Basic needs-rights kenya.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Basic Needs Basic Rights Kenya - Tech4MentalHealth"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Around 1 in 4 people will experience a mental health problem this year. Low-income countries have an estimated treatment gap of 85% (as compared with high-income countries with a gap of 35% to 50%). While Kenya has a mental illness prevalence rate that is comparable to that of high-income countries, there are still less than 500 healthcare professionals serving the country.\n",
    "\n",
    "In Kenya, there are growing concerns about mental health among young people, particularly university students that face a challenging and unique conflation of stressors that put them at risk of challenges like depression and substance abuse.\n",
    "\n",
    "From the use of app-based solutions for screening to electronically delivered therapies, the use of technologies including machine learning and AI will potentially transform the delivery of mental health services in the coming years.\n",
    "\n",
    "The objective of this challenge is to develop a machine learning model that classifies statements and questions expressed by university students in Kenya when speaking about the mental health challenges they struggle with. The four categories are depression, suicide, alchoholism, and drug abuse."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'C:\\\\Users\\\\Admin\\\\Desktop\\\\R projects\\\\Basic Needs Basic Rights Kenya - Tech4MentalHealth'"
      ]
     },
     "execution_count": 1,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pwd"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "train = pd.read_csv('Train.csv')\n",
    "data = train.copy()\n",
    "data.drop('ID', axis=1, inplace=True)\n",
    "test = pd.read_csv('Test.csv')\n",
    "sample_sub = pd.read_csv('SampleSubmission.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "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>text</th>\n",
       "      <th>label</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>I feel that it was better I dieAm happy</td>\n",
       "      <td>Depression</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>Why do I get hallucinations?</td>\n",
       "      <td>Drugs</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>I am stresseed due to lack of financial suppor...</td>\n",
       "      <td>Depression</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>Why is life important?</td>\n",
       "      <td>Suicide</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>How could I be helped to go through the depres...</td>\n",
       "      <td>Depression</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                text       label\n",
       "0            I feel that it was better I dieAm happy  Depression\n",
       "1                       Why do I get hallucinations?       Drugs\n",
       "2  I am stresseed due to lack of financial suppor...  Depression\n",
       "3                             Why is life important?     Suicide\n",
       "4  How could I be helped to go through the depres...  Depression"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "text     False\n",
      "label    False\n",
      "dtype: bool\n"
     ]
    }
   ],
   "source": [
    "print(data.isnull().any())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### categorical encoding the label variable"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "text     object\n",
       "label    object\n",
       "dtype: object"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.dtypes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'str'>\n"
     ]
    }
   ],
   "source": [
    "data['label'] = data['label'].astype('category')\n",
    "print(type(data['label'][0]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    1\n",
       "1    2\n",
       "2    1\n",
       "3    3\n",
       "4    1\n",
       "Name: label_id, dtype: int8"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data['label_id'] = data['label'].cat.codes\n",
    "data['label_id'].head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>text</th>\n",
       "      <th>label</th>\n",
       "      <th>label_id</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>I feel that it was better I dieAm happy</td>\n",
       "      <td>Depression</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>Why do I get hallucinations?</td>\n",
       "      <td>Drugs</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>I am stresseed due to lack of financial suppor...</td>\n",
       "      <td>Depression</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>Why is life important?</td>\n",
       "      <td>Suicide</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>How could I be helped to go through the depres...</td>\n",
       "      <td>Depression</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>611</td>\n",
       "      <td>What should I do to stop alcoholism?</td>\n",
       "      <td>Alcohol</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>612</td>\n",
       "      <td>How to become my oldself again</td>\n",
       "      <td>Suicide</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>613</td>\n",
       "      <td>How can someone stop it?</td>\n",
       "      <td>Alcohol</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>614</td>\n",
       "      <td>I feel unworthy</td>\n",
       "      <td>Depression</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>615</td>\n",
       "      <td>I feel so discouraged with life</td>\n",
       "      <td>Depression</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>616 rows × 3 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                  text       label  label_id\n",
       "0              I feel that it was better I dieAm happy  Depression         1\n",
       "1                         Why do I get hallucinations?       Drugs         2\n",
       "2    I am stresseed due to lack of financial suppor...  Depression         1\n",
       "3                               Why is life important?     Suicide         3\n",
       "4    How could I be helped to go through the depres...  Depression         1\n",
       "..                                                 ...         ...       ...\n",
       "611               What should I do to stop alcoholism?     Alcohol         0\n",
       "612                     How to become my oldself again     Suicide         3\n",
       "613                           How can someone stop it?     Alcohol         0\n",
       "614                                   I feel unworthy   Depression         1\n",
       "615                    I feel so discouraged with life  Depression         1\n",
       "\n",
       "[616 rows x 3 columns]"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "scrolled": true
   },
   "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>ID</th>\n",
       "      <th>text</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>304</td>\n",
       "      <td>Z9A6ACLK</td>\n",
       "      <td>Yes</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>305</td>\n",
       "      <td>ZDUOIGKN</td>\n",
       "      <td>My girlfriend dumped me</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>306</td>\n",
       "      <td>ZHQ60CCH</td>\n",
       "      <td>How can I go back to being my old self?</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>307</td>\n",
       "      <td>ZVIJMA4O</td>\n",
       "      <td>Is it true bhang is  medicinal?</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>308</td>\n",
       "      <td>ZYIFAY98</td>\n",
       "      <td>how can I overcome the problem?</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "           ID                                     text\n",
       "304  Z9A6ACLK                                      Yes\n",
       "305  ZDUOIGKN                  My girlfriend dumped me\n",
       "306  ZHQ60CCH  How can I go back to being my old self?\n",
       "307  ZVIJMA4O          Is it true bhang is  medicinal?\n",
       "308  ZYIFAY98          how can I overcome the problem?"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test.tail()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[Depression, Drugs, Suicide, Alcohol]\n",
       "Categories (4, object): [Depression, Drugs, Suicide, Alcohol]"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.label.unique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "#\n",
    "fig = plt.figure(figsize=(8,6))\n",
    "data.groupby('label').text.count().plot.bar(ylim=0)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Text Representation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(616, 98)"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.feature_extraction.text import TfidfVectorizer\n",
    "tfid = TfidfVectorizer(sublinear_tf=True, min_df=5, norm='l2', encoding='latin-1', ngram_range=(1, 2), stop_words='english')\n",
    "features = tfid.fit_transform(data.text).toarray()\n",
    "target = data.label_id\n",
    "features.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.feature_extraction.text import CountVectorizer\n",
    "from sklearn.feature_extraction.text import TfidfTransformer\n",
    "from sklearn.naive_bayes import MultinomialNB\n",
    "import random\n",
    "random.seed(1234)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(data['text'], data['label_id'], random_state = 0)\n",
    "count_vect = CountVectorizer()\n",
    "X_train_counts = count_vect.fit_transform(X_train)\n",
    "tfidf_transformer = TfidfTransformer()\n",
    "X_train_tfidf = tfidf_transformer.fit_transform(X_train_counts)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Naive Bayes Classifier: the one most suitable for word counts is the multinomial variant"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "clf = MultinomialNB().fit(X_train_tfidf, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[1]\n"
     ]
    }
   ],
   "source": [
    "print(clf.predict(count_vect.transform([\"I feel that it was better I dieAm happy\"]))) # 1 represents depression"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Model Selection"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\Admin\\Anaconda3\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:432: FutureWarning: Default solver will be changed to 'lbfgs' in 0.22. Specify a solver to silence this warning.\n",
      "  FutureWarning)\n",
      "C:\\Users\\Admin\\Anaconda3\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:469: FutureWarning: Default multi_class will be changed to 'auto' in 0.22. Specify the multi_class option to silence this warning.\n",
      "  \"this warning.\", FutureWarning)\n",
      "C:\\Users\\Admin\\Anaconda3\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:432: FutureWarning: Default solver will be changed to 'lbfgs' in 0.22. Specify a solver to silence this warning.\n",
      "  FutureWarning)\n",
      "C:\\Users\\Admin\\Anaconda3\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:469: FutureWarning: Default multi_class will be changed to 'auto' in 0.22. Specify the multi_class option to silence this warning.\n",
      "  \"this warning.\", FutureWarning)\n",
      "C:\\Users\\Admin\\Anaconda3\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:432: FutureWarning: Default solver will be changed to 'lbfgs' in 0.22. Specify a solver to silence this warning.\n",
      "  FutureWarning)\n",
      "C:\\Users\\Admin\\Anaconda3\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:469: FutureWarning: Default multi_class will be changed to 'auto' in 0.22. Specify the multi_class option to silence this warning.\n",
      "  \"this warning.\", FutureWarning)\n",
      "C:\\Users\\Admin\\Anaconda3\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:432: FutureWarning: Default solver will be changed to 'lbfgs' in 0.22. Specify a solver to silence this warning.\n",
      "  FutureWarning)\n",
      "C:\\Users\\Admin\\Anaconda3\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:469: FutureWarning: Default multi_class will be changed to 'auto' in 0.22. Specify the multi_class option to silence this warning.\n",
      "  \"this warning.\", FutureWarning)\n",
      "C:\\Users\\Admin\\Anaconda3\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:432: FutureWarning: Default solver will be changed to 'lbfgs' in 0.22. Specify a solver to silence this warning.\n",
      "  FutureWarning)\n",
      "C:\\Users\\Admin\\Anaconda3\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:469: FutureWarning: Default multi_class will be changed to 'auto' in 0.22. Specify the multi_class option to silence this warning.\n",
      "  \"this warning.\", FutureWarning)\n"
     ]
    }
   ],
   "source": [
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.ensemble import RandomForestClassifier, AdaBoostClassifier, GradientBoostingClassifier\n",
    "from sklearn.svm import LinearSVC\n",
    "from catboost import CatBoostClassifier\n",
    "from xgboost import XGBClassifier\n",
    "from sklearn.neighbors import KNeighborsClassifier\n",
    "from sklearn.model_selection import cross_val_score\n",
    "#\n",
    "models = [\n",
    "    RandomForestClassifier(n_estimators=200, max_depth=3, random_state=0),\n",
    "    LinearSVC(),\n",
    "    MultinomialNB(),\n",
    "    LogisticRegression(random_state=0),\n",
    "    AdaBoostClassifier(),\n",
    "    CatBoostClassifier(iterations=300, depth=6, learning_rate=0.1, logging_level='Silent',\n",
    "                        eval_metric='Accuracy', random_state=42),\n",
    "    KNeighborsClassifier(),\n",
    "    GradientBoostingClassifier(),\n",
    "    XGBClassifier(objective='multi:softprob', eval_metric='logloss', learning_rate=0.1, max_depth=4, seed=27,\n",
    "                  n_estimators=20)\n",
    "]\n",
    "#\n",
    "CV = 5\n",
    "cv_df = pd.DataFrame(index=range(CV * len(models)))\n",
    "entries = []\n",
    "for model in models:\n",
    "    model_name = model.__class__.__name__\n",
    "    accuracies = cross_val_score(model, features, target, scoring='accuracy', cv=CV)\n",
    "    for fold_idx, accuracy in enumerate(accuracies):\n",
    "        entries.append((model_name, fold_idx, accuracy))\n",
    "cv_df = pd.DataFrame(entries, columns=['model_name', 'fold_idx', 'accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>model_name</th>\n",
       "      <th>fold_idx</th>\n",
       "      <th>accuracy</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>RandomForestClassifier</td>\n",
       "      <td>0</td>\n",
       "      <td>0.680000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>RandomForestClassifier</td>\n",
       "      <td>1</td>\n",
       "      <td>0.653226</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>RandomForestClassifier</td>\n",
       "      <td>2</td>\n",
       "      <td>0.674797</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>RandomForestClassifier</td>\n",
       "      <td>3</td>\n",
       "      <td>0.639344</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>RandomForestClassifier</td>\n",
       "      <td>4</td>\n",
       "      <td>0.631148</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>5</td>\n",
       "      <td>LinearSVC</td>\n",
       "      <td>0</td>\n",
       "      <td>0.848000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>6</td>\n",
       "      <td>LinearSVC</td>\n",
       "      <td>1</td>\n",
       "      <td>0.862903</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>7</td>\n",
       "      <td>LinearSVC</td>\n",
       "      <td>2</td>\n",
       "      <td>0.821138</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>8</td>\n",
       "      <td>LinearSVC</td>\n",
       "      <td>3</td>\n",
       "      <td>0.811475</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>9</td>\n",
       "      <td>LinearSVC</td>\n",
       "      <td>4</td>\n",
       "      <td>0.819672</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>10</td>\n",
       "      <td>MultinomialNB</td>\n",
       "      <td>0</td>\n",
       "      <td>0.848000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>11</td>\n",
       "      <td>MultinomialNB</td>\n",
       "      <td>1</td>\n",
       "      <td>0.846774</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>12</td>\n",
       "      <td>MultinomialNB</td>\n",
       "      <td>2</td>\n",
       "      <td>0.804878</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>13</td>\n",
       "      <td>MultinomialNB</td>\n",
       "      <td>3</td>\n",
       "      <td>0.836066</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>14</td>\n",
       "      <td>MultinomialNB</td>\n",
       "      <td>4</td>\n",
       "      <td>0.844262</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>15</td>\n",
       "      <td>LogisticRegression</td>\n",
       "      <td>0</td>\n",
       "      <td>0.864000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>16</td>\n",
       "      <td>LogisticRegression</td>\n",
       "      <td>1</td>\n",
       "      <td>0.862903</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>17</td>\n",
       "      <td>LogisticRegression</td>\n",
       "      <td>2</td>\n",
       "      <td>0.796748</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>18</td>\n",
       "      <td>LogisticRegression</td>\n",
       "      <td>3</td>\n",
       "      <td>0.811475</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>19</td>\n",
       "      <td>LogisticRegression</td>\n",
       "      <td>4</td>\n",
       "      <td>0.827869</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>20</td>\n",
       "      <td>AdaBoostClassifier</td>\n",
       "      <td>0</td>\n",
       "      <td>0.840000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>21</td>\n",
       "      <td>AdaBoostClassifier</td>\n",
       "      <td>1</td>\n",
       "      <td>0.854839</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>22</td>\n",
       "      <td>AdaBoostClassifier</td>\n",
       "      <td>2</td>\n",
       "      <td>0.747967</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>23</td>\n",
       "      <td>AdaBoostClassifier</td>\n",
       "      <td>3</td>\n",
       "      <td>0.762295</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>24</td>\n",
       "      <td>AdaBoostClassifier</td>\n",
       "      <td>4</td>\n",
       "      <td>0.778689</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>25</td>\n",
       "      <td>CatBoostClassifier</td>\n",
       "      <td>0</td>\n",
       "      <td>0.872000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>26</td>\n",
       "      <td>CatBoostClassifier</td>\n",
       "      <td>1</td>\n",
       "      <td>0.854839</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>27</td>\n",
       "      <td>CatBoostClassifier</td>\n",
       "      <td>2</td>\n",
       "      <td>0.788618</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>28</td>\n",
       "      <td>CatBoostClassifier</td>\n",
       "      <td>3</td>\n",
       "      <td>0.786885</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>29</td>\n",
       "      <td>CatBoostClassifier</td>\n",
       "      <td>4</td>\n",
       "      <td>0.795082</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>30</td>\n",
       "      <td>KNeighborsClassifier</td>\n",
       "      <td>0</td>\n",
       "      <td>0.792000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>31</td>\n",
       "      <td>KNeighborsClassifier</td>\n",
       "      <td>1</td>\n",
       "      <td>0.782258</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>32</td>\n",
       "      <td>KNeighborsClassifier</td>\n",
       "      <td>2</td>\n",
       "      <td>0.780488</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>33</td>\n",
       "      <td>KNeighborsClassifier</td>\n",
       "      <td>3</td>\n",
       "      <td>0.786885</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>34</td>\n",
       "      <td>KNeighborsClassifier</td>\n",
       "      <td>4</td>\n",
       "      <td>0.745902</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>35</td>\n",
       "      <td>GradientBoostingClassifier</td>\n",
       "      <td>0</td>\n",
       "      <td>0.856000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>36</td>\n",
       "      <td>GradientBoostingClassifier</td>\n",
       "      <td>1</td>\n",
       "      <td>0.862903</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>37</td>\n",
       "      <td>GradientBoostingClassifier</td>\n",
       "      <td>2</td>\n",
       "      <td>0.764228</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>38</td>\n",
       "      <td>GradientBoostingClassifier</td>\n",
       "      <td>3</td>\n",
       "      <td>0.795082</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>39</td>\n",
       "      <td>GradientBoostingClassifier</td>\n",
       "      <td>4</td>\n",
       "      <td>0.786885</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>40</td>\n",
       "      <td>XGBClassifier</td>\n",
       "      <td>0</td>\n",
       "      <td>0.848000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>41</td>\n",
       "      <td>XGBClassifier</td>\n",
       "      <td>1</td>\n",
       "      <td>0.854839</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>42</td>\n",
       "      <td>XGBClassifier</td>\n",
       "      <td>2</td>\n",
       "      <td>0.780488</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>43</td>\n",
       "      <td>XGBClassifier</td>\n",
       "      <td>3</td>\n",
       "      <td>0.803279</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>44</td>\n",
       "      <td>XGBClassifier</td>\n",
       "      <td>4</td>\n",
       "      <td>0.819672</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                    model_name  fold_idx  accuracy\n",
       "0       RandomForestClassifier         0  0.680000\n",
       "1       RandomForestClassifier         1  0.653226\n",
       "2       RandomForestClassifier         2  0.674797\n",
       "3       RandomForestClassifier         3  0.639344\n",
       "4       RandomForestClassifier         4  0.631148\n",
       "5                    LinearSVC         0  0.848000\n",
       "6                    LinearSVC         1  0.862903\n",
       "7                    LinearSVC         2  0.821138\n",
       "8                    LinearSVC         3  0.811475\n",
       "9                    LinearSVC         4  0.819672\n",
       "10               MultinomialNB         0  0.848000\n",
       "11               MultinomialNB         1  0.846774\n",
       "12               MultinomialNB         2  0.804878\n",
       "13               MultinomialNB         3  0.836066\n",
       "14               MultinomialNB         4  0.844262\n",
       "15          LogisticRegression         0  0.864000\n",
       "16          LogisticRegression         1  0.862903\n",
       "17          LogisticRegression         2  0.796748\n",
       "18          LogisticRegression         3  0.811475\n",
       "19          LogisticRegression         4  0.827869\n",
       "20          AdaBoostClassifier         0  0.840000\n",
       "21          AdaBoostClassifier         1  0.854839\n",
       "22          AdaBoostClassifier         2  0.747967\n",
       "23          AdaBoostClassifier         3  0.762295\n",
       "24          AdaBoostClassifier         4  0.778689\n",
       "25          CatBoostClassifier         0  0.872000\n",
       "26          CatBoostClassifier         1  0.854839\n",
       "27          CatBoostClassifier         2  0.788618\n",
       "28          CatBoostClassifier         3  0.786885\n",
       "29          CatBoostClassifier         4  0.795082\n",
       "30        KNeighborsClassifier         0  0.792000\n",
       "31        KNeighborsClassifier         1  0.782258\n",
       "32        KNeighborsClassifier         2  0.780488\n",
       "33        KNeighborsClassifier         3  0.786885\n",
       "34        KNeighborsClassifier         4  0.745902\n",
       "35  GradientBoostingClassifier         0  0.856000\n",
       "36  GradientBoostingClassifier         1  0.862903\n",
       "37  GradientBoostingClassifier         2  0.764228\n",
       "38  GradientBoostingClassifier         3  0.795082\n",
       "39  GradientBoostingClassifier         4  0.786885\n",
       "40               XGBClassifier         0  0.848000\n",
       "41               XGBClassifier         1  0.854839\n",
       "42               XGBClassifier         2  0.780488\n",
       "43               XGBClassifier         3  0.803279\n",
       "44               XGBClassifier         4  0.819672"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cv_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import seaborn as sns\n",
    "fig, ax = plt.subplots()\n",
    "fig.set_size_inches(20,5)\n",
    "sns.boxplot(x='model_name', y='accuracy', data=cv_df)\n",
    "sns.stripplot(x='model_name', y='accuracy', data=cv_df, \n",
    "              size=8, jitter=True, edgecolor=\"gray\", linewidth=2)\n",
    "ax.set_xticklabels(ax.get_xticklabels(), rotation=90)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "model_name\n",
      "AdaBoostClassifier            0.796758\n",
      "CatBoostClassifier            0.819485\n",
      "GradientBoostingClassifier    0.813020\n",
      "KNeighborsClassifier          0.777507\n",
      "LinearSVC                     0.832638\n",
      "LogisticRegression            0.832599\n",
      "MultinomialNB                 0.835996\n",
      "RandomForestClassifier        0.655703\n",
      "XGBClassifier                 0.821255\n",
      "Name: accuracy, dtype: float64\n"
     ]
    }
   ],
   "source": [
    "print(cv_df.groupby('model_name').accuracy.mean())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Model Evaluation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.5693392654702557\n",
      "[[ 34  11   0   0]\n",
      " [  1 114   0   0]\n",
      " [  1   5   4   0]\n",
      " [  0  14   0   1]]\n"
     ]
    }
   ],
   "source": [
    "model = MultinomialNB()\n",
    "X_train, X_test, y_train, y_test = train_test_split(features, target, test_size=0.30, random_state=0)\n",
    "model.fit(X_train, y_train)\n",
    "y_pred = model.predict(X_test)\n",
    "y_prob = model.predict_proba(X_test)\n",
    "#\n",
    "# get the logarithmic loss for the x_test\n",
    "from sklearn.metrics import log_loss\n",
    "print(log_loss(y_test, y_prob)) # log los\n",
    "#\n",
    "from sklearn.metrics import confusion_matrix\n",
    "conf_mat = confusion_matrix(y_test, y_pred)\n",
    "print(conf_mat)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.5711954794366941\n"
     ]
    }
   ],
   "source": [
    "#catboost\n",
    "cat = CatBoostClassifier(iterations=300, depth=6, learning_rate=0.1, logging_level='Silent',\n",
    "                        eval_metric='Accuracy', random_state=42)\n",
    "cat.fit(X_train, y_train)\n",
    "cat_prob = cat.predict_proba(X_test)\n",
    "print(log_loss(y_test, cat_prob))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.6737812629422626\n"
     ]
    }
   ],
   "source": [
    "#xgboost\n",
    "xgb = XGBClassifier(objective='multi:softprob', eval_metric='logloss', learning_rate=0.1, max_depth=4, seed=27,\n",
    "                  n_estimators=20)\n",
    "xgb.fit(X_train, y_train)\n",
    "xgb_prob = xgb.predict_proba(X_test)\n",
    "print(log_loss(y_test, xgb_prob))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Test Data Text Representation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "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>text</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>How to overcome bad feelings and emotions</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>I feel like giving up in life</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>I was so depressed feel like got no strength t...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>I feel so low especially since I had no one to...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>can i be successful when I am a drug addict?</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>304</td>\n",
       "      <td>Yes</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>305</td>\n",
       "      <td>My girlfriend dumped me</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>306</td>\n",
       "      <td>How can I go back to being my old self?</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>307</td>\n",
       "      <td>Is it true bhang is  medicinal?</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>308</td>\n",
       "      <td>how can I overcome the problem?</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>309 rows × 1 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                  text\n",
       "0            How to overcome bad feelings and emotions\n",
       "1                       I feel like giving up in life \n",
       "2    I was so depressed feel like got no strength t...\n",
       "3    I feel so low especially since I had no one to...\n",
       "4         can i be successful when I am a drug addict?\n",
       "..                                                 ...\n",
       "304                                                Yes\n",
       "305                            My girlfriend dumped me\n",
       "306            How can I go back to being my old self?\n",
       "307                    Is it true bhang is  medicinal?\n",
       "308                    how can I overcome the problem?\n",
       "\n",
       "[309 rows x 1 columns]"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test.drop('ID', axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[nltk_data] Downloading package stopwords to\n",
      "[nltk_data]     C:\\Users\\Admin\\AppData\\Roaming\\nltk_data...\n",
      "[nltk_data]   Package stopwords is already up-to-date!\n"
     ]
    }
   ],
   "source": [
    "# clean the test data\n",
    "import re\n",
    "import nltk\n",
    "nltk.download('stopwords')\n",
    "from nltk.corpus import stopwords\n",
    "from nltk.stem.porter import PorterStemmer\n",
    "#\n",
    "corpus = []\n",
    "for i in range(0, len(test)):\n",
    "    clean = re.sub('[^a-zA-Z]', ' ', test['text'][i])\n",
    "    clean = clean.lower().split()\n",
    "    ps = PorterStemmer()\n",
    "    # return base words iff they are not english wods\n",
    "    clean = [ps.stem(word) for word in clean if not word in set(stopwords.words('english'))] \n",
    "    clean = ''.join(clean)\n",
    "    corpus.append(clean)\n",
    "#    \n",
    "# tokenize the corpus\n",
    "from sklearn.feature_extraction.text import CountVectorizer\n",
    "cv = CountVectorizer(max_features=98)\n",
    "valid_test = cv.fit_transform(corpus).toarray()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(309, 98)"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "valid_test.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "valid_test = pd.DataFrame(valid_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "test_pred = model.predict(valid_test)\n",
    "test_prob = model.predict_proba(valid_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [],
   "source": [
    "names = ['Depression', 'Drugs', 'Suicide', 'Alcohol']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "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>Depression</th>\n",
       "      <th>Drugs</th>\n",
       "      <th>Suicide</th>\n",
       "      <th>Alcohol</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>0.091017</td>\n",
       "      <td>0.694906</td>\n",
       "      <td>0.066255</td>\n",
       "      <td>0.147822</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>0.220418</td>\n",
       "      <td>0.549884</td>\n",
       "      <td>0.111369</td>\n",
       "      <td>0.118329</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>0.220418</td>\n",
       "      <td>0.549884</td>\n",
       "      <td>0.111369</td>\n",
       "      <td>0.118329</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>0.220418</td>\n",
       "      <td>0.549884</td>\n",
       "      <td>0.111369</td>\n",
       "      <td>0.118329</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>0.436282</td>\n",
       "      <td>0.217891</td>\n",
       "      <td>0.226353</td>\n",
       "      <td>0.119474</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Depression     Drugs   Suicide   Alcohol\n",
       "0    0.091017  0.694906  0.066255  0.147822\n",
       "1    0.220418  0.549884  0.111369  0.118329\n",
       "2    0.220418  0.549884  0.111369  0.118329\n",
       "3    0.220418  0.549884  0.111369  0.118329\n",
       "4    0.436282  0.217891  0.226353  0.119474"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pr = pd.DataFrame(test_prob, columns=names)\n",
    "pr.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {
    "scrolled": true
   },
   "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>ID</th>\n",
       "      <th>Depression</th>\n",
       "      <th>Drugs</th>\n",
       "      <th>Suicide</th>\n",
       "      <th>Alcohol</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>02V56KMO</td>\n",
       "      <td>0.091017</td>\n",
       "      <td>0.694906</td>\n",
       "      <td>0.066255</td>\n",
       "      <td>0.147822</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>03BMGTOK</td>\n",
       "      <td>0.220418</td>\n",
       "      <td>0.549884</td>\n",
       "      <td>0.111369</td>\n",
       "      <td>0.118329</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>03LZVFM6</td>\n",
       "      <td>0.220418</td>\n",
       "      <td>0.549884</td>\n",
       "      <td>0.111369</td>\n",
       "      <td>0.118329</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>0EPULUM5</td>\n",
       "      <td>0.220418</td>\n",
       "      <td>0.549884</td>\n",
       "      <td>0.111369</td>\n",
       "      <td>0.118329</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>0GM4C5GD</td>\n",
       "      <td>0.436282</td>\n",
       "      <td>0.217891</td>\n",
       "      <td>0.226353</td>\n",
       "      <td>0.119474</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         ID  Depression     Drugs   Suicide   Alcohol\n",
       "0  02V56KMO    0.091017  0.694906  0.066255  0.147822\n",
       "1  03BMGTOK    0.220418  0.549884  0.111369  0.118329\n",
       "2  03LZVFM6    0.220418  0.549884  0.111369  0.118329\n",
       "3  0EPULUM5    0.220418  0.549884  0.111369  0.118329\n",
       "4  0GM4C5GD    0.436282  0.217891  0.226353  0.119474"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = pd.concat([test.ID, pr], axis=1)\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "df.to_csv('Multinormial_Naivebayes.csv', index = False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [],
   "source": [
    "###############################################################################################"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}