGAlSVKXzGAlSY1qZ/5qBitJUi3MYCVJDWtnDmuAlSQ1ytN0JElS1wywkiTVwAArSVINHIOVJDWqrfeDjcxsug1qSESMy8wJTbdDerX8LGtRZBfx0Dau6QZIA8TPshY5BlhJkmpggJUkqQYG2KHNMSu1hZ9lLXKc5CRJUg3MYCVJqoEBtoUiYq2IuL3pdkjSUGaAlSSpBgbY9hoeEUdHxB8j4ncRsVREfDYirouIWyLiVxGxNEBEHB8RR0bE1RFxT0RsGRHHRsQdEXF8w8ehISYilomIc8vn9PaI+FhE3BcR34+I2yLi2ohYt9TdNSKuiYibIuL3EbFqKR8fESdExOURcX9EfLDj9RdExGuaPUoNBQbY9loP+FlmvgWYAXwI+HVmvjMzNwLuAPbrqL8CsAXwFWAS8GPgLcDbImLjQW25hrodgIcyc6PMfCtwQSmfmZlvA44AflLKrgA2z8y3A6cCX+/YzjrA1sBuwEnAJeX1zwI7138YGuoMsO11b2beXB7fAKwFvLX8or8N+ARVAO1xTlZTym8D/paZt2XmS8Afy2ulwXIbsG1EfC8i3pOZM0v5Lzv+blEejwUuLJ/przH3Z/r8zJxdtjeclwP1bfiZ1iAwwLbXcx2PX6S6scPxwOfLr/jvACP7qP9Sr9e+hDeF0CDKzLuATagC4aER8a2eVZ3Vyt//AY4on+n96eMzXX4ozs6Xz0n0M61BYYAdWkYBD5fxp0803RipLxGxOvBMZp4EHE4VbAE+1vH3qvJ4eeDB8nifQWuk1AV/xQ0thwDXANPK31HNNkfq09uAwyPiJWA28DngTGCFiLiVKjP9eKk7HjgjIp4ALgbWHvzmSn3zSk6SFnkRcR+waWY+1nRbpG7ZRSxJUg3MYCVJqoEZrCRJNTDASpJUAwOsJEk1MMBKQES8GBE3l2vfntFzneaF3NaWEfHb8ni3iDi4n7qjI+KfF2If4yPiXxe2jZLqZ4CVKs9m5sbl2rfPAwd0rozKAv//kpmTMvOwfqqMBhY4wEpa9BlgpVe6HFi33Ff3zxFxInA7sGZEbBcRV0XEjSXTXRYgInaIiDsj4kbggz0bioh9I+KI8njViDir3CXmloj4R+AwYJ2SPR9e6n2t3PXo1oj4Tse2vhERd0XEFcCbBu3dkLRQvJKT1CEiRgA78vKF4dcD9snMqyNiZeCbwPsz8+mIOAg4MCK+DxxNdeeWKcBp89j8fwOXZeYHImI4sCxwMPDWzNy47H+7ss/NgAAmRcR7gaeBPYGNqf6/vZHqJg6SFlEGWKmyVET03H3ocmAisDpwf2ZeXco3BzYArowIgCWorom7PtXdi+4GiIiTgHF97GNrYG+AzHwRmBkRK/Sqs11ZbirPl6UKuKOAszLzmbKPSa/qaCXVzgArVZ7tySJ7lCD6dGcRcFFmfrxXvYG8X24A383Mo3rt48sDuA9Jg8AxWKl7VwPvioh1ASJimYh4I3AnsFZErFPqfXwer59MdeF6ImJ4RCwPPMXcN124EPh0x9juGhGxCvB/wB4RsVREjAJ2HeBjkzTADLBSlzJzGrAv8MtyV5ergPUzcxZVl/C5ZZLTo/PYxJeArcrNwW8ANsjM6VRdzrdHxOGZ+TvgFOCqUu9MYFRm3kg1tnsLcD5wXW0HKmlAeC1iSZJqYAYrSVINDLCSJNXAACtJUg0MsJIk1cAAK0lSDQywkiTVwAArSVINDLCSJNXg/wMpka2OoWkE2QAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 576x576 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.metrics import confusion_matrix\n",
    "\n",
    "conf_mat = confusion_matrix(y_test, predictions)\n",
    "fig, ax = plt.subplots(figsize=(8,8))\n",
    "sns.heatmap(conf_mat, annot=True, cmap=\"Greens\", fmt='d',\n",
    "            xticklabels=['ham','spam'], \n",
    "            yticklabels=['ham','spam'])\n",
    "plt.ylabel('Actual')\n",
    "plt.xlabel('Predicted')\n",
    "plt.title(\"Confusion Matrix - Naive Bayes\\n\", size=16);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Image(filename='data/confusionmatrix.png') "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Logistic Regression\n",
    "Discriminative model: Logistic regression directly models the posterior probability of P(y|x) by learning the input to output mapping by minimising the error.\n",
    "\n",
    "Uses logit (logistic unit): y = e^(b0 + b1*x) / (1 + e^(b0 + b1*x))\n",
    "Used to model binary output variables\n",
    "\n",
    "- Loading dataset\n",
    "- Visualization\n",
    "- Preprocessing\n",
    "- Model creation\n",
    "- Evaluation"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Loading the dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.read_csv('data/financial_product_complaints.csv', index_col = 0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "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>Product</th>\n",
       "      <th>Consumer_complaint</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Credit reporting, repair, or other</td>\n",
       "      <td>it is the repeated fraud attempt from experian...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Debt collection</td>\n",
       "      <td>We received multiple voice mails from Weltman,...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Debt collection</td>\n",
       "      <td>Deactivated my car whenever I was a day late m...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Student loan</td>\n",
       "      <td>My complaint is regarding my Student Loan Serv...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Debt collection</td>\n",
       "      <td>Syndicated Office Systems is currently reporti...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                              Product  \\\n",
       "0  Credit reporting, repair, or other   \n",
       "1                     Debt collection   \n",
       "2                     Debt collection   \n",
       "3                        Student loan   \n",
       "4                     Debt collection   \n",
       "\n",
       "                                  Consumer_complaint  \n",
       "0  it is the repeated fraud attempt from experian...  \n",
       "1  We received multiple voice mails from Weltman,...  \n",
       "2  Deactivated my car whenever I was a day late m...  \n",
       "3  My complaint is regarding my Student Loan Serv...  \n",
       "4  Syndicated Office Systems is currently reporti...  "
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "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>0</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Credit reporting, repair, or other</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Debt collection</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Student loan</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Checking or savings account</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Mortgage</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>Payday loan, title loan, or personal loan</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>Vehicle loan or lease</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>Credit card or prepaid card</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>Bank account or service</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>Consumer Loan</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>Money transfer, virtual currency, or money ser...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>Other financial service</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>Money transfers</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                    0\n",
       "0                  Credit reporting, repair, or other\n",
       "1                                     Debt collection\n",
       "2                                        Student loan\n",
       "3                         Checking or savings account\n",
       "4                                            Mortgage\n",
       "5           Payday loan, title loan, or personal loan\n",
       "6                               Vehicle loan or lease\n",
       "7                         Credit card or prepaid card\n",
       "8                             Bank account or service\n",
       "9                                       Consumer Loan\n",
       "10  Money transfer, virtual currency, or money ser...\n",
       "11                            Other financial service\n",
       "12                                    Money transfers"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.DataFrame(df.Product.unique())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "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>Product</th>\n",
       "      <th>category_id</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Credit reporting, repair, or other</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Debt collection</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Student loan</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>Checking or savings account</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>Mortgage</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>Payday loan, title loan, or personal loan</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>Vehicle loan or lease</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>47</th>\n",
       "      <td>Credit card or prepaid card</td>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>67</th>\n",
       "      <td>Bank account or service</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>70</th>\n",
       "      <td>Consumer Loan</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>118</th>\n",
       "      <td>Money transfer, virtual currency, or money ser...</td>\n",
       "      <td>10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>128</th>\n",
       "      <td>Other financial service</td>\n",
       "      <td>11</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1283</th>\n",
       "      <td>Money transfers</td>\n",
       "      <td>12</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                Product  category_id\n",
       "0                    Credit reporting, repair, or other            0\n",
       "1                                       Debt collection            1\n",
       "3                                          Student loan            2\n",
       "7                           Checking or savings account            3\n",
       "10                                             Mortgage            4\n",
       "14            Payday loan, title loan, or personal loan            5\n",
       "24                                Vehicle loan or lease            6\n",
       "47                          Credit card or prepaid card            7\n",
       "67                              Bank account or service            8\n",
       "70                                        Consumer Loan            9\n",
       "118   Money transfer, virtual currency, or money ser...           10\n",
       "128                             Other financial service           11\n",
       "1283                                    Money transfers           12"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Create a new column 'category_id' with encoded categories \n",
    "df['category_id'] = df['Product'].factorize()[0]\n",
    "category_id_df = df[['Product', 'category_id']].drop_duplicates()\n",
    "\n",
    "\n",
    "category_id_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "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>Product</th>\n",
       "      <th>Consumer_complaint</th>\n",
       "      <th>category_id</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Credit reporting, repair, or other</td>\n",
       "      <td>it is the repeated fraud attempt from experian...</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Debt collection</td>\n",
       "      <td>We received multiple voice mails from Weltman,...</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Debt collection</td>\n",
       "      <td>Deactivated my car whenever I was a day late m...</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Student loan</td>\n",
       "      <td>My complaint is regarding my Student Loan Serv...</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Debt collection</td>\n",
       "      <td>Syndicated Office Systems is currently reporti...</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                              Product  \\\n",
       "0  Credit reporting, repair, or other   \n",
       "1                     Debt collection   \n",
       "2                     Debt collection   \n",
       "3                        Student loan   \n",
       "4                     Debt collection   \n",
       "\n",
       "                                  Consumer_complaint  category_id  \n",
       "0  it is the repeated fraud attempt from experian...            0  \n",
       "1  We received multiple voice mails from Weltman,...            1  \n",
       "2  Deactivated my car whenever I was a day late m...            1  \n",
       "3  My complaint is regarding my Student Loan Serv...            2  \n",
       "4  Syndicated Office Systems is currently reporti...            1  "
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Dictionaries for future use\n",
    "category_to_id = dict(category_id_df.values)\n",
    "id_to_category = dict(category_id_df[['category_id', 'Product']].values)\n",
    "\n",
    "# New dataframe\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Text preprocessing\n",
    "- convert words to vectors\n",
    "- use TF-IDF to determine word importance\n",
    "\n",
    "__TF-IDF__ is the product of the __TF__ and __IDF__ scores of the term.<br><br> $$\\text{TF-IDF}=\\frac{\\text{TF}}{\\text{IDF}}$$<br>\n",
    "\n",
    "__Term Frequency :__ This summarizes how often a given word appears within a document.\n",
    "\n",
    "$$\\text{TF} = \\frac{\\text{Number of times the term appears in the doc}}{\\text{Total number of words in the doc}}$$<br><br>\n",
    "__Inverse Document Frequency:__ This downscales words that appear a lot across documents. A term has a high IDF score if it appears in a few documents. Conversely, if the term is very common among documents (i.e., “the”, “a”, “is”), the term would have a low IDF score.<br>\n",
    "\n",
    "$$\\text{IDF} = \\ln\\left(\\frac{\\text{Number of docs}}{\\text{Number docs the term appears in}} \\right)$$<br>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.feature_extraction.text import TfidfVectorizer"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "__n-gram__ is a contiguous sequence of n items from a given sample of text \n",
    "\n",
    "__unigrams__ are single words\n",
    "\n",
    "__bigram__ are double words"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "tfidf = TfidfVectorizer(sublinear_tf=True, min_df=5, ngram_range=(1, 2), stop_words='english')\n",
    "\n",
    "# use tfidf to transform complaints into vectors\n",
    "features = tfidf.fit_transform(df.Consumer_complaint).toarray()\n",
    "\n",
    "labels = df.category_id"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(10000, 28342)"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "features.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### A metric that you might use for feature selection is chi-squared \n",
    "\n",
    "- The chi-square test measures dependence between stochastic variables\n",
    "- A \"goodness of fit\" statistic, because it measures how well the observed distribution of data fits with the distribution that is expected if the variables are independent"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " Bank account or service: \n",
      "------------\n",
      "  -->Most Correlated Unigrams are: bank, deposit, overdraft\n",
      "  -->Most Correlated Bigrams are: overdraft fee, checking account, overdraft fees\n",
      "\n",
      " Checking or savings account: \n",
      "------------\n",
      "  -->Most Correlated Unigrams are: deposited, checking, overdraft\n",
      "  -->Most Correlated Bigrams are: debit card, savings account, checking account\n",
      "\n",
      " Consumer Loan: \n",
      "------------\n",
      "  -->Most Correlated Unigrams are: dealer, vehicle, car\n",
      "  -->Most Correlated Bigrams are: credit acceptance, leased vehicle, acceptance corp\n",
      "\n",
      " Credit card or prepaid card: \n",
      "------------\n",
      "  -->Most Correlated Unigrams are: express, citi, card\n",
      "  -->Most Correlated Bigrams are: synchrony bank, american express, credit card\n",
      "\n",
      " Credit reporting, repair, or other: \n",
      "------------\n",
      "  -->Most Correlated Unigrams are: experian, report, equifax\n",
      "  -->Most Correlated Bigrams are: credit bureaus, credit file, credit report\n",
      "\n",
      " Debt collection: \n",
      "------------\n",
      "  -->Most Correlated Unigrams are: collect, collection, debt\n",
      "  -->Most Correlated Bigrams are: debt collector, collection agency, collect debt\n",
      "\n",
      " Money transfer, virtual currency, or money service: \n",
      "------------\n",
      "  -->Most Correlated Unigrams are: wire, paypal, coinbase\n",
      "  -->Most Correlated Bigrams are: xxxx coinbase, coinbase xxxx, coinbase account\n",
      "\n",
      " Money transfers: \n",
      "------------\n",
      "  -->Most Correlated Unigrams are: paypal, gram, moneygram\n",
      "  -->Most Correlated Bigrams are: contacted paypal, money gram, money soon\n",
      "\n",
      " Mortgage: \n",
      "------------\n",
      "  -->Most Correlated Unigrams are: escrow, modification, mortgage\n",
      "  -->Most Correlated Bigrams are: mortgage company, mortgage payment, loan modification\n",
      "\n",
      " Other financial service: \n",
      "------------\n",
      "  -->Most Correlated Unigrams are: authorizations, screwed, global\n",
      "  -->Most Correlated Bigrams are: according usps, rate 11, check services\n",
      "\n",
      " Payday loan, title loan, or personal loan: \n",
      "------------\n",
      "  -->Most Correlated Unigrams are: loan, loaned, payday\n",
      "  -->Most Correlated Bigrams are: loan lending, payday loans, payday loan\n",
      "\n",
      " Student loan: \n",
      "------------\n",
      "  -->Most Correlated Unigrams are: student, loans, navient\n",
      "  -->Most Correlated Bigrams are: student loan, loan forgiveness, student loans\n",
      "\n",
      " Vehicle loan or lease: \n",
      "------------\n",
      "  -->Most Correlated Unigrams are: santander, car, vehicle\n",
      "  -->Most Correlated Bigrams are: consumer usa, gap insurance, santander consumer\n"
     ]
    }
   ],
   "source": [
    "from sklearn.feature_selection import chi2\n",
    "\n",
    "# Finding the three most correlated terms with each of the product categories\n",
    "N = 3\n",
    "for Product, category_id in sorted(category_to_id.items()):\n",
    "\n",
    "    #calculate chi2 values for features\n",
    "    features_chi2 = chi2(features, labels == category_id)\n",
    "    #select indices of features\n",
    "    indices = np.argsort(features_chi2[0])\n",
    "    #get feature names corresponding to those indices\n",
    "    feature_names = np.array(tfidf.get_feature_names())[indices]\n",
    "\n",
    "    #for each value in feature, extract uni and bi grams\n",
    "    unigrams = [v for v in feature_names if len(v.split(' ')) == 1]\n",
    "    bigrams = [v for v in feature_names if len(v.split(' ')) == 2]\n",
    "\n",
    "    #print top N uni and bigrams\n",
    "    print(\"\\n %s:\" %(Product),'\\n------------')\n",
    "    print(\"  -->Most Correlated Unigrams are: %s\" %(', '.join(unigrams[-N:])))\n",
    "    print(\"  -->Most Correlated Bigrams are: %s\" %(', '.join(bigrams[-N:])))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Before we can train and test a model, the data needs to be split up into train and test subsets. Scikitlearn's train_test_split package is useful for quickly completing this step - it also has parameters that you can alter to suit your data.\n",
    "\n",
    "- test_size\n",
    "- random_state\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "X = df['Consumer_complaint'] # complaint documents\n",
    "y = df['Product'] # Target labels we want to predict (13 different product types)\n",
    "\n",
    "#split the data into train and test sets\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.25)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.linear_model import LogisticRegression\n",
    "\n",
    "X_train, X_test, y_train, y_test,indices_train,indices_test = train_test_split(features, \n",
    "                                                               labels, \n",
    "                                                               df.index, test_size=0.20)\n",
    "model = LogisticRegression()\n",
    "model.fit(X_train, y_train)\n",
    "predictions = model.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy score: 0.7705\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import accuracy_score, precision_score, recall_score, f1_score\n",
    "print('Accuracy score: {}'.format(accuracy_score(y_test, predictions)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x576 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.metrics import confusion_matrix\n",
    "\n",
    "conf_mat = confusion_matrix(y_test, predictions)\n",
    "fig, ax = plt.subplots(figsize=(8,8))\n",
    "sns.heatmap(conf_mat, annot=True, cmap=\"Greens\", fmt='d',\n",
    "            xticklabels=category_id_df.Product.values, \n",
    "            yticklabels=category_id_df.Product.values)\n",
    "plt.ylabel('Actual')\n",
    "plt.xlabel('Predicted')\n",
    "plt.title(\"Confusion Matrix for Logistic Regression\\n\", size=16);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}