notebooks/ML Pipeline Preparation.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# ML Pipeline Preparation\n",
    "Follow the instructions below to help you create your ML pipeline.\n",
    "### 1. Import libraries and load data from database.\n",
    "- Import Python libraries\n",
    "- Load dataset from database with [`read_sql_table`](https://pandas.pydata.org/pandas-docs/stable/generated/pandas.read_sql_table.html)\n",
    "- Define feature and target variables X and Y"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[nltk_data] Downloading package wordnet to\n",
      "[nltk_data]     /Users/mdranamahmud/nltk_data...\n",
      "[nltk_data]   Package wordnet is already up-to-date!\n",
      "[nltk_data] Downloading package stopwords to\n",
      "[nltk_data]     /Users/mdranamahmud/nltk_data...\n",
      "[nltk_data]   Package stopwords is already up-to-date!\n",
      "[nltk_data] Downloading package punkt to\n",
      "[nltk_data]     /Users/mdranamahmud/nltk_data...\n",
      "[nltk_data]   Package punkt is already up-to-date!\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "True"
      ]
     },
     "execution_count": 1,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# import libraries\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "from sqlalchemy import create_engine\n",
    "import nltk\n",
    "from sklearn.pipeline import Pipeline\n",
    "from nltk.tokenize import word_tokenize\n",
    "from nltk import bigrams\n",
    "from nltk.corpus import stopwords\n",
    "from nltk.tokenize import sent_tokenize\n",
    "from nltk.stem.porter import PorterStemmer\n",
    "from sklearn.multioutput import MultiOutputClassifier\n",
    "from nltk.stem.wordnet import WordNetLemmatizer\n",
    "from sklearn.feature_extraction.text import TfidfVectorizer\n",
    "from sklearn.svm import SVC\n",
    "from sklearn.model_selection import train_test_split, GridSearchCV\n",
    "from sklearn.metrics import  f1_score,precision_score,recall_score,accuracy_score,make_scorer\n",
    "import re\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from sklearn.naive_bayes import MultinomialNB\n",
    "import matplotlib.pyplot as plt\n",
    "from sklearn.multiclass import OneVsRestClassifier\n",
    "from sklearn.neighbors import KNeighborsClassifier\n",
    "nltk.download('wordnet') # download for lemmatization\n",
    "nltk.download('stopwords')\n",
    "nltk.download('punkt')\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# load data from database\n",
    "engine = create_engine('sqlite:///etl.db')\n",
    "df = pd.read_sql_table(\"message_table\",engine)\n",
    "X = df['message']\n",
    "Y = df.drop(['id', 'message', 'original', 'genre'], axis = 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(12,8))\n",
    "Y.sum().plot(kind='bar')\n",
    "plt.title(\"Distribution of Response Type\")\n",
    "plt.savefig(\"Distribution of Response Type.png\",dpi=300)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 2. Write a tokenization function to process your text data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {},
   "outputs": [],
   "source": [
    "def tokenize(text):\n",
    "    # Normalize text\n",
    "    text = re.sub(r\"[^a-zA-Z0-9]\", \" \", text.lower())\n",
    "    # Tokenize text\n",
    "    words = word_tokenize(text)\n",
    "    words = [w for w in words if w not in stopwords.words(\"english\")]\n",
    "    # Reduce words to their stems\n",
    "    words = [PorterStemmer().stem(w) for w in words]\n",
    "    words = [WordNetLemmatizer().lemmatize(w) for w in words]\n",
    "\n",
    "    return words"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 3. Build a machine learning pipeline\n",
    "This machine pipeline should take in the `message` column as input and output classification results on the other 36 categories in the dataset. You may find the [MultiOutputClassifier](http://scikit-learn.org/stable/modules/generated/sklearn.multioutput.MultiOutputClassifier.html) helpful for predicting multiple target variables."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {},
   "outputs": [],
   "source": [
    "# compute bag of word counts and tf-idf values\n",
    "vectorizer = TfidfVectorizer( tokenizer=tokenize ,use_idf=True, smooth_idf=True, sublinear_tf=False)\n",
    "\n",
    "\n",
    "# clf = MultiOutputClassifier(RandomForestClassifier(random_state = 42))\n",
    "clf = RandomForestClassifier(random_state = 42)\n",
    "\n",
    "pipeline = Pipeline([('vectorizer',vectorizer), ('clf',clf)])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 4. Train pipeline\n",
    "- Split data into train and test sets\n",
    "- Train pipeline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Pipeline(memory=None,\n",
       "         steps=[('vectorizer',\n",
       "                 TfidfVectorizer(analyzer='word', binary=False,\n",
       "                                 decode_error='strict',\n",
       "                                 dtype=<class 'numpy.float64'>,\n",
       "                                 encoding='utf-8', input='content',\n",
       "                                 lowercase=True, max_df=1.0, max_features=None,\n",
       "                                 min_df=1, ngram_range=(1, 1), norm='l2',\n",
       "                                 preprocessor=None, smooth_idf=True,\n",
       "                                 stop_words=None, strip_accents=None,\n",
       "                                 sublinear_tf=False,\n",
       "                                 token_patt...\n",
       "                 RandomForestClassifier(bootstrap=True, ccp_alpha=0.0,\n",
       "                                        class_weight=None, criterion='gini',\n",
       "                                        max_depth=None, max_features='auto',\n",
       "                                        max_leaf_nodes=None, max_samples=None,\n",
       "                                        min_impurity_decrease=0.0,\n",
       "                                        min_impurity_split=None,\n",
       "                                        min_samples_leaf=1, min_samples_split=2,\n",
       "                                        min_weight_fraction_leaf=0.0,\n",
       "                                        n_estimators=100, n_jobs=None,\n",
       "                                        oob_score=False, random_state=42,\n",
       "                                        verbose=0, warm_start=False))],\n",
       "         verbose=False)"
      ]
     },
     "execution_count": 73,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(X, Y, random_state = 42)\n",
    "pipeline.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 5. Test your model\n",
    "Report the f1 score, precision and recall for each output category of the dataset. You can do this by iterating through the columns and calling sklearn's `classification_report` on each."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {},
   "outputs": [],
   "source": [
    "# do prediction\n",
    "y_pred_train = pipeline.predict(X_train)\n",
    "y_pred_test = pipeline.predict(X_test)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_metrics(test_value, predicted_value):\n",
    "    \"\"\"\n",
    "    get_metrics calculates f1 score, accuracy and recall\n",
    "\n",
    "    Args:\n",
    "        test_value (list): list of actual values\n",
    "        predicted_value (list): list of predicted values\n",
    "\n",
    "    Returns:\n",
    "        dictionray: a dictionary with accuracy, f1 score, precision and recall\n",
    "    \"\"\"\n",
    "    accuracy = accuracy_score(test_value,predicted_value)\n",
    "    precision =round( precision_score(test_value,predicted_value,average='micro'))\n",
    "    recall = recall_score(test_value,predicted_value,average='micro')\n",
    "    f1 = f1_score(test_value,predicted_value,average='micro')\n",
    "    return {'Accuracy':accuracy, 'f1 score':f1,'Precision':precision, 'Recall':recall}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "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>Accuracy</th>\n",
       "      <th>f1 score</th>\n",
       "      <th>Precision</th>\n",
       "      <th>Recall</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.998319</td>\n",
       "      <td>0.998319</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.998319</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.999338</td>\n",
       "      <td>0.999338</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.999338</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.998981</td>\n",
       "      <td>0.998981</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.998981</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.999745</td>\n",
       "      <td>0.999745</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.999745</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0.999847</td>\n",
       "      <td>0.999847</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.999847</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>0.999898</td>\n",
       "      <td>0.999898</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.999898</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>0.999694</td>\n",
       "      <td>0.999694</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.999694</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>0.999949</td>\n",
       "      <td>0.999949</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.999949</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>0.999949</td>\n",
       "      <td>0.999949</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.999949</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>0.999949</td>\n",
       "      <td>0.999949</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.999949</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>0.999898</td>\n",
       "      <td>0.999898</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.999898</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>0.999949</td>\n",
       "      <td>0.999949</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.999949</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>0.999949</td>\n",
       "      <td>0.999949</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.999949</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>0.999949</td>\n",
       "      <td>0.999949</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.999949</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>0.999847</td>\n",
       "      <td>0.999847</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.999847</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>0.999032</td>\n",
       "      <td>0.999032</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.999032</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>0.999796</td>\n",
       "      <td>0.999796</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.999796</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>0.999593</td>\n",
       "      <td>0.999593</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.999593</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>0.999949</td>\n",
       "      <td>0.999949</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.999949</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>0.999949</td>\n",
       "      <td>0.999949</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.999949</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>0.999949</td>\n",
       "      <td>0.999949</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.999949</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26</th>\n",
       "      <td>0.999949</td>\n",
       "      <td>0.999949</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.999949</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>0.999949</td>\n",
       "      <td>0.999949</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.999949</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28</th>\n",
       "      <td>0.999389</td>\n",
       "      <td>0.999389</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.999389</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29</th>\n",
       "      <td>0.999694</td>\n",
       "      <td>0.999694</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.999694</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30</th>\n",
       "      <td>0.999796</td>\n",
       "      <td>0.999796</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.999796</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>31</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>32</th>\n",
       "      <td>0.999643</td>\n",
       "      <td>0.999643</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.999643</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>33</th>\n",
       "      <td>0.999898</td>\n",
       "      <td>0.999898</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.999898</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>34</th>\n",
       "      <td>0.999796</td>\n",
       "      <td>0.999796</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.999796</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>35</th>\n",
       "      <td>0.999287</td>\n",
       "      <td>0.999287</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.999287</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    Accuracy  f1 score  Precision    Recall\n",
       "0   0.998319  0.998319        1.0  0.998319\n",
       "1   0.999338  0.999338        1.0  0.999338\n",
       "2   1.000000  1.000000        1.0  1.000000\n",
       "3   0.998981  0.998981        1.0  0.998981\n",
       "4   0.999745  0.999745        1.0  0.999745\n",
       "5   0.999847  0.999847        1.0  0.999847\n",
       "6   0.999898  0.999898        1.0  0.999898\n",
       "7   0.999694  0.999694        1.0  0.999694\n",
       "8   0.999949  0.999949        1.0  0.999949\n",
       "9   1.000000  1.000000        1.0  1.000000\n",
       "10  0.999949  0.999949        1.0  0.999949\n",
       "11  0.999949  0.999949        1.0  0.999949\n",
       "12  0.999898  0.999898        1.0  0.999898\n",
       "13  0.999949  0.999949        1.0  0.999949\n",
       "14  1.000000  1.000000        1.0  1.000000\n",
       "15  0.999949  0.999949        1.0  0.999949\n",
       "16  0.999949  0.999949        1.0  0.999949\n",
       "17  0.999847  0.999847        1.0  0.999847\n",
       "18  0.999032  0.999032        1.0  0.999032\n",
       "19  0.999796  0.999796        1.0  0.999796\n",
       "20  0.999593  0.999593        1.0  0.999593\n",
       "21  1.000000  1.000000        1.0  1.000000\n",
       "22  1.000000  1.000000        1.0  1.000000\n",
       "23  0.999949  0.999949        1.0  0.999949\n",
       "24  0.999949  0.999949        1.0  0.999949\n",
       "25  0.999949  0.999949        1.0  0.999949\n",
       "26  0.999949  0.999949        1.0  0.999949\n",
       "27  0.999949  0.999949        1.0  0.999949\n",
       "28  0.999389  0.999389        1.0  0.999389\n",
       "29  0.999694  0.999694        1.0  0.999694\n",
       "30  0.999796  0.999796        1.0  0.999796\n",
       "31  1.000000  1.000000        1.0  1.000000\n",
       "32  0.999643  0.999643        1.0  0.999643\n",
       "33  0.999898  0.999898        1.0  0.999898\n",
       "34  0.999796  0.999796        1.0  0.999796\n",
       "35  0.999287  0.999287        1.0  0.999287"
      ]
     },
     "execution_count": 76,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_results = []\n",
    "for i,column in enumerate(y_train.columns):\n",
    "    result = get_metrics(y_train.loc[:,column].values,y_pred_train[:,i])\n",
    "    train_results.append(result)\n",
    "train_results_df = pd.DataFrame(train_results)\n",
    "train_results_df    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "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>Accuracy</th>\n",
       "      <th>f1 score</th>\n",
       "      <th>Precision</th>\n",
       "      <th>Recall</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.831322</td>\n",
       "      <td>0.831322</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.831322</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.891826</td>\n",
       "      <td>0.891826</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.891826</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.995569</td>\n",
       "      <td>0.995569</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.995569</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.760275</td>\n",
       "      <td>0.760275</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.760275</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.914897</td>\n",
       "      <td>0.914897</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.914897</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0.950344</td>\n",
       "      <td>0.950344</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.950344</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>0.970970</td>\n",
       "      <td>0.970970</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.970970</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>0.983040</td>\n",
       "      <td>0.983040</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.983040</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>0.969442</td>\n",
       "      <td>0.969442</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.969442</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>0.946830</td>\n",
       "      <td>0.946830</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.946830</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>0.928189</td>\n",
       "      <td>0.928189</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.928189</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>0.926662</td>\n",
       "      <td>0.926662</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.926662</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>0.986249</td>\n",
       "      <td>0.986249</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.986249</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>0.979985</td>\n",
       "      <td>0.979985</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.979985</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>0.989152</td>\n",
       "      <td>0.989152</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.989152</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>0.961956</td>\n",
       "      <td>0.961956</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.961956</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>0.954775</td>\n",
       "      <td>0.954775</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.954775</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>0.873644</td>\n",
       "      <td>0.873644</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.873644</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>0.928801</td>\n",
       "      <td>0.928801</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.928801</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>0.951566</td>\n",
       "      <td>0.951566</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.951566</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>0.945607</td>\n",
       "      <td>0.945607</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.945607</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>0.982277</td>\n",
       "      <td>0.982277</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.982277</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>0.993888</td>\n",
       "      <td>0.993888</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.993888</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>0.988541</td>\n",
       "      <td>0.988541</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.988541</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>0.993583</td>\n",
       "      <td>0.993583</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.993583</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26</th>\n",
       "      <td>0.986860</td>\n",
       "      <td>0.986860</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.986860</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>0.953247</td>\n",
       "      <td>0.953247</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.953247</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28</th>\n",
       "      <td>0.857907</td>\n",
       "      <td>0.857907</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.857907</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29</th>\n",
       "      <td>0.941940</td>\n",
       "      <td>0.941940</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.941940</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30</th>\n",
       "      <td>0.927884</td>\n",
       "      <td>0.927884</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.927884</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>31</th>\n",
       "      <td>0.989916</td>\n",
       "      <td>0.989916</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.989916</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>32</th>\n",
       "      <td>0.955233</td>\n",
       "      <td>0.955233</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.955233</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>33</th>\n",
       "      <td>0.979526</td>\n",
       "      <td>0.979526</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.979526</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>34</th>\n",
       "      <td>0.946218</td>\n",
       "      <td>0.946218</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.946218</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>35</th>\n",
       "      <td>0.861574</td>\n",
       "      <td>0.861574</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.861574</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    Accuracy  f1 score  Precision    Recall\n",
       "0   0.831322  0.831322        1.0  0.831322\n",
       "1   0.891826  0.891826        1.0  0.891826\n",
       "2   0.995569  0.995569        1.0  0.995569\n",
       "3   0.760275  0.760275        1.0  0.760275\n",
       "4   0.914897  0.914897        1.0  0.914897\n",
       "5   0.950344  0.950344        1.0  0.950344\n",
       "6   0.970970  0.970970        1.0  0.970970\n",
       "7   0.983040  0.983040        1.0  0.983040\n",
       "8   0.969442  0.969442        1.0  0.969442\n",
       "9   1.000000  1.000000        1.0  1.000000\n",
       "10  0.946830  0.946830        1.0  0.946830\n",
       "11  0.928189  0.928189        1.0  0.928189\n",
       "12  0.926662  0.926662        1.0  0.926662\n",
       "13  0.986249  0.986249        1.0  0.986249\n",
       "14  0.979985  0.979985        1.0  0.979985\n",
       "15  0.989152  0.989152        1.0  0.989152\n",
       "16  0.961956  0.961956        1.0  0.961956\n",
       "17  0.954775  0.954775        1.0  0.954775\n",
       "18  0.873644  0.873644        1.0  0.873644\n",
       "19  0.928801  0.928801        1.0  0.928801\n",
       "20  0.951566  0.951566        1.0  0.951566\n",
       "21  0.945607  0.945607        1.0  0.945607\n",
       "22  0.982277  0.982277        1.0  0.982277\n",
       "23  0.993888  0.993888        1.0  0.993888\n",
       "24  0.988541  0.988541        1.0  0.988541\n",
       "25  0.993583  0.993583        1.0  0.993583\n",
       "26  0.986860  0.986860        1.0  0.986860\n",
       "27  0.953247  0.953247        1.0  0.953247\n",
       "28  0.857907  0.857907        1.0  0.857907\n",
       "29  0.941940  0.941940        1.0  0.941940\n",
       "30  0.927884  0.927884        1.0  0.927884\n",
       "31  0.989916  0.989916        1.0  0.989916\n",
       "32  0.955233  0.955233        1.0  0.955233\n",
       "33  0.979526  0.979526        1.0  0.979526\n",
       "34  0.946218  0.946218        1.0  0.946218\n",
       "35  0.861574  0.861574        1.0  0.861574"
      ]
     },
     "execution_count": 77,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_results = []\n",
    "for i,column in enumerate(y_test.columns):\n",
    "    result = get_metrics(y_test.loc[:,column].values,y_pred_test[:,i])\n",
    "    test_results.append(result)\n",
    "test_results_df = pd.DataFrame(test_results)\n",
    "test_results_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5,1,'Result for Training Data')"
      ]
     },
     "execution_count": 78,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "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": [
    "train_results_df.plot()\n",
    "plt.title(\"Result for Training Data\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5,1,'Result for Test Data')"
      ]
     },
     "execution_count": 79,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "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": [
    "%matplotlib inline\n",
    "test_results_df.plot()\n",
    "plt.title(\"Result for Test Data\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Accuracy     0.999748\n",
       "f1 score     0.999748\n",
       "Precision    1.000000\n",
       "Recall       0.999748\n",
       "dtype: float64"
      ]
     },
     "execution_count": 80,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_results_df.mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Accuracy     0.944436\n",
       "f1 score     0.944436\n",
       "Precision    1.000000\n",
       "Recall       0.944436\n",
       "dtype: float64"
      ]
     },
     "execution_count": 81,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_results_df.mean()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 6. Improve your model\n",
    "Use grid search to find better parameters. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {},
   "outputs": [],
   "source": [
    "# write custom scoring for multiclass classifier\n",
    "def multi_class_score(y_true, y_pred):\n",
    "    accuracy_results = []\n",
    "    for i,column in enumerate(y_train.columns):\n",
    "        accuracy = accuracy_score(y_true.loc[:,column].values,y_pred[:,i])\n",
    "        accuracy_results.append(accuracy)\n",
    "    avg_accuracy = np.mean(accuracy_results)\n",
    "    return avg_accuracy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {},
   "outputs": [],
   "source": [
    "score = make_scorer(multi_class_score)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 5 folds for each of 20 candidates, totalling 100 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Using backend LokyBackend with 8 concurrent workers.\n",
      "[Parallel(n_jobs=-1)]: Done   2 tasks      | elapsed:  8.9min\n",
      "[Parallel(n_jobs=-1)]: Done   9 tasks      | elapsed: 15.6min\n",
      "[Parallel(n_jobs=-1)]: Done  16 tasks      | elapsed: 15.8min\n",
      "[Parallel(n_jobs=-1)]: Done  25 tasks      | elapsed: 33.8min\n",
      "[Parallel(n_jobs=-1)]: Done  34 tasks      | elapsed: 43.1min\n",
      "[Parallel(n_jobs=-1)]: Done  45 tasks      | elapsed: 51.4min\n",
      "[Parallel(n_jobs=-1)]: Done  56 tasks      | elapsed: 59.2min\n",
      "[Parallel(n_jobs=-1)]: Done  69 tasks      | elapsed: 79.9min\n",
      "[Parallel(n_jobs=-1)]: Done  82 tasks      | elapsed: 104.6min\n",
      "[Parallel(n_jobs=-1)]: Done  96 out of 100 | elapsed: 114.9min remaining:  4.8min\n",
      "[Parallel(n_jobs=-1)]: Done 100 out of 100 | elapsed: 120.0min finished\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "GridSearchCV(cv=5, error_score=nan,\n",
       "             estimator=Pipeline(memory=None,\n",
       "                                steps=[('vectorizer',\n",
       "                                        TfidfVectorizer(analyzer='word',\n",
       "                                                        binary=False,\n",
       "                                                        decode_error='strict',\n",
       "                                                        dtype=<class 'numpy.float64'>,\n",
       "                                                        encoding='utf-8',\n",
       "                                                        input='content',\n",
       "                                                        lowercase=True,\n",
       "                                                        max_df=1.0,\n",
       "                                                        max_features=None,\n",
       "                                                        min_df=1,\n",
       "                                                        ngram_range=(1, 1),\n",
       "                                                        norm='l2',\n",
       "                                                        preprocessor=None,\n",
       "                                                        smooth_idf=True,\n",
       "                                                        stop_words=None,\n",
       "                                                        stri...\n",
       "                                                               oob_score=False,\n",
       "                                                               random_state=42,\n",
       "                                                               verbose=0,\n",
       "                                                               warm_start=False))],\n",
       "                                verbose=False),\n",
       "             iid='deprecated', n_jobs=-1,\n",
       "             param_grid={'clf__criterion': ['gini', 'entropy'],\n",
       "                         'clf__max_depth': [5, 10, 20, 30, 40],\n",
       "                         'clf__max_features': ['auto', 'sqrt'],\n",
       "                         'clf__random_state': [42]},\n",
       "             pre_dispatch='2*n_jobs', refit=True, return_train_score=True,\n",
       "             scoring=make_scorer(multi_class_score), verbose=10)"
      ]
     },
     "execution_count": 84,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "parameters = {'clf__criterion':['gini','entropy'],\n",
    " 'clf__max_features': ['auto', 'sqrt'],\n",
    " 'clf__max_depth': [5,10, 20, 30, 40],\n",
    "              'clf__random_state': [42]}\n",
    "\n",
    "cv_rf_tuned = GridSearchCV(pipeline, param_grid=parameters, scoring=score,\n",
    "                  n_jobs = -1, cv = 5, refit = True, return_train_score=True,verbose=10)\n",
    "cv_rf_tuned.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'clf__criterion': 'entropy',\n",
       " 'clf__max_depth': 40,\n",
       " 'clf__max_features': 'auto',\n",
       " 'clf__random_state': 42}"
      ]
     },
     "execution_count": 85,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cv_rf_tuned.best_params_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {},
   "outputs": [],
   "source": [
    "best_model_rf_tuned = cv_rf_tuned.best_estimator_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9335693653184887"
      ]
     },
     "execution_count": 87,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cv_rf_tuned.best_score_"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 7. Test your model\n",
    "Show the accuracy, precision, and recall of the tuned model.  \n",
    "\n",
    "Since this project focuses on code quality, process, and  pipelines, there is no minimum performance metric needed to pass. However, make sure to fine tune your models for accuracy, precision and recall to make your project stand out - especially for your portfolio!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "metadata": {},
   "outputs": [],
   "source": [
    "# predict using best estimator\n",
    "# do prediction\n",
    "y_pred_train = best_model_rf_tuned.predict(X_train)\n",
    "y_pred_test = best_model_rf_tuned.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "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>Accuracy</th>\n",
       "      <th>f1 score</th>\n",
       "      <th>Precision</th>\n",
       "      <th>Recall</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.766249</td>\n",
       "      <td>0.766249</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.766249</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.860177</td>\n",
       "      <td>0.860177</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.860177</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.995518</td>\n",
       "      <td>0.995518</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.995518</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.808425</td>\n",
       "      <td>0.808425</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.808425</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.926447</td>\n",
       "      <td>0.926447</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.926447</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0.955124</td>\n",
       "      <td>0.955124</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.955124</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>0.974022</td>\n",
       "      <td>0.974022</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.974022</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>0.981815</td>\n",
       "      <td>0.981815</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.981815</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>0.967757</td>\n",
       "      <td>0.967757</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.967757</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>0.944275</td>\n",
       "      <td>0.944275</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.944275</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>0.905002</td>\n",
       "      <td>0.905002</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.905002</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>0.925835</td>\n",
       "      <td>0.925835</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.925835</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>0.985228</td>\n",
       "      <td>0.985228</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.985228</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>0.976824</td>\n",
       "      <td>0.976824</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.976824</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>0.988896</td>\n",
       "      <td>0.988896</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.988896</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>0.970507</td>\n",
       "      <td>0.970507</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.970507</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>0.959403</td>\n",
       "      <td>0.959403</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.959403</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>0.873217</td>\n",
       "      <td>0.873217</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.873217</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>0.941015</td>\n",
       "      <td>0.941015</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.941015</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>0.957977</td>\n",
       "      <td>0.957977</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.957977</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>0.955430</td>\n",
       "      <td>0.955430</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.955430</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>0.980236</td>\n",
       "      <td>0.980236</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.980236</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>0.994142</td>\n",
       "      <td>0.994142</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.994142</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>0.990016</td>\n",
       "      <td>0.990016</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.990016</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>0.996333</td>\n",
       "      <td>0.996333</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.996333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26</th>\n",
       "      <td>0.989660</td>\n",
       "      <td>0.989660</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.989660</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>0.959250</td>\n",
       "      <td>0.959250</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.959250</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28</th>\n",
       "      <td>0.820854</td>\n",
       "      <td>0.820854</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.820854</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29</th>\n",
       "      <td>0.931948</td>\n",
       "      <td>0.931948</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.931948</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30</th>\n",
       "      <td>0.918348</td>\n",
       "      <td>0.918348</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.918348</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>31</th>\n",
       "      <td>0.989711</td>\n",
       "      <td>0.989711</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.989711</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>32</th>\n",
       "      <td>0.922015</td>\n",
       "      <td>0.922015</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.922015</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>33</th>\n",
       "      <td>0.981051</td>\n",
       "      <td>0.981051</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.981051</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>34</th>\n",
       "      <td>0.949470</td>\n",
       "      <td>0.949470</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.949470</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>35</th>\n",
       "      <td>0.832264</td>\n",
       "      <td>0.832264</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.832264</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    Accuracy  f1 score  Precision    Recall\n",
       "0   0.766249  0.766249        1.0  0.766249\n",
       "1   0.860177  0.860177        1.0  0.860177\n",
       "2   0.995518  0.995518        1.0  0.995518\n",
       "3   0.808425  0.808425        1.0  0.808425\n",
       "4   0.926447  0.926447        1.0  0.926447\n",
       "5   0.955124  0.955124        1.0  0.955124\n",
       "6   0.974022  0.974022        1.0  0.974022\n",
       "7   0.981815  0.981815        1.0  0.981815\n",
       "8   0.967757  0.967757        1.0  0.967757\n",
       "9   1.000000  1.000000        1.0  1.000000\n",
       "10  0.944275  0.944275        1.0  0.944275\n",
       "11  0.905002  0.905002        1.0  0.905002\n",
       "12  0.925835  0.925835        1.0  0.925835\n",
       "13  0.985228  0.985228        1.0  0.985228\n",
       "14  0.976824  0.976824        1.0  0.976824\n",
       "15  0.988896  0.988896        1.0  0.988896\n",
       "16  0.970507  0.970507        1.0  0.970507\n",
       "17  0.959403  0.959403        1.0  0.959403\n",
       "18  0.873217  0.873217        1.0  0.873217\n",
       "19  0.941015  0.941015        1.0  0.941015\n",
       "20  0.957977  0.957977        1.0  0.957977\n",
       "21  0.955430  0.955430        1.0  0.955430\n",
       "22  0.980236  0.980236        1.0  0.980236\n",
       "23  0.994142  0.994142        1.0  0.994142\n",
       "24  0.990016  0.990016        1.0  0.990016\n",
       "25  0.996333  0.996333        1.0  0.996333\n",
       "26  0.989660  0.989660        1.0  0.989660\n",
       "27  0.959250  0.959250        1.0  0.959250\n",
       "28  0.820854  0.820854        1.0  0.820854\n",
       "29  0.931948  0.931948        1.0  0.931948\n",
       "30  0.918348  0.918348        1.0  0.918348\n",
       "31  0.989711  0.989711        1.0  0.989711\n",
       "32  0.922015  0.922015        1.0  0.922015\n",
       "33  0.981051  0.981051        1.0  0.981051\n",
       "34  0.949470  0.949470        1.0  0.949470\n",
       "35  0.832264  0.832264        1.0  0.832264"
      ]
     },
     "execution_count": 89,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_results = []\n",
    "for i,column in enumerate(y_train.columns):\n",
    "    result = get_metrics(y_train.loc[:,column].values,y_pred_train[:,i])\n",
    "    train_results.append(result)\n",
    "train_results_df = pd.DataFrame(train_results)\n",
    "train_results_df    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 90,
   "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>Accuracy</th>\n",
       "      <th>f1 score</th>\n",
       "      <th>Precision</th>\n",
       "      <th>Recall</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.757983</td>\n",
       "      <td>0.757983</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.757983</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.836975</td>\n",
       "      <td>0.836975</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.836975</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.995569</td>\n",
       "      <td>0.995569</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.995569</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.708785</td>\n",
       "      <td>0.708785</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.708785</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.914591</td>\n",
       "      <td>0.914591</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.914591</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0.948969</td>\n",
       "      <td>0.948969</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.948969</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>0.970817</td>\n",
       "      <td>0.970817</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.970817</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>0.983040</td>\n",
       "      <td>0.983040</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.983040</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>0.968220</td>\n",
       "      <td>0.968220</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.968220</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>0.932773</td>\n",
       "      <td>0.932773</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.932773</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>0.892895</td>\n",
       "      <td>0.892895</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.892895</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>0.908174</td>\n",
       "      <td>0.908174</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.908174</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>0.985638</td>\n",
       "      <td>0.985638</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.985638</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>0.979374</td>\n",
       "      <td>0.979374</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.979374</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>0.989152</td>\n",
       "      <td>0.989152</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.989152</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>0.961650</td>\n",
       "      <td>0.961650</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.961650</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>0.952330</td>\n",
       "      <td>0.952330</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.952330</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>0.869366</td>\n",
       "      <td>0.869366</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.869366</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>0.928801</td>\n",
       "      <td>0.928801</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.928801</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>0.951108</td>\n",
       "      <td>0.951108</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.951108</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>0.944843</td>\n",
       "      <td>0.944843</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.944843</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>0.982124</td>\n",
       "      <td>0.982124</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.982124</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>0.993888</td>\n",
       "      <td>0.993888</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.993888</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>0.988541</td>\n",
       "      <td>0.988541</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.988541</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>0.993583</td>\n",
       "      <td>0.993583</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.993583</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26</th>\n",
       "      <td>0.986860</td>\n",
       "      <td>0.986860</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.986860</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>0.953247</td>\n",
       "      <td>0.953247</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.953247</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28</th>\n",
       "      <td>0.776165</td>\n",
       "      <td>0.776165</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.776165</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29</th>\n",
       "      <td>0.921314</td>\n",
       "      <td>0.921314</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.921314</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30</th>\n",
       "      <td>0.907716</td>\n",
       "      <td>0.907716</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.907716</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>31</th>\n",
       "      <td>0.989916</td>\n",
       "      <td>0.989916</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.989916</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>32</th>\n",
       "      <td>0.919022</td>\n",
       "      <td>0.919022</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.919022</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>33</th>\n",
       "      <td>0.979679</td>\n",
       "      <td>0.979679</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.979679</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>34</th>\n",
       "      <td>0.945760</td>\n",
       "      <td>0.945760</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.945760</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>35</th>\n",
       "      <td>0.817112</td>\n",
       "      <td>0.817112</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.817112</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    Accuracy  f1 score  Precision    Recall\n",
       "0   0.757983  0.757983        1.0  0.757983\n",
       "1   0.836975  0.836975        1.0  0.836975\n",
       "2   0.995569  0.995569        1.0  0.995569\n",
       "3   0.708785  0.708785        1.0  0.708785\n",
       "4   0.914591  0.914591        1.0  0.914591\n",
       "5   0.948969  0.948969        1.0  0.948969\n",
       "6   0.970817  0.970817        1.0  0.970817\n",
       "7   0.983040  0.983040        1.0  0.983040\n",
       "8   0.968220  0.968220        1.0  0.968220\n",
       "9   1.000000  1.000000        1.0  1.000000\n",
       "10  0.932773  0.932773        1.0  0.932773\n",
       "11  0.892895  0.892895        1.0  0.892895\n",
       "12  0.908174  0.908174        1.0  0.908174\n",
       "13  0.985638  0.985638        1.0  0.985638\n",
       "14  0.979374  0.979374        1.0  0.979374\n",
       "15  0.989152  0.989152        1.0  0.989152\n",
       "16  0.961650  0.961650        1.0  0.961650\n",
       "17  0.952330  0.952330        1.0  0.952330\n",
       "18  0.869366  0.869366        1.0  0.869366\n",
       "19  0.928801  0.928801        1.0  0.928801\n",
       "20  0.951108  0.951108        1.0  0.951108\n",
       "21  0.944843  0.944843        1.0  0.944843\n",
       "22  0.982124  0.982124        1.0  0.982124\n",
       "23  0.993888  0.993888        1.0  0.993888\n",
       "24  0.988541  0.988541        1.0  0.988541\n",
       "25  0.993583  0.993583        1.0  0.993583\n",
       "26  0.986860  0.986860        1.0  0.986860\n",
       "27  0.953247  0.953247        1.0  0.953247\n",
       "28  0.776165  0.776165        1.0  0.776165\n",
       "29  0.921314  0.921314        1.0  0.921314\n",
       "30  0.907716  0.907716        1.0  0.907716\n",
       "31  0.989916  0.989916        1.0  0.989916\n",
       "32  0.919022  0.919022        1.0  0.919022\n",
       "33  0.979679  0.979679        1.0  0.979679\n",
       "34  0.945760  0.945760        1.0  0.945760\n",
       "35  0.817112  0.817112        1.0  0.817112"
      ]
     },
     "execution_count": 90,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_results = []\n",
    "for i,column in enumerate(y_test.columns):\n",
    "    result = get_metrics(y_test.loc[:,column].values,y_pred_test[:,i])\n",
    "    test_results.append(result)\n",
    "test_results_df = pd.DataFrame(test_results)\n",
    "test_results_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 91,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5,1,'Result for Training Data')"
      ]
     },
     "execution_count": 91,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "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": [
    "train_results_df.plot()\n",
    "plt.title(\"Result for Training Data\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5,1,'Result for Test Data')"
      ]
     },
     "execution_count": 92,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "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": [
    "%matplotlib inline\n",
    "test_results_df.plot()\n",
    "plt.title(\"Result for Test Data\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 93,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Accuracy     0.940957\n",
       "f1 score     0.940957\n",
       "Precision    1.000000\n",
       "Recall       0.940957\n",
       "dtype: float64"
      ]
     },
     "execution_count": 93,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_results_df.mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 94,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Accuracy     0.931555\n",
       "f1 score     0.931555\n",
       "Precision    1.000000\n",
       "Recall       0.931555\n",
       "dtype: float64"
      ]
     },
     "execution_count": 94,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_results_df.mean()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 8. Try improving your model further. Here are a few ideas:\n",
    "* try other machine learning algorithms\n",
    "* add other features besides the TF-IDF"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## other machine learning model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 5 folds for each of 15 candidates, totalling 75 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Using backend LokyBackend with 8 concurrent workers.\n",
      "[Parallel(n_jobs=-1)]: Done   2 tasks      | elapsed: 11.2min\n",
      "[Parallel(n_jobs=-1)]: Done   9 tasks      | elapsed: 19.3min\n",
      "[Parallel(n_jobs=-1)]: Done  16 tasks      | elapsed: 20.0min\n",
      "[Parallel(n_jobs=-1)]: Done  25 tasks      | elapsed: 36.6min\n",
      "[Parallel(n_jobs=-1)]: Done  34 tasks      | elapsed: 45.9min\n",
      "[Parallel(n_jobs=-1)]: Done  45 tasks      | elapsed: 55.4min\n",
      "[Parallel(n_jobs=-1)]: Done  56 tasks      | elapsed: 64.9min\n",
      "[Parallel(n_jobs=-1)]: Done  68 out of  75 | elapsed: 82.5min remaining:  8.5min\n",
      "[Parallel(n_jobs=-1)]: Done  75 out of  75 | elapsed: 86.7min finished\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "GridSearchCV(cv=5, error_score=nan,\n",
       "             estimator=Pipeline(memory=None,\n",
       "                                steps=[('vectorizer',\n",
       "                                        TfidfVectorizer(analyzer='word',\n",
       "                                                        binary=False,\n",
       "                                                        decode_error='strict',\n",
       "                                                        dtype=<class 'numpy.float64'>,\n",
       "                                                        encoding='utf-8',\n",
       "                                                        input='content',\n",
       "                                                        lowercase=True,\n",
       "                                                        max_df=1.0,\n",
       "                                                        max_features=None,\n",
       "                                                        min_df=1,\n",
       "                                                        ngram_range=(1, 1),\n",
       "                                                        norm='l2',\n",
       "                                                        preprocessor=None,\n",
       "                                                        smooth_idf=True,\n",
       "                                                        stop_words=None,\n",
       "                                                        stri...\n",
       "                                        KNeighborsClassifier(algorithm='auto',\n",
       "                                                             leaf_size=30,\n",
       "                                                             metric='minkowski',\n",
       "                                                             metric_params=None,\n",
       "                                                             n_jobs=None,\n",
       "                                                             n_neighbors=5, p=2,\n",
       "                                                             weights='uniform'))],\n",
       "                                verbose=False),\n",
       "             iid='deprecated', n_jobs=-1,\n",
       "             param_grid={'clf__n_neighbors': [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11,\n",
       "                                              12, 13, 14, 15]},\n",
       "             pre_dispatch='2*n_jobs', refit=True, return_train_score=True,\n",
       "             scoring=make_scorer(multi_class_score), verbose=10)"
      ]
     },
     "execution_count": 95,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# knc\n",
    "parameters = {\n",
    "#     'clf__kernel': ['linear', 'rbf'],\n",
    "    'clf__n_neighbors':[1, 2,3,4,5,6,7,8,9,10, 11, 12, 13,14,15]\n",
    "}\n",
    "knc_clf = KNeighborsClassifier()\n",
    "pipeline = Pipeline([('vectorizer',vectorizer), ('clf',knc_clf)])\n",
    "\n",
    "\n",
    "clf_kn = GridSearchCV(pipeline, param_grid=parameters, scoring=score,\n",
    "                  n_jobs = -1, cv = 5, refit = True, return_train_score=True,verbose=10)\n",
    "\n",
    "clf_kn.fit(X_train, y_train)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 96,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.939397411660754"
      ]
     },
     "execution_count": 96,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "clf_kn.best_score_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 97,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Pipeline(memory=None,\n",
       "         steps=[('vectorizer',\n",
       "                 TfidfVectorizer(analyzer='word', binary=False,\n",
       "                                 decode_error='strict',\n",
       "                                 dtype=<class 'numpy.float64'>,\n",
       "                                 encoding='utf-8', input='content',\n",
       "                                 lowercase=True, max_df=1.0, max_features=None,\n",
       "                                 min_df=1, ngram_range=(1, 1), norm='l2',\n",
       "                                 preprocessor=None, smooth_idf=True,\n",
       "                                 stop_words=None, strip_accents=None,\n",
       "                                 sublinear_tf=False,\n",
       "                                 token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\n",
       "                                 tokenizer=<function tokenize at 0x7fa0f51f2d40>,\n",
       "                                 use_idf=True, vocabulary=None)),\n",
       "                ('clf',\n",
       "                 KNeighborsClassifier(algorithm='auto', leaf_size=30,\n",
       "                                      metric='minkowski', metric_params=None,\n",
       "                                      n_jobs=None, n_neighbors=15, p=2,\n",
       "                                      weights='uniform'))],\n",
       "         verbose=False)"
      ]
     },
     "execution_count": 97,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "clf_kn.best_estimator_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 98,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'clf__n_neighbors': 15}"
      ]
     },
     "execution_count": 98,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "clf_kn.best_params_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 99,
   "metadata": {},
   "outputs": [],
   "source": [
    "best_model_kn = clf_kn.best_estimator_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 100,
   "metadata": {},
   "outputs": [],
   "source": [
    "# predict using best estimator\n",
    "# do prediction\n",
    "y_pred_train = best_model_kn.predict(X_train)\n",
    "y_pred_test = best_model_kn.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 101,
   "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>Accuracy</th>\n",
       "      <th>f1 score</th>\n",
       "      <th>Precision</th>\n",
       "      <th>Recall</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.820650</td>\n",
       "      <td>0.820650</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.820650</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.893337</td>\n",
       "      <td>0.893337</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.893337</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.995518</td>\n",
       "      <td>0.995518</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.995518</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.741188</td>\n",
       "      <td>0.741188</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.741188</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.924613</td>\n",
       "      <td>0.924613</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.924613</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0.953087</td>\n",
       "      <td>0.953087</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.953087</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>0.973411</td>\n",
       "      <td>0.973411</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.973411</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>0.981663</td>\n",
       "      <td>0.981663</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.981663</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>0.968215</td>\n",
       "      <td>0.968215</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.968215</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>0.948095</td>\n",
       "      <td>0.948095</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.948095</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>0.917584</td>\n",
       "      <td>0.917584</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.917584</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>0.924358</td>\n",
       "      <td>0.924358</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.924358</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>0.986298</td>\n",
       "      <td>0.986298</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.986298</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>0.976467</td>\n",
       "      <td>0.976467</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.976467</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>0.988488</td>\n",
       "      <td>0.988488</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.988488</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>0.968368</td>\n",
       "      <td>0.968368</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.968368</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>0.958792</td>\n",
       "      <td>0.958792</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.958792</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>0.869448</td>\n",
       "      <td>0.869448</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.869448</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>0.936940</td>\n",
       "      <td>0.936940</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.936940</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>0.956398</td>\n",
       "      <td>0.956398</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.956398</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>0.953902</td>\n",
       "      <td>0.953902</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.953902</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>0.980236</td>\n",
       "      <td>0.980236</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.980236</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>0.993938</td>\n",
       "      <td>0.993938</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.993938</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>0.989405</td>\n",
       "      <td>0.989405</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.989405</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>0.996027</td>\n",
       "      <td>0.996027</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.996027</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26</th>\n",
       "      <td>0.988641</td>\n",
       "      <td>0.988641</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.988641</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>0.956958</td>\n",
       "      <td>0.956958</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.956958</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28</th>\n",
       "      <td>0.820650</td>\n",
       "      <td>0.820650</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.820650</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29</th>\n",
       "      <td>0.928280</td>\n",
       "      <td>0.928280</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.928280</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30</th>\n",
       "      <td>0.922015</td>\n",
       "      <td>0.922015</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.922015</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>31</th>\n",
       "      <td>0.989303</td>\n",
       "      <td>0.989303</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.989303</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>32</th>\n",
       "      <td>0.940403</td>\n",
       "      <td>0.940403</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.940403</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>33</th>\n",
       "      <td>0.980644</td>\n",
       "      <td>0.980644</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.980644</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>34</th>\n",
       "      <td>0.948350</td>\n",
       "      <td>0.948350</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.948350</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>35</th>\n",
       "      <td>0.855185</td>\n",
       "      <td>0.855185</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.855185</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    Accuracy  f1 score  Precision    Recall\n",
       "0   0.820650  0.820650        1.0  0.820650\n",
       "1   0.893337  0.893337        1.0  0.893337\n",
       "2   0.995518  0.995518        1.0  0.995518\n",
       "3   0.741188  0.741188        1.0  0.741188\n",
       "4   0.924613  0.924613        1.0  0.924613\n",
       "5   0.953087  0.953087        1.0  0.953087\n",
       "6   0.973411  0.973411        1.0  0.973411\n",
       "7   0.981663  0.981663        1.0  0.981663\n",
       "8   0.968215  0.968215        1.0  0.968215\n",
       "9   1.000000  1.000000        1.0  1.000000\n",
       "10  0.948095  0.948095        1.0  0.948095\n",
       "11  0.917584  0.917584        1.0  0.917584\n",
       "12  0.924358  0.924358        1.0  0.924358\n",
       "13  0.986298  0.986298        1.0  0.986298\n",
       "14  0.976467  0.976467        1.0  0.976467\n",
       "15  0.988488  0.988488        1.0  0.988488\n",
       "16  0.968368  0.968368        1.0  0.968368\n",
       "17  0.958792  0.958792        1.0  0.958792\n",
       "18  0.869448  0.869448        1.0  0.869448\n",
       "19  0.936940  0.936940        1.0  0.936940\n",
       "20  0.956398  0.956398        1.0  0.956398\n",
       "21  0.953902  0.953902        1.0  0.953902\n",
       "22  0.980236  0.980236        1.0  0.980236\n",
       "23  0.993938  0.993938        1.0  0.993938\n",
       "24  0.989405  0.989405        1.0  0.989405\n",
       "25  0.996027  0.996027        1.0  0.996027\n",
       "26  0.988641  0.988641        1.0  0.988641\n",
       "27  0.956958  0.956958        1.0  0.956958\n",
       "28  0.820650  0.820650        1.0  0.820650\n",
       "29  0.928280  0.928280        1.0  0.928280\n",
       "30  0.922015  0.922015        1.0  0.922015\n",
       "31  0.989303  0.989303        1.0  0.989303\n",
       "32  0.940403  0.940403        1.0  0.940403\n",
       "33  0.980644  0.980644        1.0  0.980644\n",
       "34  0.948350  0.948350        1.0  0.948350\n",
       "35  0.855185  0.855185        1.0  0.855185"
      ]
     },
     "execution_count": 101,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_results = []\n",
    "for i,column in enumerate(y_train.columns):\n",
    "    result = get_metrics(y_train.loc[:,column].values,y_pred_train[:,i])\n",
    "    train_results.append(result)\n",
    "train_results_df = pd.DataFrame(train_results)\n",
    "train_results_df    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 102,
   "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>Accuracy</th>\n",
       "      <th>f1 score</th>\n",
       "      <th>Precision</th>\n",
       "      <th>Recall</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.798625</td>\n",
       "      <td>0.798625</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.798625</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.872880</td>\n",
       "      <td>0.872880</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.872880</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.995569</td>\n",
       "      <td>0.995569</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.995569</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.698090</td>\n",
       "      <td>0.698090</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.698090</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.915355</td>\n",
       "      <td>0.915355</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.915355</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0.952788</td>\n",
       "      <td>0.952788</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.952788</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>0.971123</td>\n",
       "      <td>0.971123</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.971123</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>0.983040</td>\n",
       "      <td>0.983040</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.983040</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>0.969290</td>\n",
       "      <td>0.969290</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.969290</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>0.939037</td>\n",
       "      <td>0.939037</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.939037</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>0.913063</td>\n",
       "      <td>0.913063</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.913063</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>0.915814</td>\n",
       "      <td>0.915814</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.915814</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>0.986555</td>\n",
       "      <td>0.986555</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.986555</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>0.979679</td>\n",
       "      <td>0.979679</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.979679</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>0.989152</td>\n",
       "      <td>0.989152</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.989152</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>0.961956</td>\n",
       "      <td>0.961956</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.961956</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>0.955080</td>\n",
       "      <td>0.955080</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.955080</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>0.869977</td>\n",
       "      <td>0.869977</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.869977</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>0.928648</td>\n",
       "      <td>0.928648</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.928648</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>0.953400</td>\n",
       "      <td>0.953400</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.953400</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>0.945302</td>\n",
       "      <td>0.945302</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.945302</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>0.982429</td>\n",
       "      <td>0.982429</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.982429</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>0.993888</td>\n",
       "      <td>0.993888</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.993888</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>0.988541</td>\n",
       "      <td>0.988541</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.988541</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>0.993583</td>\n",
       "      <td>0.993583</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.993583</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26</th>\n",
       "      <td>0.986860</td>\n",
       "      <td>0.986860</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.986860</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>0.953247</td>\n",
       "      <td>0.953247</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.953247</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28</th>\n",
       "      <td>0.797555</td>\n",
       "      <td>0.797555</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.797555</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29</th>\n",
       "      <td>0.928189</td>\n",
       "      <td>0.928189</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.928189</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30</th>\n",
       "      <td>0.919175</td>\n",
       "      <td>0.919175</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.919175</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>31</th>\n",
       "      <td>0.990069</td>\n",
       "      <td>0.990069</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.990069</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>32</th>\n",
       "      <td>0.939649</td>\n",
       "      <td>0.939649</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.939649</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>33</th>\n",
       "      <td>0.979832</td>\n",
       "      <td>0.979832</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.979832</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>34</th>\n",
       "      <td>0.946066</td>\n",
       "      <td>0.946066</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.946066</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>35</th>\n",
       "      <td>0.841711</td>\n",
       "      <td>0.841711</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.841711</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    Accuracy  f1 score  Precision    Recall\n",
       "0   0.798625  0.798625        1.0  0.798625\n",
       "1   0.872880  0.872880        1.0  0.872880\n",
       "2   0.995569  0.995569        1.0  0.995569\n",
       "3   0.698090  0.698090        1.0  0.698090\n",
       "4   0.915355  0.915355        1.0  0.915355\n",
       "5   0.952788  0.952788        1.0  0.952788\n",
       "6   0.971123  0.971123        1.0  0.971123\n",
       "7   0.983040  0.983040        1.0  0.983040\n",
       "8   0.969290  0.969290        1.0  0.969290\n",
       "9   1.000000  1.000000        1.0  1.000000\n",
       "10  0.939037  0.939037        1.0  0.939037\n",
       "11  0.913063  0.913063        1.0  0.913063\n",
       "12  0.915814  0.915814        1.0  0.915814\n",
       "13  0.986555  0.986555        1.0  0.986555\n",
       "14  0.979679  0.979679        1.0  0.979679\n",
       "15  0.989152  0.989152        1.0  0.989152\n",
       "16  0.961956  0.961956        1.0  0.961956\n",
       "17  0.955080  0.955080        1.0  0.955080\n",
       "18  0.869977  0.869977        1.0  0.869977\n",
       "19  0.928648  0.928648        1.0  0.928648\n",
       "20  0.953400  0.953400        1.0  0.953400\n",
       "21  0.945302  0.945302        1.0  0.945302\n",
       "22  0.982429  0.982429        1.0  0.982429\n",
       "23  0.993888  0.993888        1.0  0.993888\n",
       "24  0.988541  0.988541        1.0  0.988541\n",
       "25  0.993583  0.993583        1.0  0.993583\n",
       "26  0.986860  0.986860        1.0  0.986860\n",
       "27  0.953247  0.953247        1.0  0.953247\n",
       "28  0.797555  0.797555        1.0  0.797555\n",
       "29  0.928189  0.928189        1.0  0.928189\n",
       "30  0.919175  0.919175        1.0  0.919175\n",
       "31  0.990069  0.990069        1.0  0.990069\n",
       "32  0.939649  0.939649        1.0  0.939649\n",
       "33  0.979832  0.979832        1.0  0.979832\n",
       "34  0.946066  0.946066        1.0  0.946066\n",
       "35  0.841711  0.841711        1.0  0.841711"
      ]
     },
     "execution_count": 102,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_results = []\n",
    "for i,column in enumerate(y_test.columns):\n",
    "    result = get_metrics(y_test.loc[:,column].values,y_pred_test[:,i])\n",
    "    test_results.append(result)\n",
    "test_results_df = pd.DataFrame(test_results)\n",
    "test_results_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 103,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5,1,'Result for Training Data')"
      ]
     },
     "execution_count": 103,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "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": [
    "train_results_df.plot()\n",
    "plt.title(\"Result for Training Data\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5,1,'Result for Test Data')"
      ]
     },
     "execution_count": 104,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "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": [
    "%matplotlib inline\n",
    "test_results_df.plot()\n",
    "plt.title(\"Result for Test Data\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 105,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Accuracy     0.937089\n",
       "f1 score     0.937089\n",
       "Precision    1.000000\n",
       "Recall       0.937089\n",
       "dtype: float64"
      ]
     },
     "execution_count": 105,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_results_df.mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 106,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Accuracy     0.942413\n",
       "f1 score     0.942413\n",
       "Precision    1.000000\n",
       "Recall       0.942413\n",
       "dtype: float64"
      ]
     },
     "execution_count": 106,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_results_df.mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 108,
   "metadata": {},
   "outputs": [],
   "source": [
    "# function for making bigram token\n",
    "def tokenize_bigram(text):\n",
    "    # Normalize text\n",
    "    text = re.sub(r\"[^a-zA-Z0-9]\", \" \", text.lower())\n",
    "    # Tokenize text\n",
    "    words = word_tokenize(text)\n",
    "    words = bigrams(words)\n",
    "    return words"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 109,
   "metadata": {},
   "outputs": [],
   "source": [
    "# compute bag of word counts and tf-idf values\n",
    "vectorizer_bigram = TfidfVectorizer( tokenizer=tokenize_bigram ,use_idf=True, smooth_idf=True, sublinear_tf=False)\n",
    "\n",
    "# clf = MultiOutputClassifier(RandomForestClassifier(random_state = 42))\n",
    "clf_bigram = RandomForestClassifier(random_state = 42)\n",
    "\n",
    "pipeline_bigram = Pipeline([('vectorizer',vectorizer_bigram), ('clf',clf_bigram)])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 110,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 5 folds for each of 20 candidates, totalling 100 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Using backend LokyBackend with 8 concurrent workers.\n",
      "[Parallel(n_jobs=-1)]: Done   2 tasks      | elapsed:  1.1min\n",
      "[Parallel(n_jobs=-1)]: Done   9 tasks      | elapsed:  2.3min\n",
      "[Parallel(n_jobs=-1)]: Done  16 tasks      | elapsed:  2.5min\n",
      "[Parallel(n_jobs=-1)]: Done  25 tasks      | elapsed:  4.8min\n",
      "[Parallel(n_jobs=-1)]: Done  34 tasks      | elapsed:  6.4min\n",
      "[Parallel(n_jobs=-1)]: Done  45 tasks      | elapsed:  8.4min\n",
      "[Parallel(n_jobs=-1)]: Done  56 tasks      | elapsed: 10.0min\n",
      "[Parallel(n_jobs=-1)]: Done  69 tasks      | elapsed: 12.2min\n",
      "[Parallel(n_jobs=-1)]: Done  82 tasks      | elapsed: 14.7min\n",
      "[Parallel(n_jobs=-1)]: Done  96 out of 100 | elapsed: 17.9min remaining:   44.8s\n",
      "[Parallel(n_jobs=-1)]: Done 100 out of 100 | elapsed: 18.6min finished\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "GridSearchCV(cv=5, error_score=nan,\n",
       "             estimator=Pipeline(memory=None,\n",
       "                                steps=[('vectorizer',\n",
       "                                        TfidfVectorizer(analyzer='word',\n",
       "                                                        binary=False,\n",
       "                                                        decode_error='strict',\n",
       "                                                        dtype=<class 'numpy.float64'>,\n",
       "                                                        encoding='utf-8',\n",
       "                                                        input='content',\n",
       "                                                        lowercase=True,\n",
       "                                                        max_df=1.0,\n",
       "                                                        max_features=None,\n",
       "                                                        min_df=1,\n",
       "                                                        ngram_range=(1, 1),\n",
       "                                                        norm='l2',\n",
       "                                                        preprocessor=None,\n",
       "                                                        smooth_idf=True,\n",
       "                                                        stop_words=None,\n",
       "                                                        stri...\n",
       "                                                               oob_score=False,\n",
       "                                                               random_state=42,\n",
       "                                                               verbose=0,\n",
       "                                                               warm_start=False))],\n",
       "                                verbose=False),\n",
       "             iid='deprecated', n_jobs=-1,\n",
       "             param_grid={'clf__criterion': ['gini', 'entropy'],\n",
       "                         'clf__max_depth': [5, 10, 20, 30, 40],\n",
       "                         'clf__max_features': ['auto', 'sqrt'],\n",
       "                         'clf__random_state': [42]},\n",
       "             pre_dispatch='2*n_jobs', refit=True, return_train_score=True,\n",
       "             scoring=make_scorer(multi_class_score), verbose=10)"
      ]
     },
     "execution_count": 110,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "parameters = {'clf__criterion':['gini','entropy'],\n",
    " 'clf__max_features': ['auto', 'sqrt'],\n",
    " 'clf__max_depth': [5,10, 20, 30, 40],\n",
    "              'clf__random_state': [42]}\n",
    "\n",
    "cv_bigrams = GridSearchCV(pipeline_bigram, param_grid=parameters, scoring=score,\n",
    "                  n_jobs = -1, cv = 5, refit = True, return_train_score=True,verbose=10)\n",
    "cv_bigrams.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 111,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9283921174844432"
      ]
     },
     "execution_count": 111,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cv_bigrams.best_score_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 112,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9283921174844432"
      ]
     },
     "execution_count": 112,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cv_bigrams.best_score_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 113,
   "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>Accuracy</th>\n",
       "      <th>f1 score</th>\n",
       "      <th>Precision</th>\n",
       "      <th>Recall</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.820650</td>\n",
       "      <td>0.820650</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.820650</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.893337</td>\n",
       "      <td>0.893337</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.893337</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.995518</td>\n",
       "      <td>0.995518</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.995518</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.741188</td>\n",
       "      <td>0.741188</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.741188</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.924613</td>\n",
       "      <td>0.924613</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.924613</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0.953087</td>\n",
       "      <td>0.953087</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.953087</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>0.973411</td>\n",
       "      <td>0.973411</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.973411</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>0.981663</td>\n",
       "      <td>0.981663</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.981663</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>0.968215</td>\n",
       "      <td>0.968215</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.968215</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>0.948095</td>\n",
       "      <td>0.948095</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.948095</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>0.917584</td>\n",
       "      <td>0.917584</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.917584</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>0.924358</td>\n",
       "      <td>0.924358</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.924358</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>0.986298</td>\n",
       "      <td>0.986298</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.986298</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>0.976467</td>\n",
       "      <td>0.976467</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.976467</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>0.988488</td>\n",
       "      <td>0.988488</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.988488</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>0.968368</td>\n",
       "      <td>0.968368</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.968368</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>0.958792</td>\n",
       "      <td>0.958792</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.958792</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>0.869448</td>\n",
       "      <td>0.869448</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.869448</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>0.936940</td>\n",
       "      <td>0.936940</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.936940</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>0.956398</td>\n",
       "      <td>0.956398</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.956398</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>0.953902</td>\n",
       "      <td>0.953902</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.953902</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>0.980236</td>\n",
       "      <td>0.980236</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.980236</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>0.993938</td>\n",
       "      <td>0.993938</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.993938</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>0.989405</td>\n",
       "      <td>0.989405</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.989405</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>0.996027</td>\n",
       "      <td>0.996027</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.996027</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26</th>\n",
       "      <td>0.988641</td>\n",
       "      <td>0.988641</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.988641</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>0.956958</td>\n",
       "      <td>0.956958</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.956958</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28</th>\n",
       "      <td>0.820650</td>\n",
       "      <td>0.820650</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.820650</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29</th>\n",
       "      <td>0.928280</td>\n",
       "      <td>0.928280</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.928280</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30</th>\n",
       "      <td>0.922015</td>\n",
       "      <td>0.922015</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.922015</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>31</th>\n",
       "      <td>0.989303</td>\n",
       "      <td>0.989303</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.989303</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>32</th>\n",
       "      <td>0.940403</td>\n",
       "      <td>0.940403</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.940403</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>33</th>\n",
       "      <td>0.980644</td>\n",
       "      <td>0.980644</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.980644</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>34</th>\n",
       "      <td>0.948350</td>\n",
       "      <td>0.948350</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.948350</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>35</th>\n",
       "      <td>0.855185</td>\n",
       "      <td>0.855185</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.855185</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    Accuracy  f1 score  Precision    Recall\n",
       "0   0.820650  0.820650        1.0  0.820650\n",
       "1   0.893337  0.893337        1.0  0.893337\n",
       "2   0.995518  0.995518        1.0  0.995518\n",
       "3   0.741188  0.741188        1.0  0.741188\n",
       "4   0.924613  0.924613        1.0  0.924613\n",
       "5   0.953087  0.953087        1.0  0.953087\n",
       "6   0.973411  0.973411        1.0  0.973411\n",
       "7   0.981663  0.981663        1.0  0.981663\n",
       "8   0.968215  0.968215        1.0  0.968215\n",
       "9   1.000000  1.000000        1.0  1.000000\n",
       "10  0.948095  0.948095        1.0  0.948095\n",
       "11  0.917584  0.917584        1.0  0.917584\n",
       "12  0.924358  0.924358        1.0  0.924358\n",
       "13  0.986298  0.986298        1.0  0.986298\n",
       "14  0.976467  0.976467        1.0  0.976467\n",
       "15  0.988488  0.988488        1.0  0.988488\n",
       "16  0.968368  0.968368        1.0  0.968368\n",
       "17  0.958792  0.958792        1.0  0.958792\n",
       "18  0.869448  0.869448        1.0  0.869448\n",
       "19  0.936940  0.936940        1.0  0.936940\n",
       "20  0.956398  0.956398        1.0  0.956398\n",
       "21  0.953902  0.953902        1.0  0.953902\n",
       "22  0.980236  0.980236        1.0  0.980236\n",
       "23  0.993938  0.993938        1.0  0.993938\n",
       "24  0.989405  0.989405        1.0  0.989405\n",
       "25  0.996027  0.996027        1.0  0.996027\n",
       "26  0.988641  0.988641        1.0  0.988641\n",
       "27  0.956958  0.956958        1.0  0.956958\n",
       "28  0.820650  0.820650        1.0  0.820650\n",
       "29  0.928280  0.928280        1.0  0.928280\n",
       "30  0.922015  0.922015        1.0  0.922015\n",
       "31  0.989303  0.989303        1.0  0.989303\n",
       "32  0.940403  0.940403        1.0  0.940403\n",
       "33  0.980644  0.980644        1.0  0.980644\n",
       "34  0.948350  0.948350        1.0  0.948350\n",
       "35  0.855185  0.855185        1.0  0.855185"
      ]
     },
     "execution_count": 113,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_results = []\n",
    "for i,column in enumerate(y_train.columns):\n",
    "    result = get_metrics(y_train.loc[:,column].values,y_pred_train[:,i])\n",
    "    train_results.append(result)\n",
    "train_results_df = pd.DataFrame(train_results)\n",
    "train_results_df    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 114,
   "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>Accuracy</th>\n",
       "      <th>f1 score</th>\n",
       "      <th>Precision</th>\n",
       "      <th>Recall</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.798625</td>\n",
       "      <td>0.798625</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.798625</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.872880</td>\n",
       "      <td>0.872880</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.872880</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.995569</td>\n",
       "      <td>0.995569</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.995569</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.698090</td>\n",
       "      <td>0.698090</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.698090</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.915355</td>\n",
       "      <td>0.915355</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.915355</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0.952788</td>\n",
       "      <td>0.952788</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.952788</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>0.971123</td>\n",
       "      <td>0.971123</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.971123</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>0.983040</td>\n",
       "      <td>0.983040</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.983040</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>0.969290</td>\n",
       "      <td>0.969290</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.969290</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>0.939037</td>\n",
       "      <td>0.939037</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.939037</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>0.913063</td>\n",
       "      <td>0.913063</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.913063</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>0.915814</td>\n",
       "      <td>0.915814</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.915814</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>0.986555</td>\n",
       "      <td>0.986555</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.986555</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>0.979679</td>\n",
       "      <td>0.979679</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.979679</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>0.989152</td>\n",
       "      <td>0.989152</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.989152</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>0.961956</td>\n",
       "      <td>0.961956</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.961956</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>0.955080</td>\n",
       "      <td>0.955080</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.955080</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>0.869977</td>\n",
       "      <td>0.869977</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.869977</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>0.928648</td>\n",
       "      <td>0.928648</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.928648</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>0.953400</td>\n",
       "      <td>0.953400</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.953400</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>0.945302</td>\n",
       "      <td>0.945302</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.945302</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>0.982429</td>\n",
       "      <td>0.982429</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.982429</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>0.993888</td>\n",
       "      <td>0.993888</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.993888</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>0.988541</td>\n",
       "      <td>0.988541</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.988541</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>0.993583</td>\n",
       "      <td>0.993583</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.993583</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26</th>\n",
       "      <td>0.986860</td>\n",
       "      <td>0.986860</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.986860</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>0.953247</td>\n",
       "      <td>0.953247</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.953247</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28</th>\n",
       "      <td>0.797555</td>\n",
       "      <td>0.797555</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.797555</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29</th>\n",
       "      <td>0.928189</td>\n",
       "      <td>0.928189</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.928189</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30</th>\n",
       "      <td>0.919175</td>\n",
       "      <td>0.919175</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.919175</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>31</th>\n",
       "      <td>0.990069</td>\n",
       "      <td>0.990069</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.990069</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>32</th>\n",
       "      <td>0.939649</td>\n",
       "      <td>0.939649</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.939649</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>33</th>\n",
       "      <td>0.979832</td>\n",
       "      <td>0.979832</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.979832</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>34</th>\n",
       "      <td>0.946066</td>\n",
       "      <td>0.946066</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.946066</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>35</th>\n",
       "      <td>0.841711</td>\n",
       "      <td>0.841711</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.841711</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    Accuracy  f1 score  Precision    Recall\n",
       "0   0.798625  0.798625        1.0  0.798625\n",
       "1   0.872880  0.872880        1.0  0.872880\n",
       "2   0.995569  0.995569        1.0  0.995569\n",
       "3   0.698090  0.698090        1.0  0.698090\n",
       "4   0.915355  0.915355        1.0  0.915355\n",
       "5   0.952788  0.952788        1.0  0.952788\n",
       "6   0.971123  0.971123        1.0  0.971123\n",
       "7   0.983040  0.983040        1.0  0.983040\n",
       "8   0.969290  0.969290        1.0  0.969290\n",
       "9   1.000000  1.000000        1.0  1.000000\n",
       "10  0.939037  0.939037        1.0  0.939037\n",
       "11  0.913063  0.913063        1.0  0.913063\n",
       "12  0.915814  0.915814        1.0  0.915814\n",
       "13  0.986555  0.986555        1.0  0.986555\n",
       "14  0.979679  0.979679        1.0  0.979679\n",
       "15  0.989152  0.989152        1.0  0.989152\n",
       "16  0.961956  0.961956        1.0  0.961956\n",
       "17  0.955080  0.955080        1.0  0.955080\n",
       "18  0.869977  0.869977        1.0  0.869977\n",
       "19  0.928648  0.928648        1.0  0.928648\n",
       "20  0.953400  0.953400        1.0  0.953400\n",
       "21  0.945302  0.945302        1.0  0.945302\n",
       "22  0.982429  0.982429        1.0  0.982429\n",
       "23  0.993888  0.993888        1.0  0.993888\n",
       "24  0.988541  0.988541        1.0  0.988541\n",
       "25  0.993583  0.993583        1.0  0.993583\n",
       "26  0.986860  0.986860        1.0  0.986860\n",
       "27  0.953247  0.953247        1.0  0.953247\n",
       "28  0.797555  0.797555        1.0  0.797555\n",
       "29  0.928189  0.928189        1.0  0.928189\n",
       "30  0.919175  0.919175        1.0  0.919175\n",
       "31  0.990069  0.990069        1.0  0.990069\n",
       "32  0.939649  0.939649        1.0  0.939649\n",
       "33  0.979832  0.979832        1.0  0.979832\n",
       "34  0.946066  0.946066        1.0  0.946066\n",
       "35  0.841711  0.841711        1.0  0.841711"
      ]
     },
     "execution_count": 114,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_results = []\n",
    "for i,column in enumerate(y_test.columns):\n",
    "    result = get_metrics(y_test.loc[:,column].values,y_pred_test[:,i])\n",
    "    test_results.append(result)\n",
    "test_results_df = pd.DataFrame(test_results)\n",
    "test_results_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 115,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5,1,'Result for Training Data')"
      ]
     },
     "execution_count": 115,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "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": [
    "train_results_df.plot()\n",
    "plt.title(\"Result for Training Data\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 116,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5,1,'Result for Test Data')"
      ]
     },
     "execution_count": 116,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYAAAAEICAYAAABWJCMKAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi41LCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvSM8oowAAIABJREFUeJzsnXd4XNW1t991pqtZ1U2yLfeGjW1MSSBA6HCpaZCQG+AGCCQkpBJICC0JLXDJTULgIwkhkAQChBZCDb2Dce/dlmRbVq+jaWd/f5wjI9kqM6MZTdF+n0ePR+fsskaWZp2991q/JUopNBqNRjPyMFJtgEaj0WhSg3YAGo1GM0LRDkCj0WhGKNoBaDQazQhFOwCNRqMZoWgHoNFoNCMU7QA0GYeIHCsi1TH2uVxEakWkXURKkmWbRpNJaAegGRIisl1E/PYH6x4ReUBE8lJgwwkD3HcB/wucpJTKU0o1DHG+9h5fZo/33y4i5w9h3PdF5KsD3J8lIqrHXHtE5BkR+WwMc1wmIv+J10ZNdqEdgCYRnKGUygMWAAuBa1Jsz/6MAbzAmlg7ikWvvxPbieTZ73kn9vu3v/6WGJP7JdJj7oXAm8CzInJekufVZCHaAWgShlJqD/AiliMAQEQ8InKHiOy0t2DuFRGffa9URJ4VkWYRaRSRt7o/bO0n3Wk9xnlARH6x/5wi8hAwEfiX/VR81X73ZwAb7G+bReRV+/qnReQjEWmx//10jz6vi8gvReQdoBOYEsvPQUQcIvIzEdkqIvUi8jcRKbTv5YrII/b7bRaRD0SkSETuBA4F/mi/jzsHm0cptVspdQdwC/CrHvNfJyLbRKRNRFaLyH/Z1xcCvwaO7V5B2NfPEZEVItJq/z/9JJb3q8lctAPQJAwRqQBOBTb3uHwrMAPLKUwDyoHr7Hs/AKqBMqyn9J8AMWmTKKX+m95P4bfvd38jMNf+tlApdZyIFAP/Bn4DlGBtD/17v7OB/wYuBfKBHbHYBPwQOAk4CqgAQsBd9r2LASfWz6EUuAIIKqV+AHwEXGy/jx/EMN8TQIWITLa/3wB8GhgF3AY8IiKlSqllwHeB1+05xtrtW4GvAIXA2cAPReSUGN+zJgPRDkCTCJ4SkTagCtgLXA/W9gnWh+j3lFKNSqk24Gage7siBIwDJimlQkqpt9TwiFP9F7BJKfWQUiqslHoYWA+c0aPNA0qpNfb9UIzjXwZcrZTapZTqAm4EzrV/HiEshzfVHvsjpVTHEN/PLvvfYgCl1D/s1YGplHoIqAEO6a+zUuoV+72aSqmlwKPAMUO0SZMBaAegSQRnK6XygWOBWVhPtmB90OUAH9vbHc3AC/Z1sLYtNgMv2dslVw+TveM58Kl+B9ZTeTdV8Qxsf8hPAJ7r8Z6XYf2tlQB/At4AHheRahG5WUQc8czVg267G20bvi4iK3vMP41P/k/6svlIEXlDROpEpAW4cKD2muxBOwBNwlBKvQE8ANxhX6oH/MBcpVSh/TXKPsBEKdWmlPqBUmoKcCbwfRE53u7bieU8uhlL/8S6atgFTNrv2kSsJ+V4x7Q6WSuYGuC4Hu+5UCnlVUrVK6UCSqnrlFKzgKOBL/LJiije1c85QLVSapt95vFbrJVXsVKqEMvJygBzPAr8A5iglBqF9X8ofbTTZBnaAWgSza+BE0XkYKWUCfwBuEtERgOISLmInGy/Pl1EptlPzS1ABDDtcZYDX7EPVE9h4C2JWmI7qH0OmCEiXxERp4icC8wBno1hjIG4F7hVRCYAiMhoETnDfn2CiMyxD7tbgTCfvOeY3oeIjBWR72FFXXWvnvLs8eoAQ0Quw1oBdFMLTBArNLZ7xZIHNCiluuzD8C/G86Y1mYd2AJqEopSqAx7kk4PeH2M9gb4vIq3Af4CZ9r3p9vftwHvA75VSr9n3rsTak28GzgeeGmDaW4Br7S2PH0ZhYwNwOtYhdANwFXC6Uqo+2vc5CLdjva9X7bORd4FF9r1y4GmgDViN5Yz+Yd+7C/iaiDSJyO30jcOO4OkAVgDHA2d1h5/ae/j3AkuA3cBk+3U3LwDbgb0iUm2vWC4D7rBtvQp4bIjvX5MhiC4Io9FoNCMTvQLQaDSaEYp2ABqNRjNC0Q5Ao9FoRijaAWg0Gs0IxZlqA/antLRUVVZWptoMjUajySg+/vjjeqVU2eAtPyHtHEBlZSVLliwZvKFGo9Fo9iEisWpW6S0gjUajGaloB6DRaDQjFO0ANBqNZoSiHYBGo9GMULQD0Gg0mhHKoA5ARO4Xkb0isrqf+yIivxGRzbYG+aIe9y4QkU321wWJNFyj0Wg0QyOaFcADwEDl4U7FUnWcjqVBfg+AXXbveuBw4DDgehEpGoqxGo1Go0kcg+YBKKXeFJHKAZqcBTxoy8q+LyKFIjIOqzrUy0qp7ipFL2M5kocHmm9763YueuGi6KzXRIUEw8x4bTObjpmK6XWl2hyNRpMmJOIMoJze5fOq7Wv9XT8AEblURJaIyJJQMNbyq5rBmPnKJs56vpkp72xLtSkajSaNSItMYKXUfcB9AIsXL1Z/PuXPKbYoewh2tLP1J4tRCKdvgoN/pX+2Gk028gAPxNwnESuAGqwi2N1U2Nf6uz4shPydLLv08zTtHNlPvWtu/wmqUxCvwlPTmGpzNBpNGpEIB/AMVhk7EZEjgBal1G7gReAkESmyD39Psq8NC5uf/jveN9ey/S+/Ha4p045IKITv+ZeRQqHz0wej2mDvmuWpNkuTJjRsWs/SH32dLc89PuSxOhvrWffQPbTX7UmAZZrhYtAtIBF5GOtAt1REqrEie1wASql7sWqanoZV97UTuMi+1ygiPwc+soe6qftAeDgIbF6PB1BtbcM1Zdqx6q4b8bRC1/mnkDtzPuarK9n19MOMnrsg1aZpUszGxx9E3Xwzvk4h+K93Wf/z6+hcfDAVl/046t+PYEc76+//P3j5WdxbmyAsVP/q1/gPmcGEa26ldMbcJL8LzVBJu5rAixcvVolQA136zXPxvboS/wkHs+h3jyTAsszCjETY8Jn5SDDC1LeWokyTrYctpmvuOBY8+trgA2iylmXXXIbv6dcRL4QuvZjghlX4PlwCDSaIwhzvJXjsccy4/BpyS3urC4cDATb89V7Cz/0Tz6ZaVNBAXIrglBIihxyG6+03cOz0Iw5FYPYYir93HROOPD5F73RkISIfK6UWx9InLQ6Bk4GjrtZ64fen1pAUsfYPd+FoNPGfdTQuXw4AkdEevFW1KbYseax/5E/kjJ/IxKNPTLUpaUlncyObLjoT77oGIuM8jL3vH5RMn7Xv/rZX/k3zg3fjW7UV79+ep+ofzxGYVoLzjC/hyM2j66m/41tfg9EleByKUOUo1PGnMuPi7+EtKOw1Tutvb8G9tpb2i7/Fqsp83P9zBTO/mPm5oCvOPgoM4eAn3kq1KQkha1cAq09chKPKT2BRBQv+/nICLMss1h47H6MlSOUb7+/741z69bPxvbOesS89T9HEySm2MHE0bdtE1XcvwLWhCclTTHljCe7cvFSblVbsWvIuLd+5BBpNug6fzLz7nsTp8fTZNhIKse7B32E+8zjuLQ2osFg3DEWkIofw0ccx7Rs/JK9s7IBz1q5ayq7bf4pv+TZUSFBjXYQ/fy4HffNqDIcj0W8x6ZiRCBsXzAUTJr7zLjmFxak2qRfxrACyVgvI2dYFgHR1pdiS4Wfd3/+A7AnRedTCXk9m7k8fAwhVT/0tdcbtx9JLzmHpN8+lpaZq8MZ9sPy2a6g95wxcGxuJTM5FtQurb7gywVZmNqvvvYPW/7kIaYsQvPgLLPzLc/1++AM4XC4O+vr3mP/0O0x4/S26vnoaXV88gbEvPM9BLy1lwbV3DPrhDzBm3iIWPvQ84154Gf8JB+NoCeK6+6+sO2ERZiSSyLc4LFS/86rlyCLC1of/kGpzEkLWOgA6TACMYDDFhgw/kT/fg7hNpl97Z6/rk8/6qvUU99E7KbKsN007t+F7az2+V1ey+5QTWXbBf1G3flVUfWtXLWX1KYvx/PkplM/AedsNzHn2AxjtJPfld3Q0CtYT69LLvoDj//4IPgPfb+/i4B/+PKYxckvLWHjtnSz8+W/jXjWOKp/Aot89QuVbH9H1qWkYu4Os/dOv4xorlTS88fy+1+E3XkqhJYkjKx1AS00VKmQtWyWQmZnFrburWfm/1xMOBGLqt+W5x3FU+ek8dDb5Y8b3updbWgbFTrzb43vaTjQ1z/0TAP/x8wmX5+H9YAsNn/8iK754LNXvv9lnHzMSYem136LpK1/BubMd/2cPYtqrS5h+5nnWtsI3rsDsEjb+9PLhfCtpR0tNFWtPOwzf62sITy6g/JlXmHTsqSm1yZOXz8y7/oJ4THgsfVah0SJrVyIOhTnejXfLrlSbkxCy0gHUr/zkDMEIhVNoSXxUvfMKu84+Edd9j7Luc58h2NEedd/O3/8KcSomX/urPu93VZZDQ5jOxvpEmRs3wWXvATD5R7dw0AtL8P3+fwlML8G9Zg9tF13KqtOPYOuLT+1rX/3eG6w7YRG+x1/FLHLh+/1dLLrnsX2H3ACzz/8GkYk+ct5dR8Om9cP+ntIBMxJh13mn4tjZgf+kRcz913sHPAykipzCYvwLp+Oo6mTnm5l1NuepqUUVOQksWIhqgx1ZsArISgfQsdHaRhCXQoKZtde4+t476Lj8m+A3CR40GueWNjae/ZmoPrCr3nkF55YWug6eSNHk6X22cRx6JJjCtidT/wTm3r4dyVMUVU4BoPK40zj4qXco/NtfCCyowLW9mcCV17D6pEUsvfhs2i/+Bo66AP7TP8XMV5b2+0Sb+8MbUBGouu6K4Xw7acOmJ/4KdRH8Jx/Got/8Le0OXMf/4CYwoPHuW1JtStSE/J1IQ4hA+WjKPm9FMzU88eCQxmzcsoHq995IhHlxk5UOIFy1HQBV6ETCZmqNiRIzEmHplV/dt1+be8+9HPz4G/hP/xSOKj/bP3ccrburBxyj8c4bEIFxV9/ab5sJ5/w3oAi8l/pcAGddB6ExBQdcH7fwcBY8/B9GP/00/iNn4NzTge/tDUTGeil44H4W3XE/Dlf/qqaTTzqT4MwSPCuqqfkwO8L1YqHr4T8iDsXUH/0y1ab0yZh5iwhPKcS7Zlfch//Dzc7XX0RFBDV7HhOO/CxSAJ5V0Z1X9ceuy79M+2XfSOlqPCsdgOzZBYYiVFSAhNIrzLUvulqbWf25o/G9+DGRCbmMf+plJhz5WQAW3XE/XeedgtQGqfnCyTRt29TnGHtWfox7fR2B2WMYM29Rn20AiiZORgoNPFtSq5G0d+1KVKcQmjK13zbFU2ey6E9PU/7yazh/dSOzX/qY8Ys/HdX4ZT+7AxFo+MVViTI5I2ir3YVnYx3BaSWMKp8weIcU4b7gclRY2HL7Nak2JSpa3ra2q4qOPhmArqkVGHsCcQcb1K1fhaOqExUQNtz+04TZGStZ6QCcTU1IDpg+LyrNjwBqVy1l22lH4drQSNenpjH73+9TMK6iV5uFN/ya0CXnIk0Ras87k9pVSw8YZ8+tV4OC4u/fMOicgQljcNQFCPk7E/U2YmbPS9bevnfxUYO2zR8znulnfCmmrYzxh3yKrgUTcW1sYvOzQ9e6yRQ233UjKiy4vvS1VJsyIDO/eAGUOch5Z0nMgQ6pwLFhDeJSTPiMlWToPO5UMIWtf703rvGq7r0dlCA+he+1N1MWFpudDqC1AzPXjfL6UGHSNuZ4w2N/oflrX0GawgQuOpuFf/5Xv1sb879/I+HvXgIdiqaLvtIrSqZxywa8K6sITy3ct3IYCLVwMSosbH32sYS9l1iJrPgIRDHxtM8nbY5Jv7gbcUHg1+m5FZIMvG+/jYyCmeddnGpTBiVw6qmodmHt3Ten2pRBce+qxyxx7/v7nP7lSxCnQr39alzj+T5cihQJ/lM/i2qBtff/XyLNjZqsdABGe5jwqDyUzwdK8DelnwzyshuuRF1/C8oA522/YMGPBz8QO+jSH6Cu/SEEoOPyS9n2yr8B2PHLH6HCQu63r45q7nFnfBmAjjdfiP8NDBH3ziqkQKJKKIqXosnT6TxqHkZ1F2sf+F3S5kkXtjz3ONSbdB6+KO0OfvtizpXXY/gUzmeeTLUpA9LV2gzNEQIV4/Zd8+TlEy7PxbNtb8wPmNte+Tc0mnQuPpiZV/0S8Sh47K+JNjsqss4BBDvaMbsgUlwMOZYcQGd9eiUFLbvgNLyPvIRZ5qbskaeYdvoXou47+7yLcdx8E8qEwPe/z9q/3I1vyQYiE3OYcvLZUY0xZt4iJE/h3rAh3rcwJMxIBEd9F8GxyU+ln/XLezB8CvnjvWm7EkwUbQ/cDYai8srrUm1KVLhz8+g4Yh6yJ5QQSepksfOlZ8AUZO7CXtdDCxejOoVtLz0T03jND94Noii/9EdWWOyCqTh2dlL1zvAHZmSdA9i7aikogXEViK0H09WwN8VWfcLSqy7B+8E2grNLmPbcu73EuKJl+hlfwvd/d4FDkFt+iwoauC7+dkxjBMtLcO7pIBIa/kS5XR+9gwoI4ekzkz5XTnEpnScfA/URVt55fdLnSxWdjfV41+0mPLmA4qnJ/7kmislX3Yw4FO33pW9mcPv7rwNQ+tnTel0fa2+ztTwzYJnzXkRCIXJWbcUc790XrDHuBzeBQMPvhn+rMuscQMu6FQC4Jk3ByLdCDLsaG1Jp0j7W/f0P+J59C7Pcw9xHXhmSYNmkY0+l4L77kQIhMtHHzC9dGFP/8EEHo4IGO197fvDGCab+VWvrKvfwY4ZlvoOuvwspAO/j/0zpwXcy2fCbX6BCgpx9XqpNiYmiydMJzCzDs6k+bRP3nJs2IB7F+EOP7HV93IJDkSLBu3Zd1GOt/+u9mJ1C8JhPzurGzj+E8JRReFdXDxrqnWiyzgEEt1thkvmz5uHIHwVAuDn1DqB21VKM2+9EcmD8n54YUIwrWsYv/jTT31nJzKffjrlvmX342vTy00O2I1bU2uVgKCadfNawzOfy5dD1hc+jWmH1jd8bljmHG89rryB5itkXxrYSTAfyv/EDVETY+avUhUMOhLu2iUipt89zFf+MSmRviObqHVGNFX76H4hTMeObP+l13fnfl6BCwuZfXZsQm6Ml6xyA2mUllpTNPxTHKGuPOdSc2kPgQHsbjd/8GioErpt+sS/zNRE4XK5eUgjRUn7EsRhehWvd6oTZEi2e6t1Q6OilVJps5v/gRih1kPPiG3TU1w3bvMPBjtefR2rDdC4+aMAEuXRlyslnY45z4/toTUyyJ8NBe90eVItJcGLfORXeE84EJWx/8PeDjtXZWI9nSz3BKcUHFNqZfd7FUGKQ89YHw7otm3UOwNFQj3gUOcWluItKATDbWlNq0/qLzoC6CF1fPp2pp0V/4JtMDIeD4PgC3DUtw3o4GgmFMBqCBMeXDtucYL1fdfFlmH5hw8++OaxzJ5umP9xlHSp+6yeDN05TImd/HuUX1vxveh1gV73wJCjBMf+QPu9P+8LXELeJ8cHgGecbf38rKiQYp/cd+tx18kmoNlhz7+1DsjkWss4BOJtbUbnWUs1TVAKAak+dA1j2k8txr6olcMgEFlx7R8rs6IvQrDmYfmHXMMpD73zzJVRIiMwc/nqxcy68ArPCS87bq9jw6APDPn8yCLS34V29k8iE3AEzwNOdud+8BskH74svptqUXvhtKZGxJ/UdYefy5RCaMAr3jsZBn9xdr7+C4VPMuqBvjao5378Jw6twPDl8+TlZ5wAc7UEiBT4AcsrGAKA6U7Os3PDoA3ifeg1znJu5f3xq8A7DTMFxpwNQ9/zwheA1v2n9gY86KjVlG3OuutGqe3vdbaw9dj6r770jo8ND1919MyogmKefk2pThoTD5aLzyEOh3mTdI39MtTn7cG3dguFTjJ67oN824cVHoLqELf/6R79t6tavwqjx0zl3Ur/nf568fDoWz8LYFdiX45NsssoBmJEIdCjCRdbecrcDIAUOYO/alXDLrYgPxv3x0bj26ZPN5JPPQlwKY+WB0hLJQtatQpyKScedNnjjJDD5pDMZ/+Ir+I+fj6M5iOPXf2LDZ+az/JarM0KSYH9cLz6P4VPM+caPUm3KkJn245sRlyL80H2pNmUfrr2thEbnDthmwvnfABTtA+QydEs/FJx/2cBj/fDnYCha7+1bzj3RZJUDaNq6CRURzNFWdqm3oBAMhfiHtyxksKOd+svPRwXBcf0NaRuX7XC5CI/14akZvigpz646zGJXQqKg4qVgXAWL7v4Hk958n66zj0aCETx/eZrNRy1g2U8uJ9DeljLbYqH6/TcxdnXRuWB6Sn+eiaJgXAVdB1Xg3NrK7uUfpdocmnZuQ7VBaFLlgO1KZ8yFUge+9X0LNcIn0g9TTx14pVY2ax6h6cV41u2haWfyBRuzygE0rv4YAKO8ct81cSqMLv+w2rH2f85EasP4zz2V6Wemd1x2cNp0VCtRl2Ic0lwd7UhTmGB58uQfYsFbUMjCW/8f095ZQeBrp6OcgveJ19n+mUNZ+u2v0Fab3lWf6uzDwtGXZ4/iackVPwUFe+5KfdJed8U616LDB23rnzkd6iPUb1xzwL2e0g/R4Pv6d1BhYduvkn+on1UOoGOTlZCRO332vmviFGQYl/bLfvZtPCt2E1hYzsLr7hq2eePFd8xJANQ8/fekz7XjlWctTfU585I+Vyw4PR4W/ORXzHx7FeFvX0CkwI3v5WXs/HJqtqmiIRwIkLNiM2a5l/LDPpNqcxLGhCM/S2RiDr5lW1Jeta67Yl35KYNH7uWd+jlAqPr7gdtXPaUfomH6meehxjjJeXdZ0rclo3IAInKKiGwQkc0icoDimIhMEpFXRGSliLwuIhU97kVEZLn9FZtoRoyYNdsBKD7ok5At5RKM4PDE1W565hF8T7yMGuNi7v1JfasJY8rp5yIOBcs+TPpc7e9Yyoklx6S2Nm1/GA4H8751NXPeWElw3hgctV3Ddi7QUlPF+sNns/rERWx7+V+Dtl97722YfiF8cvo6qXiRcy9ABQ023HVDSu1w79jRq2LdQEw968uIR+H86P1e1/dJP4zzxhSlFTztdFSHsOa3v4jZ7lgY1AGIiAO4GzgVmAN8WUTm7NfsDuBBpdR84Cagp7SlXym1wP46M0F294mxdw/iUBRN+aQconIZw1YYvuueO8GhKPt/f0vLQ9++8OTlY5a58ezcnfS5jE3rELdJxZHHJX2uoRKZMQcVEXZ90Hdx+kSz7Yf/g2pROHd30vWdH7Hi88dYgQT9YPz7KcSjmPXNzCioEguzL7zCOrvbujGldjj3tvdZsa4vHC4XwUnFuKqaez007JN+OPrYmOae+52fYeQoXP9KbqZ+NCuAw4DNSqmtSqkg8Aiwfw7/HKBbGPu1Pu4PC86mZsiVXinbpsuJEUp+mF/9xjU4d7TRNaeCslnptcUxGIEpldBkJr08n2d3A5GSvlPq043cBda+b9OHyXcA217+F57lVQRnl1L6xD8JHjQO99paGr/0JZZees4Bmcu1q5biqOrEP28Snrz8pNs33BgOB4YXjNbUHcZHU7Fuf8wjjkQFDTY9+ud917qlH6Z/Kzqp9m5cvhw6jjgYqQ0ltaBRNA6gHOj5yVBtX+vJCuBz9utzgHwRKbG/94rIEhF5X0T6zKYQkUvtNkvq6uJP03e0+onkuXtdUy4nMgwOYOdvfwGmkN9Pkkc64zriGFDCzieTp0ne2ViPajYJTtz/Vyc9KT/WKv0X2ZBcqQwzEqHz5msRJ5Tf9v8onTGXgx97Dd/d/0t4fC6+N9dTfeJnWPazK/Y9Wdb87peghOJLfpBU21KJmePA0Z4ZFeu6mfTVb4Iouv5jbf/2lH6Ip+7F5KtuQRyKjj8mTyk1UYfAPwSOEZFlwDFADdD9qTtJKbUY+ArwaxE5wKUqpe5TSi1WSi0uKyvb/3bUSEeEcGHvJZvpcUM4uXWBzUiEnA+WQ6lj0DCvdKTynPNBFOEkPu3usFPqjXl9p9SnG3llY5E8haumJqnzrLrzZxi7g3Qef5gVTmhTedxpHPTSx6hrrsDMceB97BU2H7uQ1ffeTs7Ha1BjXEyyD/CzkUiOB4d/+KXK982/cgmIYsIp0f89F02cjBrtwrdxOzC49MOg41VOITBrdFKVUqNxADVATyWkCvvaPpRSu5RSn1NKLQR+al9rtv+tsf/dCrwO9K6qkCA6G+tRASFS0ltjRnnckOTC8Bse+SOqFbo+E/3TQjqRVzYWShx4tydPitZvO5fRJyT1GCihhItzcDUkbxuio74O7yNPIIXCvFv6ri0754JvMeP15XR99TQImDh+/WdUuxD4bPqfowyFSH4e+FOXod1dsS5/zPiY+nXNmYVqUuxe/tGg0g/RkH/p91ERYce9t8U9xkBE4wA+AqaLyGQRcQPnAb1CXESkVES6x7oGuN++XiQinu42wJHA2kQZ35O6lVbiiIzvrdpner2osCRjyn2EHn0QcSqmfTe9hKxiITCpHOpDdCZJOdW1eROGVzFuwaFJGT8ZhMaNRbWqpCWGbfzR/2B2CpFvXD5g0IDD5WLhtXcy8dV38R8/n/DkPGZm8O9aNJijClEBSYk6qBmJ4KjrJDi2KOa+o844F4Dd99w6qPRDNFSecIaVrV+1Pe4xBmJQB6CUCgNXAC8C64BHlVJrROQmEel+nDsW2CAiG4ExQHdpm9nAEhFZgXU4fKtSKikOoM1OZHJXTu91XflywJSk/RG31FTh3txAYHppzE8L6YRxyKfAFLY/lZx8ANfeFkJlmREZtY9ps0AJ1W/9J+FDV7/3Br4PNhGems/ci6LT8M8pLGbR3f9g3vMfkVOY/HKaqUSVlAFC45bhL1u6++N3UQGD8LTYM/gnn3wOhk+R8/aaqKQfBsNwOFD5Bq6GpiGN0+/40TRSSj2nlJqhlJqqlPqlfe06pdQz9uvHlVLT7TYXK6UC9vV3lVLzlFIH2//+KSnvAgjt2ArAqNn7ZdvZT1addcmpC7z1Nz9HRQTPuRclZfzhYsI5XwWg691XEj52S00VqlURnDQx4WMnk4JDPg1A68exF9wZjObrvw8Cpb/4bcLHzgZz/ty3AAAgAElEQVQc9sNU25bhrxJWZwux5R5xbMx9DYeDrsllqIhEJf0QDeFRuThak5OPkj2ZwLurQRSj90u2kFwrTM7fkJwiIN6330FGwYwvXpiU8YeLosnTkVHg2bw14WNXvfA4ILgWDJ5Sn05UHH0yiIIEH8Ctvuc2HDs76TxqLuMWZtbPZLhwV0wCwF+VfD2c/VFrhlaxTo6yzmeilX4YjHBZKapdJaWcadY4AEdjI4aXA+rsJrMw/OZnH4cGk87DF2VEbPtg+KdPwtgdYMvzTyZ03OASK6V+/MmZFSHlyctHCgTX7sQlyQXa23D++QEkTzHnV0lbEGc8+ZUzAIjsSW4UVl94qncNqWLd7Mt/jP/MI5l5/W8SY1BFJSihNgkCeVnjAJwt7Zh5zgOuG/Z/YrAp8YqX7Q/93npS+PbPEj52Kqj8+W8Rl6LrVzcmdFz39q1IrqJk+qyEjjschIrzcDYm7slrzY8vRrVC8GtfHdaSmJlG8QxbbKA+8Q9uA7GvYt24+CvWuXw5LLr9jweUfYwXjx0e3Lzig4SM15OscQBGR5BwwYG63U77jyzcnNhDlM7mRrzraghXFmTkB1tfFE2eTuexizF2BVj1u1sG7xAlztp2QqMzM2M1VF6Oardqww6VPSs/JueNFUQmeJn/neEt/p1p5BSXIi6FDHM976q3XkaFBHPm/mo3qaN40acACG5O/HlIVjiASCiE6oRw0YFhW65CKyE53Nqc0Dk3/PYXqKCBnPmlhI6baubefK+VAPXQgwnZc6zfuMZOqR9cUCsdccyYAwi73nhpyGPV/eRbKBMKfjZ8NV8zGp/gaBveMNCmN63/5/xPHz+s8w7E6LkLEIfCSIJUS1Y4gLp1K8EU1NgDwzA9xdZSTnUkti6w55X/ILmK2V+/MqHjphpPXj6B876IaoFV10UXnjgQu16wzhM8MaTUpxOFttRy+/KhLb/XPXQvzs0tdB06hYlHp6YcZqZh+pw4Ooa3mJOsX4U4FJOOTx+VVcPhgHwDZ0Pit7GzwgE0rbZKGjonVB5wz1ti7cOpBOYBVL3zCrInROei2ThcroSNmy7M+94Nlh75S2/TXL1jSGOFVywBYkupTyfKP3UcYiiMIShThgMB5Pe/wfAppt9xfwKty24iuT4Mf3hY5/TU1KKKnWmn5hsu8OFsSXxhq6xwAIGtVrJI7owDVThz7PKQdHYkbL76++4EUYz7ZmwKf5mC4XDg+u7VqICw7epvDGks984dSIFQMK5i8MZpiNPjQY0ycO2JP4x41Y3fRTUp/J8/PS5RsJFKpCAf/EOXcVn9p7tY/8jgEVchfyfSGCZQPmbIcyaacFkJtCkiocTqI2WFAzBrdgJQOn/xAfdyissAhSQohjbk78S3citmuS+rY7inn3M+oRnFeJdup/q9N+Iaw4xEcNb5CY7N7GiXUOkonE3xb0W4P3wfKYD51yRHzyVbMQuLUGEZ8gG88//uRd1wByvOPmrAzOIdrzxnVaxLQzl3s3wiyhT22rsdiSIrHICjfi/iUowqn3DAPcPhQJwgCaoLvO7/3YHyC+FTTk/IeOlM2U3/hwg0/Ty6Unb7s2f5h6guITx1RoItG17CFRMw/ULT9tiT5EL+Thx7/AQmjs6KXJFhpdR6Em/auC7uIdrr9ljBGoWCe0M9ez93Fkuv/RZm5EChubZ3LMmPYlsKPJ1wT7PK3DYufX+QlrGRHQ6guQVy+38rlgNITCq18dxTiFcx+/IfJ2S8dGbcgkPxHz4D59Y21v39DzH3r3vFKm2Yc/jRiTZtWHHOng/AnrdejLnv1mcfQ4UFteiwRJuV9bjGWtuG7dvjP3+pX2U9MftPORH3nb9AFTjxPf4q649bwLaXepdtNTauRVyKiiNPiN/oJFG0wPr9CW5KrJRaVjgAZ1uAcL633/vKJRgJqO26rxLTQZUHZBxnK9Nvuw/Dq+D3v+7zqWkgzFXLQBQTM/QAuJvST1up/f6VS2Lu2/HacwCMP/u/E2rTSMA7cTIAwZr4AxHaN1ofmO5J05h62heY8eoyus45BqMpSNeVV7Hsa6fuq7jm3l2PWepOy8COMQcfBoaCIQZl7E9WOAA6TCKF/dfuVC4DIzT0aIKau2+2KzF9f8hjZQp5ZWPpOP14qDdZcctVMfX1VNdAkZHxypVjFx6OOBWOOLaAPBs3IgUwes78JFiW3YyaaiVYRmrjl+IIVln/ZwUzDgJsae1b7qX08ScITS/G++F2qk7+DMtvvxaaTQLl6ano6/R4kDzBVV+f0HEz3gE0V+9AhYRIWf8n98plIMGhOQAzEiFnyWrUGGdWV2Lqi4Ov/zUUCTlP/JvOxuh+ASOhEEb90FLq0wXD4UAVOnHtjS0OO+TvxFHrJzBhdJIsy26Kp80GUchQhBxrdwFQNq93HarSGXOZ/8y7mD+6DByC5/5/ginIvKTUq0oIkQIvzpbECsJlvANosAvBGOMPPADuxnQPvS7wugd+NyIqMfWFw+UicvElmJ3C+p9Ep29e/e5rVkr99NlJtm54CI4uwmgKxLQNtu3fj6NCglqYOUVw0gmnx4N4wGhtiXsMR2MD4lH96i7N/fqVTH71Q7qOngMlBhO/cGHccyWbUGkxtJkxb8UORMY7gA77UMQzrX89HuV2ISFzSPOYTz6MuBQzv5Mdwm+xctDXv0dkoo+ct1ezd+3KAds2bFpPw9/uASDvyPRJqR8KkYmVqKBB/YY1Ufdpf83SlR9v11rQxIHPwNEWfw6Ps6UNNUCACFjZ7wvv+yez31lD8dTYi8AMF+a4ClRYYvodHIyMdwDhnZZeeNGc/pdupscDQ9gBCrS34drWTGDmGHKKM39LI17yf/JLVAR2/8yqcWpGIlS98xrLb7uGZReezuoTF7Jh4Sz2nnEOvjfXIx5F5QnZES7rnmv9fu19++Xo+3Tv/89dkCyzsp6Iz42jM/4ADkd7gEhe/wEimYTLPhNp+PjdhI15oH5yhiG1u8FQlM3u/5BNebxDcgDN2zahIkJk8rT4B8kCJh17KssPvgPPihrWHT0PaQqhQoJV9FkhBUKwvJDQpEm45x1K+WmfT7uU+ngZ85kTqb/zDwTXLIuqfTgQwLnHT2CG3v8fCpG8HBxDUQTtVEQmjkqcQSmkYN4hdPIgXRtXJ2zMjHcAzqZGJIcBQ7eUXRg+EgrFFeLVubsaAGNU7EWis41Jt9xL7VfOQilFYNZYzCkzyFl8JBOO/6+sXh2VzZpHg8fEsXN7VO23/vsx6wxE7/8PiUhBAaqrETMSiTmRrnV3tRUgUpodv5djD/kUW0VBAgvEZ74DaO3EzHMP2EblWE+hHfW1cWnSdNVaVYmcJYkp8JDJFE2eTtF7iU1GyRTMIg/uuuhkxdtffRYfMP6s85NrVJajikvA3EHLzq0UTZ4eU9+G1dZqTTJUh2p/PHn5SC446xJXJCfjzwCM9jDhgkGSsnKs+511tXHNEaq3+rlKtZDXSCY4ugRpDkcVheHeYO3/j9mvRrUmNsQWc2zeFLscRMdmO0BkUvZs3UYKPDibEidsmdEOINjRjtkFkeKBE426C8MHGuPznJFGK/47Z3x2PElo4sOsnIYKC7sHOYSz9v87CVRkx9ZDKnGPmwhAx47NMfcN2Vsl3Ulg2UC4pAhpT5xEdkY7gL2rloISGGSJZ+R1O4D4CiqoVqucZN74iXH112QHvoMttdmG914bsN225x7X+/8JIqfSenoP7Y69GpbalwSWPauwyNjxqKBB07ZNCRkvox1Ay7oVALgmDVxu0GEf3obiLAwvbVY1sXztAEY04445BYDw+lUDtmt79Vmrvd7/HzKF063avCoOSWhHYwPiVXjyMrMedV847XOQvR+9k5DxonIAInKKiGwQkc0ickAVFBGZJCKviMhKEXldRCp63LtARDbZXxckxGqb4HbLC+YPot/dXRc40hZfRqG0tyEuhdPjiau/JjsorJiE5ChcVQM/jbo3bEDyYez8Q4bJsuxlVPlEqyJbHKt3Z0s7Kie7JLjzD7J+pzrXr0jIeIM6ABFxAHcDpwJzgC+LyJz9mt0BPKiUmg/cBNxi9y0GrgcOBw4DrheRhMVSql3WH2LZ/IGX2m5bjCzSFl9heEenHwYONNKMECJFXlwN/deXDgcCOHd3Epig9/8TgeFwgBeM1thLujo6AkQGUAnORMYcdhSgUDu2JWS8aFYAhwGblVJblVJB4BHgrP3azAFetV+/1uP+ycDLSqlGpVQT8DJwytDNtnA01CMeNWj8ubfUTsZpb49rHqMrgPJk9G6ZJkGExo6GFpNQPxXmtj3/hLX/v+DA6nSa+FA5DhztcYigdSrCo/pXCc5EcgqLkRxw7o0vonF/ovlUKwd6rnmr7Ws9WQF8zn59DpAvIiVR9kVELhWRJSKypK4ueuU/Z3MrKnfwJZ63xHIAqjO+wvBGIITyZHzKhCYBqCkzUKZQ897rfd5vs4vgjD3zK8NoVXYTyfHg8MdWC7elpspyxKXZl4ltFrhxNcf3WbY/iXqs/SFwjIgsA44BaoCoJeuUUvcppRYrpRaXlUWfbOVoDxIp8A3aLq9bKrozvrKQEohgevUekAZyFx4BQPNHb/V5371hPZKnGLdARwAlikh+HvhjU8BsWP0xAJKFoduh4lFIa2JCQaNxADVAT63lCvvaPpRSu5RSn1NKLQR+al9rjqZvvJiRCHQowkWDFxx3+XIQh8LoilNLO6iI+AZ3NJrsp/zokwGFufHAbOhIKIRzT4fe/08w5qhCVEAIdkS/hduxeT0A3ixKAusmMmYsKiC01MQeGrs/0TiAj4DpIjJZRNzAeUCvYpoiUioi3WNdA9xvv34ROElEiuzD35Psa0Omaasl0GaOjjI716mQrq6Y5zEjEVQQzJzsEDXTDI3c0jIkD9y7DnyO2fr8P1FBA/Ngvf+fSFRJGSA0btkQdZ9wlXVIWjAze5LAujHs3Ii9H7099LEGa6CUCgNXYH1wrwMeVUqtEZGbRORMu9mxwAYR2QiMAX5p920Efo7lRD4CbrKvDZmGVVZ9VqO8MroOLsEIBGOex9/UCKag8rPrMEkTP+HiXJwNB6bjt/3H3v8/68vDbVJW4xhjlWls27I++k57dwOKsrnpW+ErXnJt6fuOtcuHPFZUJ5tKqeeA5/a7dl2P148Dj/fT934+WREkjM7N6/EAudFWnHIaSDB2B9Bavd16kT/4VpNmZBAaPw5H1WY6mxt71Tt2r7f3/xcenkLrsg93xSQA/FXRhz4ajQ0YXnDnDqITloGMPewoqgEzDnmM/cnY2Eaz2vplKD4oumQb5TIw4qgL3LlrBwBGhhc21yQOmTYLlFDz1n/2XbP2/9u1/k8SyK+cAUBkT/THh86WdswoIgQzkfwx4zG8Ckdt7NnR+5OxDsCoq0UciqIp0UnEmq746gIH9lo/ZGexloLWWBQcehQA7T1E4ba+8KS1/79AZ/8mmuIZdt5pffRijo6OIJH87A3cMAtcuJr6T0iMlox1AM6mZsiVqItEKI8zrrrA3VLQnjHjYu6ryU7KjzweDAU9DiXb/mPFRYw547xUmZW15BSXIi6FRFkZzIxErCSwwuyoBNYXocJ8HK2x5Ub0RcY6AEern0he9No8ptuNhFTM80TsXzrf2OyLJ9bEhycvH8kXXHs+WYK7169D8hTjD/lUCi3LYnyCsy265KeWmp2osGCWZu+qPTxmDKZf6KiPPnG2LzLWAUhHmHBh9Cp/pseDiid3osWSgs4t10qgmk8IlebjbLQSCyOhEM7d7QQqSlJsVfZi+pw4OqIrDt+4xoqOMbL4oU1sBeQ9QwwFzUgH0NlYjwoYMdX6VD4fKkxU1Zx6YklBq7hKSWqyl1B5OapDaKvdxbYXn7bj//X+f7KI5PmQzui2PPZVAouxhGQmkTvrYADa1ywd0jgZ6QD2Lv8QABk3YZCWn6C8PlBCl/1EHy1GRzviVnEVk9dkLw67ytSuN16i9T9PA3r/P5lE8vMhyjzOsF0JbNTMgWXiM5myxVYgQmTbxiGNk5EOoH3jagA8k2dE36m7LnB9bCp6ht8Pbompjyb7Kf7UsQB0Lv/A2v/PVYxf/OnUGpXFmEXFqLC14hoM2bsbRFE2d8EwWJYaiiqnIG4Tx56hhYJmpAMIbd8CQMGs+dF3shNC/HWx1QU2/AGUOzvjiTXxM/6woxGHQrZuxLmrjaDe/08qUmoJOjZvHrw4vKOpEcNraYBlMyrfibMxvhon3WSkA2BPDYhidAy1Pg1byiEQY1lIIxDG1FLQmv1wuFyoUQ58G2pQQYPI/OypO5uOOO0D3fatg9fCdbR2ZG0SWE/CRXk4WqM7GO+PjHQAjjjSvB15VkxwqKk+prm0FLSmP0JlozD91vbg6DO+lGJrshvvRCvqJWBn5g9EtieBdRMuG43qhK7W+FcBGekAXM3tmPmxHco6C61KlOEYD4EJmloJVNMn4QlWaLDkKsoP+0yKrcluRk2dBYBZu3vAdvtk4gtHgHbXxMmghD1L3h28bT9knAOIhEJIY4jgmNj2XN1FsReGt6SgBTNXOwDNgbhmW4eMwQqtE5VsiqfOBFFI48Ar+JaqHZZMfBYngXXjtaOcWm1l5HjIOAdQ9dbLqLAQmbF/XfqB8RRZvxBme/T6GR31taAElaeloDUHMu7EMxCHwjzyuFSbkvU4PR4MDxgtA293NKy24uKNcdmfuFlmR52Ft8YfCppxp5tN7/wHL5B/xLEx9fOVjqENIIaqQm3V1n6jFIyA5aQmZkpnzCXnrbfJKdYKoMOB8hk42g+sw9CTzi1rcQGeKEUiM5niqbOodyqM3fEXWcy4FYCsX40YiknHnRZTP1+ZVRxaOgf+BepJxy6r5JoU6SW+pm/0h//wEclx4+gYuKZHuGYnAIVZnATWjeFwoPINXA0xnmv2HCOB9gwL7po9qCJHzIUevPmjQBTEUBc4uNc6cHKWjo5pLo1Gk3gieTkYgxSHl1o7CWxO9iaB9SRcmDukUNCMcgBmJILRECQ4NvakG8PhQJxg+KOvCxxqsJT2PGVaClqjSTWRglGYASsQpD8cTU0YXuvMYCQQLitFtStC/ugfbHuSUQ6g5v03USEhMnVmfAO4QALRe0vVbCWN5Ywtj28+jUaTMFRxCZhCy86t/bZxtrZj5mbc0Wb8VFRaoaBLP4ire0Y5gIY3XgAg74hj4hvAJRgx1AVWLVYtgLxoC89rNJqkYZSNBQaWgzA6QoRHQBJYN54ZcwFoWflhXP0zygGodStBFBNPOD2+/k4DCUZfRUfa20AU+WPHxzWfRqNJHK5yqzh8544tfd7vrgQWGQlJYDZFCw4HINijOl0sZJQDcFfvQgoNvHGGZSq3AyOGusBGRwfiJuqykxqNJnnkTpoKQMiOztuf5u2brSSwspETtDF67gLEoTBq+v6ZDEbGOAAzEsHR0EVgTPwhmVZh+OjrAmspaI0mfSi0tztUfd8SyA1rlgFgjKDqfQ6XC/IEZ0NsIpfdROUAROQUEdkgIptF5Oo+7k8UkddEZJmIrBSR0+zrlSLiF5Hl9te9cVkJ7Fn2gaW6ODX+BA/ldkNMDiCIcmeMj9RospqCcRXW025T38Xh/XZGrHfStOE0K+WER+XgbPHH1XfQ43IRcQB3AycC1cBHIvKMUmptj2bXAo8qpe4RkTnAc0ClfW+LUmrIQbl7X30OD+A79Mi4xzA9HoihLrARCGF6dCUwjSYdMBwO8IKjpW85l3D1dlxA4QjJAegmXFaCszr6BNeeRPN4exiwWSm1VSkVBB4BztqvjQK6BXNGAYOX7YkRtXY5iGLSSWfHP4bXA9GfAVtS0L6REU+s0WQCyufE6Oj7aVfq9oAoSmceNMxWpRZz/ASUGd9WdTQOoBzoecJQbV/ryQ3AV0WkGuvp/9s97k22t4beEJE+NXNF5FIRWSIiS+rq6vo0wl1dgxTIkFLvTV8OyhSC0eoBBRVmzsgJKdNo0p1IjgdHZ9+h3I7GJgzfyEkC68Y9dXbcfRO1wf1l4AGlVAVwGvCQiBjAbmCiUmoh8H3g7yJygLSmUuo+pdRipdTisrIDZVzNSARHvZ/gmKKhWWmXiOuIoiykGYmgQoKZkzu0OTUaTcKI5OdBl+rznrO1Y2QlgdkULjgs7r7ROIAaYEKP7yvsaz35OvAogFLqPcALlCqlAkqpBvv6x8AWIIZK7hZ71yxHdQnhKVNj7doLyc0HwN8weGH4tj27LCno/PwhzanRaBJHpLAQFeh7FW90BAnnj7zaHWMXHg5G305xMKJxAB8B00Vksoi4gfOAZ/ZrsxM4HkBEZmM5gDoRKbMPkRGRKcB0oP887n7Y++q/AfAd8ulYu/YmzxKQCzT0vc3Uk7aqbQBIwRBXHRqNJnGUWDsEDZt6ZwObkQj4IVI0cpLAunF6PEheks4AlFJh4ArgRWAdVrTPGhG5SUTOtJv9ALhERFYADwMXKqUUcDSwUkSWA48Dlyml+o7hGoDwqo8BRcVJ+589x4Yj36oLHGwe3AT/HmuRY2i5X40mbXCMtoQZ27b2znxt3LLeTgIbkwqzUk6kwBtXv6g2zJRSz2Ed7va8dl2P12uBA+IzlVL/BP4Zl2U9cFdVIQVC/pihSTI4R1lJZKHmwZMmAnW7cQDOkpGTVajRpDueikoU0FXVeyOhae1KABzjJ/TRK/sJlcS3U5ERWU6uuk5Co4deltE1yvohRVoHLisHEG7UUtAaTbqRP8U6Qozs6R1p7t+6HgDv5DiVgjMciTM/Ku0dwN61KzH9QmjylCGP5S6Ovi6warSloMdpKWiNJl0o6paD2C+QY18lsFnZXwmsLw7+4c/j6pf2DqDWPgD2LDhiyGN57f181d42aFtlrxLyKyYNeV6NRpMYcgqLEZfC0dR7FS9794ChKB2hDiBe0t4BhGyd6/KT488A7iZntKUnThR1gaW9DQxFbunIPFTSaNIWn+DY7yHO2dSE+GxxNE3UpL0DcO+sQvIUhQl4Es+xt4CiKQxvdHYgLi0FrdGkG6bPhaOjd2U/R1sHKld/+MdK2jsAZ107obLEJGM5XC7EqZDA4HWBjU4/eLQUtEaTbkTyvEhnb1EvqxLYyEsCGypp7QAat2xAdQihyskJG1OcIF1ROICuIMqtn/41mnQjkp8PPf6EI6EQqhPCIzAJbKiktQPY9fK/AHAfHL/Wxf4ol2AEBq8LbATCmF69pNRo0g2zqBgVFtpqrVDQxi3rwRSUPq+LmbR2AMEV1gHw+BPPSNiYyiUYUdQFlmAE0zuyVAU1mkxA7A/67uLwTWtXAODQEXsxk9YOwLV9G5KrKJ6auOQO5XIgoSiqwgQVpk9LQWs06YZzbAUA7Vs3AeC3ZSF8U0ZmEthQSG8HUNdGuCwvoWMqlxMZpDB8JBRCBQUzV0tBazTphneilRQa2LUDgMi+JLD5KbMpU0lbB9BcvQPVLgQnJlbbw3S7kNDA0qmtNVWAoPKHLj+h0WgSy6ipswAwa3cDdiUwQ1EyPf7CKCOVtHUANS8+BYBrfuIOgAFMj3tQB9BeYz1ZSGFxQufWaDRDp3jqTBCFNNYD4Gxu1klgcZK2DiCw/H0Axhz3XwkdV3m9qEGOAPy1lhS0o6gkoXNrNJqh4/R4MDxgtFhyEI7WTsw8/eEfD2nrAFzbtmL4FKPnJHZfT3m8qLAQCfUfCRTcuwfQUtAaTbqifAaOdiuj3+gIEcnX53XxkL4OYG8robLEZ/Ypu8ZvZ2P/VcH2SUGPHlr9AY1GkxwiOW4cHUErYMMP4UKdBBYPaekA2mp3oVoVwQlJKO7Q7QDsp/y+MJu7paArEj+/RqMZMpG8HAx/hIaNa6wksG6hR01MpKUDqH7paUBwzjsk4WNLnqUr1GUfIPWJLQU9amLiJCg0Gk3iiIwahRmApjXLAXDqJLC4SEsH4P/4XQBGJ/gAGMDIs0I7gwNsAUl7G2IocnQ9YI0mLVHFpWAKnUvfA8A3eUaKLcpM0tIBOLduQbyK0XMXJHxsR4G1Vxhqaeq3jdHZibgTPrVGo0kQhr3lY2xcC0BREj4rRgJp6QDctU1ESn1J0eJ3FVqhneGWxn7bODr9KLeWgtZo0hX3eGvLx72nATEUxXZymCY20s4BmOGwdQBckZxavG47uSvS1tJvG6MriPJoKWiNJl3JrZxqvWgyIUcXboqXtHMAwdZmUILMSc6SzlNiVQWjvb3fNhIIE9FS0BpN2lI4/SDrhRLMXL1fGy9p5wDMTuuDefRxpyVlfF+JJSWrOvovDC9BE9PrTcr8Go1m6OSPHY84LEmXiK4EFjdROQAROUVENojIZhG5uo/7E0XkNRFZJiIrReS0HveusfttEJGTB52rqwtxm4xdeHhs7yRKcsvs7F5/Z/+NgmDmaClojSZdMRwOsJ/RwsVFqTUmgxnUAYiIA7gbOBWYA3xZRObs1+xa4FGl1ELgPOD3dt859vdzgVOA39vj9T9fOEKkxJu0PT13bh5iKIx+HEA4EECFBJWbWBlqjUaTWJTPaf07elyKLclcolkBHAZsVkptVUoFgUeAs/Zro4Bu7eRRwC779VnAI0qpgFJqG7DZHq9/IopgRZIlGFwggUCft1qrtwOgCkYl1waNRjMkIrlWxT5n+cQUW5K5ROMAyoGqHt9X29d6cgPwVRGpBp4Dvh1DX0TkUhFZIiJLUILMTnJhBycYXX07gHa7uISM0stKjSadieRbq3SfDgGNm0QdAn8ZeEApVQGcBjwkIlGPrZS6Tym1WCm1GIei5JhTEmRWP7gMJNh3YXi/XWjaUaSzgDWadCZSaD2kFc/WSWDx4oyiTQ3QU5Wtwr7Wk69j7fGjlHpPRLxAaZR9e+GdfRATjvxsFGbFj3IZGP3UBQ7V78EBuLQUtEaT1oz/zs/YXf4XSqbrFUC8RPOU/hEwXUQmi4gb61D3mf3a7ASOBxCR2Qt+xL0AABg+SURBVFjn83V2u/NExCMik4HpwIeJMj5eTJcDCfZdFzjUsBcA79jkJKJpNJrEMHb+ISy88TepNiOjGXQFoJQKi8gVwIuAA7hfKbVGRG4CliilngF+APxBRL6HdSB8oVJKAWtE5FFgLRAGvqWUGrgi+zCg3C6ko+8tINVsaQTljk+CFLVGo9GkEdFsAaGUeg7rcLfntet6vF4LHNlP318CvxyCjQnHdLsh1HcmsLRaEhEFEyqH0SKNRqMZftIuE3g4UB4P9FMRUjraEIfCW6ArDGk0muxmRDoA02cVhjcjB+5GSUcnaGkRjUYzAhiRDgBvDiihqw9FUIffD1oKWqPRjACiOgPINroLw/vr9pJjy0N3Y3SFMD0j8sei0SSdUChEdXU1XV1dqTYlY/F6vVRUVOByDV2xeGR+0tk6P/76WtgvhtgIhgnnayVQjSYZVFdXk5+fT2VlJSJ6pR0rSikaGhqorq5m8uSh1ywfkVtA3XWBA00H1gWWgJaC1miSRVdXFyUlJfrDP05EhJKSkoStoEakA3DkW0JvwaaGA+4pLQWt0SQV/eE/NBL58xuRDsBpC72Fm3sXhg92tKPCgpmXnwqzNBqNZlgZkQ7AZQu9Rdp7RwG11uywXuRrKWiNJpt56qmnEBHWr1+falNSyoh0AJ6iEgDMttZe1zt2aSlojWYk8PDDD3PUUUfx8MMPJ22OSB95RunGiIwC8pWNph2go7cchH+PLQVdrKWgNZpkc+O/1rB2V+vgDWNgzvgCrj9j7oBt2tvbefvtt3nttdc444wzuPHGGwG47bbb+Otf/4phGJx66qnceuutbN68mcsuu4y6ujocDgePPfYYVVVV3HHHHTz77LMAXHHFFSxevJgLL7yQyspKzj33XF5++WWuuuoq2trauO+++wgGg0ybNo2HHnqInJwcamtrueyyy9i6dSsA99xzDy+88ALFxcV897vfBeCnP/0po0eP5sorr0zoz6gnI9IB5JRaheHp7O0AgnV7cAGu0WOH3yiNRjMsPP3005xyyinMmDGDkpISPv74Y/bu3cvTTz/NBx98QE5ODo2NjQCcf/75XH311Zxzzjl0dXVhmiZVVVUDjl9SUsLSpUsBaGho4JJLLgHg2muv5U9/+hPf/va3+c53vsMxxxzDk08+SSQSob29nfHjx/O5z32O7373u5imySOPPMKHHyZXPHlEOgDvqCIQhXT5e12PNNXjAnyjk1ySUqPRDPqkniwefvjhfU/V5513Hg8//DBKKS666CJycnIAKC4upq2tjZqaGs455xzASsCKhnPPPXff69WrV3PttdfS3NxMe3s7J598MgCvvvoqDz74IPz/9u49Oqr6WuD4d2cyCW8I4SEaTIN1CQiB2ABGRIssUGqLlyog10ux4tKWQrm67qra+kCbtpZGC7kLxdZiFLkYoF4vtfjiVXzWBAwEw1tAiECANJoAJiSz7x9zEieBhAQm8+Dsz1qszJzHzJ5fyNlzfr9z9g/weDx07tyZzp07k5iYyCeffMLhw4dJS0sjMTExaJ/7TFyZAGI8HiQW5GT9BGCloI25sJWWlrJmzRoKCwsREWpqahARJkyY0OzXiI2Nxefz1T1veE1++/bt6x7feeedvPbaawwaNIicnBzWrVvX5Gvffffd5OTkcOjQIe66665mx3SuXDkIDCCxENNwYninNlDn3ud/h50xJvIsX76cKVOmsG/fPvbu3cv+/ftJSUmhc+fOvPDCC5w4cQLwJ4qOHTuSlJTEa6+9BkBlZSUnTpwgOTmZoqIiKisrKSsrY/Xq1Y2+X3l5Ob169eLUqVMsXry4bvmoUaN49tlnAf9g8Zdf+o8948eP58033yQvL6/ubKE1uTYBqFeIaTAvcExFBRKrxDmlIowxF5YlS5bUdenUuvXWWzl48CDjxo0jPT2dwYMHk5WVBcCiRYvIzs4mNTWVa665hkOHDtG7d28mTpzIgAEDmDhxImlpaY2+369//WuGDRvG8OHD6dv3m7Iz8+bNY+3atQwcOJDvfOc7FBUVARAXF8fIkSOZOHEiHo+nFVqgPvFP3BU50tPTNT8/v9XfZ+vwK/HFe7lyTUHdss23XEPc/lL6bnT3tcHGtJatW7fSr1+/cIcRsXw+H1dddRXLli3j8ssvb3S7M7WjiGxQ1fSWvJ+LzwA8p00MH3Pya4hzbZMYY8KoqKiIb3/724waNarJg38wuXIQGMAXF0vsV/UHb2IqT+GLb/3TLmOMaah///519wWEimu/7mpcHFT76i2LqazBF2/TgRlj3MG1CcAXHw+n6o9/SKUPX1srBW2McQfXJgBtEw/V9cuqahX42rcLU0TGGBNark0Avjbt0Brh1En/db+VFeVojeDrYJeAGmPcwbUJAGfSl4ojhwH48vM9/uUdu4QrImNMCGRnZ9OvXz/uuOMOtm3bRkZGBvHx8XXX/rtJs64CEpGbgHmAB3heVZ9ssP6PwEjnaTugh6p2cdbVAIXOus9VdVwwAj9f0s4/6cvXx0rg0hSOH/QXeJIuVgramAvZM888w6pVq0hKSqKkpITs7Oy6u31Dobq6mtjYyLgA86xRiIgHmA+MBg4AeSKyQlWLardR1fsCtp8JBN4ad1JVBwcv5CBxunq+Plri/3noAACxXXuELSRjXOWNB+FQ4dm3a4mLBsLYJxtdXVuCeezYsdx1113cd9999OjRg7///e+N7lNTU8O0adPIz89HROr2O1Op6D59+vCLX/yCN954AxHh4YcfZtKkSaxbt45HHnmEhIQEtm3bxo4dO3j55ZfJzs6mqqqKYcOG8cwzz4Tk7t9AzUlDQ4FdqvoZgIi8AtwCFDWy/WTgseCE13o8TldPVZm/7Oupo4fxAnHdrRS0MReqBQsW8Oabb7J27Vq6dWvevB8FBQUUFxezZcsWAMrKyoAzl4p+9dVXKSgoYNOmTRw9epQhQ4Zw3XXXAbBx40a2bNlCSkoKW7duJTc3l/fffx+v18v06dNZvHgxP/rRj1rngzeiOQngEiCwAPYBYNiZNhSRZCAFWBOwuI2I5APVwJOqetq5lojcA9wDcOmllzYv8vPkcWb9OlXmnxi+5l/H8AJtelopaGNCoolv6pGkT58+fPbZZ8ycOZObb76ZMWPGNFoq+r333mPy5Ml4PB569uzJ9ddfT15eHp06dWLo0KGkpPgLTa5evZoNGzYwZMgQAE6ePEmPHqHvfQh2R9TtwHJVDZwLLVlVi0WkD7BGRApVdXfgTqr6J+BP4K8FFOSYzsjbOYEaoOZLfwno2lLQHS4OTQIyxkSHhIQENm3axFtvvcWCBQtYunQp8+bNa/HrBJaJVlWmTp3K7373u2CG2mLNuQqoGAgskJ/kLDuT24F6k2yqarHz8zNgHfXHB8ImvqszL3BFuX+BMz9wpyRLAMaYbxw9ehSfz8ett95KZmYmGzdubLRU9IgRI8jNzaWmpoYjR46wfv16hg4detprjho1iuXLl1NS4h+DLC0tZd++fSH9XNC8M4A84HIRScF/4L8d+PeGG4lIXyAB+DBgWQJwQlUrRaQbMByYE4zAz1d81x6cAPS4PwHEHC9HvIq3rd0IZoxbHDp0iPT0dL766itiYmKYO3cuRUVFdOrUqW6b4uJifvzjH9dNAlP7rX3RokXce++9PProo3i9XpYtW8b48eP58MMPGTRoECLCnDlzuOiii9i2rX6F4f79+5OZmcmYMWPw+Xx4vV7mz59PcnJy6D48zSwHLSLfA+bivwx0oar+RkSeAPJVdYWzzWygjao+GLDfNcBzgA//2cZcVf1LU+8VqnLQXx08QPHI0ZwclcpV83PZPC6DuAP/slLQxrQiKwcdHMEqB92sMQBVXQmsbLDs0QbPZ59hvw+AgS0JKFTaOxPDizMDUMzJSoh3731xxhj3ce0Rz+P1IrGKOPN5eipP4YuPjJszjDEmFFybAMA/L7BU+hOAVFbja2OloI0x7uHqBKABE8NLpVJjpaCNMS7i7gTgjUGqTvkfnwJtZ1cAGWPcw+UJwENMVTUnykqdUtAdwx2SMcaEjLsTQFwscspH+f69/gWdrBS0MRc6j8fD4MGDGTBgABMmTOCEcyXg+cjPz+fnP/95o+u/+OILbrvttvN+n2BzdQLwef0J4PgXnwMQk9A1zBEZY1pb27ZtKSgoYMuWLcTFxbFgwYJ661W17qav5kpPTyc7O7vR9RdffDHLly8/p3hbk6uve/TFx0N1OZUlBwGI7do9zBEZ4x6///j3bCsN7o2Xfbv25YGhDzR7+xEjRrB582b27t3LjTfeyLBhw9iwYQMrV65k+/btPPbYY1RWVnLZZZfxwgsv0KFDB/Ly8pg1axbHjx8nPj6+rrBbVlYWr7/+Ov/4xz+YNWsWACLC+vXrOXbsGN///vfZsmULX3/9NT/96U/Jz88nNjaWp59+mpEjR5KTk8OKFSs4ceIEu3fvZvz48cyZ07qFE1x9BqBt4uGUvxQ0QFw3KwVtjFtUV1fzxhtvMHCg/17VnTt3Mn36dD799FPat29PZmYmq1atYuPGjaSnp/P0009TVVXFpEmTmDdvHps2bWLVqlW0bdu23utmZWUxf/58CgoKePfdd09bP3/+fESEwsJClixZwtSpU/nauR+poKCA3NxcCgsLyc3NZf/+/bQmV58BaJu2aDXUlB7FC7TtdUm4QzLGNVryTT2YTp48yeDB/jmqRowYwbRp0/jiiy9ITk7m6quvBuCjjz6iqKiI4cOHA1BVVUVGRgbbt2+nV69edWWcA2sG1Ro+fDj3338/d9xxBz/84Q9JSkqqt/69995j5syZAPTt25fk5GR27NgB+IvEde7cGfDXC9q3bx+9e/emtbg7AbRtBwgc9ncBWSloYy58tWMADTUs1zx69GiWLKlX3JjCwrPPYPbggw9y8803s3LlSoYPH85bb71VN1/A2cTHx9c99ng8VFdXN2u/c+XqLiDa+X/hMceOAEpHSwDGGODqq6/m/fffZ9euXQAcP36cHTt2cMUVV3Dw4EHy8vIAKC8vP+0gvXv3bgYOHMgDDzzAkCFDTqsEOmLECBYvXgzAjh07+Pzzz7niiitC8KlO5+oEIM51/7FlXyFeiA3IvsYY9+revTs5OTlMnjyZ1NRUMjIy2LZtG3FxceTm5jJz5kwGDRrE6NGj6/rva82dO5cBAwaQmpqK1+tl7Nix9dZPnz4dn8/HwIEDmTRpEjk5OfW++YdSs8pBh1KoykEDFMx5mPiFf0U6A9VK3w1WCtqY1mTloIMjWOWgXX0GEFt749cJRa0UtDHGZVx91Ivt4p8YXk8JvnhvmKMxxpjQcnUCiEtIrHtspaCNMW7j6gTQJrFH3WOflYI2xriMqxNA28RvSj/42rVvYktjjLnwuDoBtOv+TekH7WiloI0x7uLqBBDfoSPE+C+DlU4JYY7GGBMKgeWgf/CDH1BWVhbU18/JyWHGjBkAzJ49m6ysrKC+fjC5OgEASKyTALpYKWhj3CCwHHTXrl2ZP39+uEMKG1fXAgLAK1AFsYlWCtqYUDr0299SuTW4N1/G9+vLRb/8ZbO3z8jIYPPmzXXP//CHP7B06VIqKysZP348jz/+OAAvvfQSWVlZiAipqaksWrSIv/3tb2RmZlJVVUViYiKLFy+mZ8+eQf08rc0SQKwAEN/dSkEb4yY1NTWsXr2aadOmAfD222+zc+dOPv74Y1SVcePGsX79ehITE8nMzOSDDz6gW7dulJaWAnDttdfy0UcfISI8//zzzJkzh6eeeiqcH6nFmpUAROQmYB7gAZ5X1ScbrP8jMNJ52g7ooapdnHVTgYeddZmq+mIwAg8W9XqAGtpelHTWbY0xwdOSb+rBVFsOuri4mH79+jF69GjAnwDefvtt0tLSAKioqGDnzp1s2rSJCRMm0K1bNwC6dvV3Fx84cIBJkyZx8OBBqqqqSElJCcvnOR9nHQMQEQ8wHxgL9Acmi0j/wG1U9T5VHayqg4H/Bl519u0KPAYMA4YCj4lIRI22+rweADr0/lZ4AzHGhETtGMC+fftQ1boxAFXloYceoqCggIKCAnbt2lV3dnAmM2fOZMaMGRQWFvLcc8+dVhQuGjRnEHgosEtVP1PVKuAV4JYmtp8M1BbRvhF4R1VLVfVfwDvATecTcLBpvBdE6djz4nCHYowJoXbt2pGdnc1TTz1FdXU1N954IwsXLqSiogKA4uJiSkpKuOGGG1i2bBnHjh0DqOsC+vLLL7nkEv8kUi++GFEdG83WnC6gS4DAeckO4P9GfxoRSQZSgDVN7HvatFsicg9wD8Cll4a2Jr8vLo5Yr+LxWi0gY9wmLS2N1NRUlixZwpQpU9i6dSsZGRkAdOjQgZdffpkrr7ySX/3qV1x//fV4PB7S0tLIyclh9uzZTJgwgYSEBG644Qb27NkT5k/TcmctBy0itwE3qerdzvMpwDBVnXGGbR8AklR1pvP8v4A2qprpPH8EOKmqjV4YG8py0ABbFy3gZOEGrprz55C9pzFuZeWggyNY5aCbcwZQDAROSpnkLDuT24GfNdj3uw32Xdf88Fpfvyk/CXcIxhgTFs0ZA8gDLheRFBGJw3+QX9FwIxHpCyQAHwYsfgsYIyIJzuDvGGeZMcaYMDvrGYCqVovIDPwHbg+wUFU/FZEngHxVrU0GtwOvaECfkqqWisiv8ScRgCdUtTS4H8EYE01UFREJdxhRK5izOLp6SkhjTGjt2bOHjh07kpiYaEngHKgqx44do7y8/LT7DlprDMAYY4IiKSmJAwcOcOTIkXCHErXatGlDUlJwbly1BGCMCRmv1xuVd8xeqFxfDdQYY9zKEoAxxriUJQBjjHGpiLsKSETKge3hjuM8dAOOhjuI82Dxh5fFHz7RHDvAFaraorltI3EQeHtLL2WKJCKSb/GHj8UfXtEcfzTHDv74W7qPdQEZY4xLWQIwxhiXisQE8KdwB3CeLP7wsvjDK5rjj+bY4Rzij7hBYGOMMaERiWcAxhhjQsASgDHGuFREJQARuUlEtovILhF5MNzxtJSI7BWRQhEpOJdLskJNRBaKSImIbAlY1lVE3hGRnc7PhHDG2JRG4p8tIsXO76BARL4XzhgbIyK9RWStiBSJyKciMstZHhXt30T80dL+bUTkYxHZ5MT/uLM8RUT+6RyDcp05UCJOE/HniMiegPYf3OTrRMoYgIh4gB3AaPxzB+cBk1W1KKyBtYCI7AXSVTUqbiYRkeuACuAlVR3gLJsDlKrqk04STlDVB8IZZ2MaiX82UNHUtKORQER6Ab1UdaOIdAQ2AP8G3EkUtH8T8U8kOtpfgPaqWiEiXuA9YBZwP/Cqqr4iIguATar6bDhjPZMm4v8J8LqqLm/O60TSGcBQYJeqfqaqVcArwC1hjumCpqrrgYYT9NwCvOg8fhH/H3VEaiT+qKCqB1V1o/O4HNgKXEKUtH8T8UcF9atwnnqdfwrcANQePCO5/RuLv0UiKQFcAuwPeH6AKPoP5VDgbRHZICL3hDuYc9RTVQ86jw8BPcMZzDmaISKbnS6iiOxCCSQi3wLSgH8She3fIH6IkvYXEY+IFAAlwDvAbqBMVaudTSL6GNQwflWtbf/fOO3/RxGJb+o1IikBXAiuVdWrgLHAz5wuiqjlTO8ZGX2EzfcscBkwGDgIPBXecJomIh2AvwL/qapfBa6LhvY/Q/xR0/6qWqOqg4Ek/D0QfcMcUos0jF9EBgAP4f8cQ4CuQJPdh5GUAIqB3gHPk5xlUUNVi52fJcD/4v9PFW0OO/27tf28JWGOp0VU9bDzh+ED/kwE/w6cvtu/AotV9VVncdS0/5nij6b2r6WqZcBaIAPoIiK1NdKi4hgUEP9NTtecqmol8AJnaf9ISgB5wOXOKHwc/knmV5xln4ghIu2dwTBEpD0wBtjS9F4RaQUw1Xk8Ffi/MMbSYrUHT8d4IvR34Azi/QXYqqpPB6yKivZvLP4oav/uItLFedwW/8UnW/EfSG9zNovk9j9T/NsCvjwI/vGLJts/Yq4CAnAuGZsLeICFqvqbMIfUbCLSB/+3fvBXWf2fSI9fRJYA38VfBvcw8BjwGrAUuBTYB0xU1YgcaG0k/u/i735QYC9wb0CfesQQkWuBd4FCwOcs/iX+fvSIb/8m4p9MdLR/Kv5BXg/+L8JLVfUJ5+/4FfzdJ58A/+F8m44oTcS/BugOCFAA/CRgsPj014mkBGCMMSZ0IqkLyBhjTAhZAjDGGJeyBGCMMS5lCcAYY1zKEoAxxriUJQBjjHEpSwDGGONS/w+eMsi2CzdSKwAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "%matplotlib inline\n",
    "test_results_df.plot()\n",
    "plt.title(\"Result for Test Data\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 117,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Accuracy     0.942413\n",
       "f1 score     0.942413\n",
       "Precision    1.000000\n",
       "Recall       0.942413\n",
       "dtype: float64"
      ]
     },
     "execution_count": 117,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_results_df.mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 118,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Accuracy     0.937089\n",
       "f1 score     0.937089\n",
       "Precision    1.000000\n",
       "Recall       0.937089\n",
       "dtype: float64"
      ]
     },
     "execution_count": 118,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_results_df.mean()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 9. Export your model as a pickle file"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 119,
   "metadata": {},
   "outputs": [],
   "source": [
    "# we'll choose random forest with \n",
    "import pickle\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 120,
   "metadata": {},
   "outputs": [],
   "source": [
    "# save the model to disk\n",
    "filename = 'best_model_bi.pkl'\n",
    "pickle.dump(best_model_bi, open(filename, 'wb'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 121,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "filename = 'best_model_kn.pkl'\n",
    "pickle.dump(best_model_kn, open(filename, 'wb'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 122,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "filename = 'best_model_rf_tuned.pkl'\n",
    "pickle.dump(best_model_rf_tuned, open(filename, 'wb'))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "k neighbours classifier and random forest peroforms almost similarlry. We'll choose random forest for classification as the best model."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 10. Use this notebook to complete `train.py`\n",
    "Use the template file attached in the Resources folder to write a script that runs the steps above to create a database and export a model based on a new dataset specified by the user."
   ]
  },
  {
   "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": "3.7.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}