Fadhel Ijlal Falah_m3_s22_text_mining.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Exercise: Text Mining\n",
    "\n",
    " \n",
    "\n",
    "Gunakan dataset Tweets.csv\n",
    "\n",
    " \n",
    "\n",
    "Dataset ini berisi tweet yang ditujukan (mention) ke beberapa maskapai di US. \n",
    "\n",
    "Anda bekerja sebagai data scientist yang ditugaskan untuk memprediksi apakah sebuah tweet termasuk ke dalam sentiment negative (class 1) atau positive or neutral (class 0). \n",
    "\n",
    "Tujuan dari klasifikasi tweet ini adalah untuk memberikan alarm/warning ketika muncul sebuah tweet negatif sehingga tim PR (Public Relation) bisa menghandle tweet negatif tsb dengan cepat.\n",
    "\n",
    " \n",
    "\n",
    "- Lakukan preprocessing pada semua tweet \n",
    "\n",
    "- Lakukan EDA. Apakah ada perbedaan antara tweet dengan sentiment negative (class 1) dengan tweet dengan sentiment positive or neutral (class 0)?\n",
    "\n",
    "- Definisikan FP dan FN dalam kasus sentiment analysis ini\n",
    "\n",
    "- Tentukan metric yang cocok untuk kasus ini"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "# general librariers\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "# visualisations\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>tweet_id</th>\n",
       "      <th>airline_sentiment</th>\n",
       "      <th>airline_sentiment_confidence</th>\n",
       "      <th>negativereason</th>\n",
       "      <th>negativereason_confidence</th>\n",
       "      <th>airline</th>\n",
       "      <th>airline_sentiment_gold</th>\n",
       "      <th>name</th>\n",
       "      <th>negativereason_gold</th>\n",
       "      <th>retweet_count</th>\n",
       "      <th>text</th>\n",
       "      <th>tweet_coord</th>\n",
       "      <th>tweet_created</th>\n",
       "      <th>tweet_location</th>\n",
       "      <th>user_timezone</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>570306133677760513</td>\n",
       "      <td>neutral</td>\n",
       "      <td>1.0000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Virgin America</td>\n",
       "      <td>NaN</td>\n",
       "      <td>cairdin</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>@VirginAmerica What @dhepburn said.</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2015-02-24 11:35:52 -0800</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Eastern Time (US &amp; Canada)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>570301130888122368</td>\n",
       "      <td>positive</td>\n",
       "      <td>0.3486</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>Virgin America</td>\n",
       "      <td>NaN</td>\n",
       "      <td>jnardino</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>@VirginAmerica plus you've added commercials t...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2015-02-24 11:15:59 -0800</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Pacific Time (US &amp; Canada)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>570301083672813571</td>\n",
       "      <td>neutral</td>\n",
       "      <td>0.6837</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Virgin America</td>\n",
       "      <td>NaN</td>\n",
       "      <td>yvonnalynn</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>@VirginAmerica I didn't today... Must mean I n...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2015-02-24 11:15:48 -0800</td>\n",
       "      <td>Lets Play</td>\n",
       "      <td>Central Time (US &amp; Canada)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>570301031407624196</td>\n",
       "      <td>negative</td>\n",
       "      <td>1.0000</td>\n",
       "      <td>Bad Flight</td>\n",
       "      <td>0.7033</td>\n",
       "      <td>Virgin America</td>\n",
       "      <td>NaN</td>\n",
       "      <td>jnardino</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>@VirginAmerica it's really aggressive to blast...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2015-02-24 11:15:36 -0800</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Pacific Time (US &amp; Canada)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>570300817074462722</td>\n",
       "      <td>negative</td>\n",
       "      <td>1.0000</td>\n",
       "      <td>Can't Tell</td>\n",
       "      <td>1.0000</td>\n",
       "      <td>Virgin America</td>\n",
       "      <td>NaN</td>\n",
       "      <td>jnardino</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>@VirginAmerica and it's a really big bad thing...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2015-02-24 11:14:45 -0800</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Pacific Time (US &amp; Canada)</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             tweet_id airline_sentiment  airline_sentiment_confidence  \\\n",
       "0  570306133677760513           neutral                        1.0000   \n",
       "1  570301130888122368          positive                        0.3486   \n",
       "2  570301083672813571           neutral                        0.6837   \n",
       "3  570301031407624196          negative                        1.0000   \n",
       "4  570300817074462722          negative                        1.0000   \n",
       "\n",
       "  negativereason  negativereason_confidence         airline  \\\n",
       "0            NaN                        NaN  Virgin America   \n",
       "1            NaN                     0.0000  Virgin America   \n",
       "2            NaN                        NaN  Virgin America   \n",
       "3     Bad Flight                     0.7033  Virgin America   \n",
       "4     Can't Tell                     1.0000  Virgin America   \n",
       "\n",
       "  airline_sentiment_gold        name negativereason_gold  retweet_count  \\\n",
       "0                    NaN     cairdin                 NaN              0   \n",
       "1                    NaN    jnardino                 NaN              0   \n",
       "2                    NaN  yvonnalynn                 NaN              0   \n",
       "3                    NaN    jnardino                 NaN              0   \n",
       "4                    NaN    jnardino                 NaN              0   \n",
       "\n",
       "                                                text tweet_coord  \\\n",
       "0                @VirginAmerica What @dhepburn said.         NaN   \n",
       "1  @VirginAmerica plus you've added commercials t...         NaN   \n",
       "2  @VirginAmerica I didn't today... Must mean I n...         NaN   \n",
       "3  @VirginAmerica it's really aggressive to blast...         NaN   \n",
       "4  @VirginAmerica and it's a really big bad thing...         NaN   \n",
       "\n",
       "               tweet_created tweet_location               user_timezone  \n",
       "0  2015-02-24 11:35:52 -0800            NaN  Eastern Time (US & Canada)  \n",
       "1  2015-02-24 11:15:59 -0800            NaN  Pacific Time (US & Canada)  \n",
       "2  2015-02-24 11:15:48 -0800      Lets Play  Central Time (US & Canada)  \n",
       "3  2015-02-24 11:15:36 -0800            NaN  Pacific Time (US & Canada)  \n",
       "4  2015-02-24 11:14:45 -0800            NaN  Pacific Time (US & Canada)  "
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = pd.read_csv('../datasets/Tweets.csv')\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "airline\n",
       "United            3822\n",
       "US Airways        2913\n",
       "American          2759\n",
       "Southwest         2420\n",
       "Delta             2222\n",
       "Virgin America     504\n",
       "Name: count, dtype: int64"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['airline'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "df['airline_sentiment'] = np.where((df['airline_sentiment'] == 'neutral'), 'positive', df['airline_sentiment'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "airline_sentiment\n",
       "negative    9178\n",
       "positive    5462\n",
       "Name: count, dtype: int64"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['airline_sentiment'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "airline_sentiment\n",
       "negative    9178\n",
       "positive    5462\n",
       "Name: count, dtype: int64"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['airline_sentiment'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tweet_id                            0\n",
       "airline_sentiment                   0\n",
       "airline_sentiment_confidence        0\n",
       "negativereason                   5462\n",
       "negativereason_confidence        4118\n",
       "airline                             0\n",
       "airline_sentiment_gold          14600\n",
       "name                                0\n",
       "negativereason_gold             14608\n",
       "retweet_count                       0\n",
       "text                                0\n",
       "tweet_coord                     13621\n",
       "tweet_created                       0\n",
       "tweet_location                   4733\n",
       "user_timezone                    4820\n",
       "dtype: int64"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.isna().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tweet_id                         0.000000\n",
       "airline_sentiment                0.000000\n",
       "airline_sentiment_confidence     0.000000\n",
       "negativereason                  37.308743\n",
       "negativereason_confidence       28.128415\n",
       "airline                          0.000000\n",
       "airline_sentiment_gold          99.726776\n",
       "name                             0.000000\n",
       "negativereason_gold             99.781421\n",
       "retweet_count                    0.000000\n",
       "text                             0.000000\n",
       "tweet_coord                     93.039617\n",
       "tweet_created                    0.000000\n",
       "tweet_location                  32.329235\n",
       "user_timezone                   32.923497\n",
       "dtype: float64"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.isna().sum()/df.shape[0]*100"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "37"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.duplicated().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>airline_sentiment</th>\n",
       "      <th>text</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>positive</td>\n",
       "      <td>@VirginAmerica What @dhepburn said.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>positive</td>\n",
       "      <td>@VirginAmerica plus you've added commercials t...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>positive</td>\n",
       "      <td>@VirginAmerica I didn't today... Must mean I n...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>negative</td>\n",
       "      <td>@VirginAmerica it's really aggressive to blast...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>negative</td>\n",
       "      <td>@VirginAmerica and it's a really big bad thing...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  airline_sentiment                                               text\n",
       "0          positive                @VirginAmerica What @dhepburn said.\n",
       "1          positive  @VirginAmerica plus you've added commercials t...\n",
       "2          positive  @VirginAmerica I didn't today... Must mean I n...\n",
       "3          negative  @VirginAmerica it's really aggressive to blast...\n",
       "4          negative  @VirginAmerica and it's a really big bad thing..."
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = df[['airline_sentiment', 'text']]\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Preprocessing"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 1. Lowercase"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "to_lower = lambda text: text.lower()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 2. Contraction"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "import contractions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "# contractions.fix()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 3. Remove Number"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "def remove_number(text): return ''.join(i for i in text if not i.isdigit())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 4. Remove Punctuation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "from string import punctuation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'!\"#$%&\\'()*+,-./:;<=>?@[\\\\]^_`{|}~'"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "punctuation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "import re\n",
    "def remove_punctuation(text): \n",
    "    res = re.sub('@[^\\s]+','',text)\n",
    "    res = ''.join(i for i in res if i not in punctuation)\n",
    "    return res"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'I do not care'"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "contractions.fix(\"I don't care\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 5. Remove Stopwords"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "import nltk\n",
    "from nltk.corpus import stopwords"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['i',\n",
       " 'me',\n",
       " 'my',\n",
       " 'myself',\n",
       " 'we',\n",
       " 'our',\n",
       " 'ours',\n",
       " 'ourselves',\n",
       " 'you',\n",
       " \"you're\",\n",
       " \"you've\",\n",
       " \"you'll\",\n",
       " \"you'd\",\n",
       " 'your',\n",
       " 'yours',\n",
       " 'yourself',\n",
       " 'yourselves',\n",
       " 'he',\n",
       " 'him',\n",
       " 'his',\n",
       " 'himself',\n",
       " 'she',\n",
       " \"she's\",\n",
       " 'her',\n",
       " 'hers',\n",
       " 'herself',\n",
       " 'it',\n",
       " \"it's\",\n",
       " 'its',\n",
       " 'itself',\n",
       " 'they',\n",
       " 'them',\n",
       " 'their',\n",
       " 'theirs',\n",
       " 'themselves',\n",
       " 'what',\n",
       " 'which',\n",
       " 'who',\n",
       " 'whom',\n",
       " 'this',\n",
       " 'that',\n",
       " \"that'll\",\n",
       " 'these',\n",
       " 'those',\n",
       " 'am',\n",
       " 'is',\n",
       " 'are',\n",
       " 'was',\n",
       " 'were',\n",
       " 'be',\n",
       " 'been',\n",
       " 'being',\n",
       " 'have',\n",
       " 'has',\n",
       " 'had',\n",
       " 'having',\n",
       " 'do',\n",
       " 'does',\n",
       " 'did',\n",
       " 'doing',\n",
       " 'a',\n",
       " 'an',\n",
       " 'the',\n",
       " 'and',\n",
       " 'but',\n",
       " 'if',\n",
       " 'or',\n",
       " 'because',\n",
       " 'as',\n",
       " 'until',\n",
       " 'while',\n",
       " 'of',\n",
       " 'at',\n",
       " 'by',\n",
       " 'for',\n",
       " 'with',\n",
       " 'about',\n",
       " 'against',\n",
       " 'between',\n",
       " 'into',\n",
       " 'through',\n",
       " 'during',\n",
       " 'before',\n",
       " 'after',\n",
       " 'above',\n",
       " 'below',\n",
       " 'to',\n",
       " 'from',\n",
       " 'up',\n",
       " 'down',\n",
       " 'in',\n",
       " 'out',\n",
       " 'on',\n",
       " 'off',\n",
       " 'over',\n",
       " 'under',\n",
       " 'again',\n",
       " 'further',\n",
       " 'then',\n",
       " 'once',\n",
       " 'here',\n",
       " 'there',\n",
       " 'when',\n",
       " 'where',\n",
       " 'why',\n",
       " 'how',\n",
       " 'all',\n",
       " 'any',\n",
       " 'both',\n",
       " 'each',\n",
       " 'few',\n",
       " 'more',\n",
       " 'most',\n",
       " 'other',\n",
       " 'some',\n",
       " 'such',\n",
       " 'no',\n",
       " 'nor',\n",
       " 'not',\n",
       " 'only',\n",
       " 'own',\n",
       " 'same',\n",
       " 'so',\n",
       " 'than',\n",
       " 'too',\n",
       " 'very',\n",
       " 's',\n",
       " 't',\n",
       " 'can',\n",
       " 'will',\n",
       " 'just',\n",
       " 'don',\n",
       " \"don't\",\n",
       " 'should',\n",
       " \"should've\",\n",
       " 'now',\n",
       " 'd',\n",
       " 'll',\n",
       " 'm',\n",
       " 'o',\n",
       " 're',\n",
       " 've',\n",
       " 'y',\n",
       " 'ain',\n",
       " 'aren',\n",
       " \"aren't\",\n",
       " 'couldn',\n",
       " \"couldn't\",\n",
       " 'didn',\n",
       " \"didn't\",\n",
       " 'doesn',\n",
       " \"doesn't\",\n",
       " 'hadn',\n",
       " \"hadn't\",\n",
       " 'hasn',\n",
       " \"hasn't\",\n",
       " 'haven',\n",
       " \"haven't\",\n",
       " 'isn',\n",
       " \"isn't\",\n",
       " 'ma',\n",
       " 'mightn',\n",
       " \"mightn't\",\n",
       " 'mustn',\n",
       " \"mustn't\",\n",
       " 'needn',\n",
       " \"needn't\",\n",
       " 'shan',\n",
       " \"shan't\",\n",
       " 'shouldn',\n",
       " \"shouldn't\",\n",
       " 'wasn',\n",
       " \"wasn't\",\n",
       " 'weren',\n",
       " \"weren't\",\n",
       " 'won',\n",
       " \"won't\",\n",
       " 'wouldn',\n",
       " \"wouldn't\"]"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "stop_words = stopwords.words('english')\n",
    "stop_words"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "# stop_words.remove(\"not\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "stop_words.append('u')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "def remove_stopwords(text): \n",
    "    return ' '.join(i for i in text.split() if i not in stop_words)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 6. Remove Whitespace"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "def remove_whitespace(text):\n",
    "    return ' '.join(text.split())"
   ]
  },
  {
   "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>airline_sentiment</th>\n",
       "      <th>text</th>\n",
       "      <th>1_lower</th>\n",
       "      <th>2_contraction</th>\n",
       "      <th>3_number</th>\n",
       "      <th>4_punct</th>\n",
       "      <th>5_stopword</th>\n",
       "      <th>6_white</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>positive</td>\n",
       "      <td>@VirginAmerica What @dhepburn said.</td>\n",
       "      <td>@virginamerica what @dhepburn said.</td>\n",
       "      <td>@virginamerica what @dhepburn said.</td>\n",
       "      <td>@virginamerica what @dhepburn said.</td>\n",
       "      <td>what  said</td>\n",
       "      <td>said</td>\n",
       "      <td>said</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>positive</td>\n",
       "      <td>@VirginAmerica plus you've added commercials t...</td>\n",
       "      <td>@virginamerica plus you've added commercials t...</td>\n",
       "      <td>@virginamerica plus you have added commercials...</td>\n",
       "      <td>@virginamerica plus you have added commercials...</td>\n",
       "      <td>plus you have added commercials to the experi...</td>\n",
       "      <td>plus added commercials experience tacky</td>\n",
       "      <td>plus added commercials experience tacky</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>positive</td>\n",
       "      <td>@VirginAmerica I didn't today... Must mean I n...</td>\n",
       "      <td>@virginamerica i didn't today... must mean i n...</td>\n",
       "      <td>@virginamerica i did not today... must mean i ...</td>\n",
       "      <td>@virginamerica i did not today... must mean i ...</td>\n",
       "      <td>i did not today must mean i need to take anot...</td>\n",
       "      <td>today must mean need take another trip</td>\n",
       "      <td>today must mean need take another trip</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>negative</td>\n",
       "      <td>@VirginAmerica it's really aggressive to blast...</td>\n",
       "      <td>@virginamerica it's really aggressive to blast...</td>\n",
       "      <td>@virginamerica it is really aggressive to blas...</td>\n",
       "      <td>@virginamerica it is really aggressive to blas...</td>\n",
       "      <td>it is really aggressive to blast obnoxious en...</td>\n",
       "      <td>really aggressive blast obnoxious entertainmen...</td>\n",
       "      <td>really aggressive blast obnoxious entertainmen...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>negative</td>\n",
       "      <td>@VirginAmerica and it's a really big bad thing...</td>\n",
       "      <td>@virginamerica and it's a really big bad thing...</td>\n",
       "      <td>@virginamerica and it is a really big bad thin...</td>\n",
       "      <td>@virginamerica and it is a really big bad thin...</td>\n",
       "      <td>and it is a really big bad thing about it</td>\n",
       "      <td>really big bad thing</td>\n",
       "      <td>really big bad thing</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  airline_sentiment                                               text  \\\n",
       "0          positive                @VirginAmerica What @dhepburn said.   \n",
       "1          positive  @VirginAmerica plus you've added commercials t...   \n",
       "2          positive  @VirginAmerica I didn't today... Must mean I n...   \n",
       "3          negative  @VirginAmerica it's really aggressive to blast...   \n",
       "4          negative  @VirginAmerica and it's a really big bad thing...   \n",
       "\n",
       "                                             1_lower  \\\n",
       "0                @virginamerica what @dhepburn said.   \n",
       "1  @virginamerica plus you've added commercials t...   \n",
       "2  @virginamerica i didn't today... must mean i n...   \n",
       "3  @virginamerica it's really aggressive to blast...   \n",
       "4  @virginamerica and it's a really big bad thing...   \n",
       "\n",
       "                                       2_contraction  \\\n",
       "0                @virginamerica what @dhepburn said.   \n",
       "1  @virginamerica plus you have added commercials...   \n",
       "2  @virginamerica i did not today... must mean i ...   \n",
       "3  @virginamerica it is really aggressive to blas...   \n",
       "4  @virginamerica and it is a really big bad thin...   \n",
       "\n",
       "                                            3_number  \\\n",
       "0                @virginamerica what @dhepburn said.   \n",
       "1  @virginamerica plus you have added commercials...   \n",
       "2  @virginamerica i did not today... must mean i ...   \n",
       "3  @virginamerica it is really aggressive to blas...   \n",
       "4  @virginamerica and it is a really big bad thin...   \n",
       "\n",
       "                                             4_punct  \\\n",
       "0                                         what  said   \n",
       "1   plus you have added commercials to the experi...   \n",
       "2   i did not today must mean i need to take anot...   \n",
       "3   it is really aggressive to blast obnoxious en...   \n",
       "4          and it is a really big bad thing about it   \n",
       "\n",
       "                                          5_stopword  \\\n",
       "0                                               said   \n",
       "1            plus added commercials experience tacky   \n",
       "2             today must mean need take another trip   \n",
       "3  really aggressive blast obnoxious entertainmen...   \n",
       "4                               really big bad thing   \n",
       "\n",
       "                                             6_white  \n",
       "0                                               said  \n",
       "1            plus added commercials experience tacky  \n",
       "2             today must mean need take another trip  \n",
       "3  really aggressive blast obnoxious entertainmen...  \n",
       "4                               really big bad thing  "
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "def cleaning_all(urdf):\n",
    "    urdf['1_lower'] = urdf['text'].apply(to_lower)\n",
    "    urdf['2_contraction'] = urdf['1_lower'].apply(contractions.fix)\n",
    "    urdf['3_number'] = urdf['2_contraction'].apply(remove_number)\n",
    "    urdf['4_punct'] = urdf['3_number'].apply(remove_punctuation)\n",
    "    urdf['5_stopword'] = urdf['4_punct'].apply(remove_stopwords)\n",
    "    urdf['6_white'] = urdf['5_stopword'].apply(remove_whitespace)\n",
    "    return urdf\n",
    "df = cleaning_all(df)\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Preprocessing Part 2\n",
    "\n",
    "Mengubah kata menjadi kata dasar\n",
    "\n",
    "Pilihan:\n",
    "- Stemming\n",
    "- Lemmatization"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "from nltk.stem import SnowballStemmer, WordNetLemmatizer"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Stemming"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "stemmer = SnowballStemmer('english')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "def stem(text):\n",
    "    list_stemmed_word = []\n",
    "\n",
    "    for kalimat in nltk.sent_tokenize(text): # pecah paragraph menjadi kalimat\n",
    "        for kata in nltk.word_tokenize(kalimat): # pecah kalimat menjadi kata\n",
    "            stemmed_word = stemmer.stem(kata) # ubah kata ke kata dasarnya\n",
    "            list_stemmed_word.append(stemmed_word)\n",
    "\n",
    "    return ' '.join(list_stemmed_word)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Lemmatization"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "lemmatizer = WordNetLemmatizer()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [],
   "source": [
    "def lemmatize(text):\n",
    "    list_lemmatized_word = []\n",
    "\n",
    "    for kalimat in nltk.sent_tokenize(text): # pecah paragraph menjadi kalimat\n",
    "        for kata in nltk.word_tokenize(kalimat): # pecah kalimat menjadi kata\n",
    "            lemmatized_word = lemmatizer.lemmatize(kata) # ubah kata ke kata dasarnya\n",
    "            list_lemmatized_word.append(lemmatized_word)\n",
    "\n",
    "    return ' '.join(list_lemmatized_word)"
   ]
  },
  {
   "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>airline_sentiment</th>\n",
       "      <th>text</th>\n",
       "      <th>1_lower</th>\n",
       "      <th>2_contraction</th>\n",
       "      <th>3_number</th>\n",
       "      <th>4_punct</th>\n",
       "      <th>5_stopword</th>\n",
       "      <th>6_white</th>\n",
       "      <th>stem</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>positive</td>\n",
       "      <td>@VirginAmerica What @dhepburn said.</td>\n",
       "      <td>@virginamerica what @dhepburn said.</td>\n",
       "      <td>@virginamerica what @dhepburn said.</td>\n",
       "      <td>@virginamerica what @dhepburn said.</td>\n",
       "      <td>what  said</td>\n",
       "      <td>said</td>\n",
       "      <td>said</td>\n",
       "      <td>said</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>positive</td>\n",
       "      <td>@VirginAmerica plus you've added commercials t...</td>\n",
       "      <td>@virginamerica plus you've added commercials t...</td>\n",
       "      <td>@virginamerica plus you have added commercials...</td>\n",
       "      <td>@virginamerica plus you have added commercials...</td>\n",
       "      <td>plus you have added commercials to the experi...</td>\n",
       "      <td>plus added commercials experience tacky</td>\n",
       "      <td>plus added commercials experience tacky</td>\n",
       "      <td>plus ad commerci experi tacki</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  airline_sentiment                                               text  \\\n",
       "0          positive                @VirginAmerica What @dhepburn said.   \n",
       "1          positive  @VirginAmerica plus you've added commercials t...   \n",
       "\n",
       "                                             1_lower  \\\n",
       "0                @virginamerica what @dhepburn said.   \n",
       "1  @virginamerica plus you've added commercials t...   \n",
       "\n",
       "                                       2_contraction  \\\n",
       "0                @virginamerica what @dhepburn said.   \n",
       "1  @virginamerica plus you have added commercials...   \n",
       "\n",
       "                                            3_number  \\\n",
       "0                @virginamerica what @dhepburn said.   \n",
       "1  @virginamerica plus you have added commercials...   \n",
       "\n",
       "                                             4_punct  \\\n",
       "0                                         what  said   \n",
       "1   plus you have added commercials to the experi...   \n",
       "\n",
       "                                5_stopword  \\\n",
       "0                                     said   \n",
       "1  plus added commercials experience tacky   \n",
       "\n",
       "                                   6_white                           stem  \n",
       "0                                     said                           said  \n",
       "1  plus added commercials experience tacky  plus ad commerci experi tacki  "
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['stem'] = df['6_white'].apply(stem)\n",
    "df.head(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "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>airline_sentiment</th>\n",
       "      <th>text</th>\n",
       "      <th>1_lower</th>\n",
       "      <th>2_contraction</th>\n",
       "      <th>3_number</th>\n",
       "      <th>4_punct</th>\n",
       "      <th>5_stopword</th>\n",
       "      <th>6_white</th>\n",
       "      <th>stem</th>\n",
       "      <th>lemmatize</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>positive</td>\n",
       "      <td>@VirginAmerica What @dhepburn said.</td>\n",
       "      <td>@virginamerica what @dhepburn said.</td>\n",
       "      <td>@virginamerica what @dhepburn said.</td>\n",
       "      <td>@virginamerica what @dhepburn said.</td>\n",
       "      <td>what  said</td>\n",
       "      <td>said</td>\n",
       "      <td>said</td>\n",
       "      <td>said</td>\n",
       "      <td>said</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>positive</td>\n",
       "      <td>@VirginAmerica plus you've added commercials t...</td>\n",
       "      <td>@virginamerica plus you've added commercials t...</td>\n",
       "      <td>@virginamerica plus you have added commercials...</td>\n",
       "      <td>@virginamerica plus you have added commercials...</td>\n",
       "      <td>plus you have added commercials to the experi...</td>\n",
       "      <td>plus added commercials experience tacky</td>\n",
       "      <td>plus added commercials experience tacky</td>\n",
       "      <td>plus ad commerci experi tacki</td>\n",
       "      <td>plus added commercial experience tacky</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  airline_sentiment                                               text  \\\n",
       "0          positive                @VirginAmerica What @dhepburn said.   \n",
       "1          positive  @VirginAmerica plus you've added commercials t...   \n",
       "\n",
       "                                             1_lower  \\\n",
       "0                @virginamerica what @dhepburn said.   \n",
       "1  @virginamerica plus you've added commercials t...   \n",
       "\n",
       "                                       2_contraction  \\\n",
       "0                @virginamerica what @dhepburn said.   \n",
       "1  @virginamerica plus you have added commercials...   \n",
       "\n",
       "                                            3_number  \\\n",
       "0                @virginamerica what @dhepburn said.   \n",
       "1  @virginamerica plus you have added commercials...   \n",
       "\n",
       "                                             4_punct  \\\n",
       "0                                         what  said   \n",
       "1   plus you have added commercials to the experi...   \n",
       "\n",
       "                                5_stopword  \\\n",
       "0                                     said   \n",
       "1  plus added commercials experience tacky   \n",
       "\n",
       "                                   6_white                           stem  \\\n",
       "0                                     said                           said   \n",
       "1  plus added commercials experience tacky  plus ad commerci experi tacki   \n",
       "\n",
       "                                lemmatize  \n",
       "0                                    said  \n",
       "1  plus added commercial experience tacky  "
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['lemmatize'] = df['6_white'].apply(lemmatize)\n",
    "df.head(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(14640, 10)"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(14138, 10)"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.drop_duplicates(subset='lemmatize').shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [],
   "source": [
    "df.drop_duplicates(subset='lemmatize', inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [],
   "source": [
    "def word_freq(text):\n",
    "\n",
    "    df_text = text.str.extractall('(\\w+)') # ambil tiap kata dari tiap document\n",
    "\n",
    "    df_text = df_text.reset_index()[0].value_counts()\n",
    "\n",
    "    df_text = df_text.to_frame().reset_index()\n",
    "\n",
    "    df_text = df_text.rename(columns={0:'word', 'count':'freq'})\n",
    "    return df_text\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "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>word</th>\n",
       "      <th>freq</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>flight</td>\n",
       "      <td>4458</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>get</td>\n",
       "      <td>1360</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>hour</td>\n",
       "      <td>1134</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>thanks</td>\n",
       "      <td>1067</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>cancelled</td>\n",
       "      <td>1030</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12616</th>\n",
       "      <td>httptcopdntgbxc</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12617</th>\n",
       "      <td>deed</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12618</th>\n",
       "      <td>curse</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12619</th>\n",
       "      <td>notmadeofmoney</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12620</th>\n",
       "      <td>blackberry</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>12621 rows × 2 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                  word  freq\n",
       "0               flight  4458\n",
       "1                  get  1360\n",
       "2                 hour  1134\n",
       "3               thanks  1067\n",
       "4            cancelled  1030\n",
       "...                ...   ...\n",
       "12616  httptcopdntgbxc     1\n",
       "12617             deed     1\n",
       "12618            curse     1\n",
       "12619   notmadeofmoney     1\n",
       "12620       blackberry     1\n",
       "\n",
       "[12621 rows x 2 columns]"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "word_freq(df['lemmatize'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "airline_sentiment\n",
       "negative    9034\n",
       "positive    5104\n",
       "Name: count, dtype: int64"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['airline_sentiment'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Top 10 most frequent words\n",
    "\n",
    "spam_data = df[df['airline_sentiment'] == 'positive']['lemmatize']\n",
    "plt.title(\"Frequency of positive words\")\n",
    "\n",
    "sns.barplot(\n",
    "    data=word_freq(spam_data).loc[:10],\n",
    "    x='freq',\n",
    "    y='word',\n",
    "    hue='word'\n",
    ")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Axes: title={'center': 'Frequency of negative words'}, xlabel='freq', ylabel='word'>"
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Top 10 most frequent words\n",
    "\n",
    "not_spam_data = df[df['airline_sentiment'] == 'negative']['lemmatize']\n",
    "plt.title(\"Frequency of negative words\")\n",
    "\n",
    "sns.barplot(\n",
    "    data=word_freq(not_spam_data).loc[:10],\n",
    "    x='freq',\n",
    "    y='word',\n",
    "    hue='word'\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Insight\n",
    "\n",
    "- Pada tweet positive paling sering muncul kata: flight, thanks, thank\n",
    "- pada tweet negative paling sering muncul kata: flight, hour, get\n",
    "- kata 'flight','u', 'get' jika mau dianggap sebagai stopwords, maka bisa kita hapus"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 2. Length of Document"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [],
   "source": [
    "df['len_char'] = df['lemmatize'].apply(len)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "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>airline_sentiment</th>\n",
       "      <th>text</th>\n",
       "      <th>1_lower</th>\n",
       "      <th>2_contraction</th>\n",
       "      <th>3_number</th>\n",
       "      <th>4_punct</th>\n",
       "      <th>5_stopword</th>\n",
       "      <th>6_white</th>\n",
       "      <th>stem</th>\n",
       "      <th>lemmatize</th>\n",
       "      <th>len_char</th>\n",
       "      <th>len_word</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>positive</td>\n",
       "      <td>@VirginAmerica What @dhepburn said.</td>\n",
       "      <td>@virginamerica what @dhepburn said.</td>\n",
       "      <td>@virginamerica what @dhepburn said.</td>\n",
       "      <td>@virginamerica what @dhepburn said.</td>\n",
       "      <td>what  said</td>\n",
       "      <td>said</td>\n",
       "      <td>said</td>\n",
       "      <td>said</td>\n",
       "      <td>said</td>\n",
       "      <td>4</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>positive</td>\n",
       "      <td>@VirginAmerica plus you've added commercials t...</td>\n",
       "      <td>@virginamerica plus you've added commercials t...</td>\n",
       "      <td>@virginamerica plus you have added commercials...</td>\n",
       "      <td>@virginamerica plus you have added commercials...</td>\n",
       "      <td>plus you have added commercials to the experi...</td>\n",
       "      <td>plus added commercials experience tacky</td>\n",
       "      <td>plus added commercials experience tacky</td>\n",
       "      <td>plus ad commerci experi tacki</td>\n",
       "      <td>plus added commercial experience tacky</td>\n",
       "      <td>38</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>positive</td>\n",
       "      <td>@VirginAmerica I didn't today... Must mean I n...</td>\n",
       "      <td>@virginamerica i didn't today... must mean i n...</td>\n",
       "      <td>@virginamerica i did not today... must mean i ...</td>\n",
       "      <td>@virginamerica i did not today... must mean i ...</td>\n",
       "      <td>i did not today must mean i need to take anot...</td>\n",
       "      <td>today must mean need take another trip</td>\n",
       "      <td>today must mean need take another trip</td>\n",
       "      <td>today must mean need take anoth trip</td>\n",
       "      <td>today must mean need take another trip</td>\n",
       "      <td>38</td>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  airline_sentiment                                               text  \\\n",
       "0          positive                @VirginAmerica What @dhepburn said.   \n",
       "1          positive  @VirginAmerica plus you've added commercials t...   \n",
       "2          positive  @VirginAmerica I didn't today... Must mean I n...   \n",
       "\n",
       "                                             1_lower  \\\n",
       "0                @virginamerica what @dhepburn said.   \n",
       "1  @virginamerica plus you've added commercials t...   \n",
       "2  @virginamerica i didn't today... must mean i n...   \n",
       "\n",
       "                                       2_contraction  \\\n",
       "0                @virginamerica what @dhepburn said.   \n",
       "1  @virginamerica plus you have added commercials...   \n",
       "2  @virginamerica i did not today... must mean i ...   \n",
       "\n",
       "                                            3_number  \\\n",
       "0                @virginamerica what @dhepburn said.   \n",
       "1  @virginamerica plus you have added commercials...   \n",
       "2  @virginamerica i did not today... must mean i ...   \n",
       "\n",
       "                                             4_punct  \\\n",
       "0                                         what  said   \n",
       "1   plus you have added commercials to the experi...   \n",
       "2   i did not today must mean i need to take anot...   \n",
       "\n",
       "                                5_stopword  \\\n",
       "0                                     said   \n",
       "1  plus added commercials experience tacky   \n",
       "2   today must mean need take another trip   \n",
       "\n",
       "                                   6_white  \\\n",
       "0                                     said   \n",
       "1  plus added commercials experience tacky   \n",
       "2   today must mean need take another trip   \n",
       "\n",
       "                                   stem  \\\n",
       "0                                  said   \n",
       "1         plus ad commerci experi tacki   \n",
       "2  today must mean need take anoth trip   \n",
       "\n",
       "                                lemmatize  len_char  len_word  \n",
       "0                                    said         4         1  \n",
       "1  plus added commercial experience tacky        38         5  \n",
       "2  today must mean need take another trip        38         7  "
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['len_word'] = df['lemmatize'].apply(lambda x: len(x.split()))\n",
    "\n",
    "df.head(3)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# EDA"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.title('Distribution of Number of Character in Positive vs Negative Tweet')\n",
    "sns.boxenplot(data=df, x='len_char', y='airline_sentiment');"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.title('Distribution of Number of Character in Positive vs Negative Tweet')\n",
    "sns.histplot(data=df, x='len_char', hue='airline_sentiment');"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Insight:\n",
    "- Tweet Negatif cenderung lebih banyak jumlah karakternya dibanding Tweet Positif"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Tujuan dari klasifikasi tweet ini adalah untuk memberikan alarm/warning ketika muncul sebuah tweet negatif sehingga tim PR (Public Relation) bisa menghandle tweet negatif tsb dengan cepat."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Tujuan selalu berada di positif**, karena tujuannya itu tweet negatif maka :\n",
    "\n",
    "- class 0 (1) : itu merupakan tweet positif\n",
    "\n",
    "- class 1 (+) : itu merupakan tweet negatif"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "berdasarkan tujuan dari klasifikasi ini akan lebih baik jika menggunakan recall karena ingin mencegah terjadinya error pada FN"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "FP: Diprediksi tweet negatif -> alarm nyala -> actualnya tweet positive -> cost : waktu untuk cek tweet\n",
    "\n",
    "FN: Diprediksi tweet positif -> alarm tidak menyala -> actualnya tweet negatif -> cost : tweet negatif tidak di handle (merusak reputasi)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "---\n",
    "# Bag of Words (Document Term Matrix)\n",
    "\n",
    "- TF (Term Frequency)\n",
    "- TF-IDF (Term Frequency - Inverse Document Frequency)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.feature_extraction.text import CountVectorizer, TfidfVectorizer"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Vectorizer"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 1. TF (Term Frequency)\n",
    "\n",
    "CountVectorizer\n",
    "\n",
    "Berapa kali sebuah kata muncul dalam sebuah dokumen (tweet/email/dll)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [],
   "source": [
    "# define vectorizer\n",
    "vect_tf = CountVectorizer(ngram_range=(1,1))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 2. TF-IDF\n",
    "\n",
    "- TF -> Semakin sering sebuah kata muncul dalam 1 dokumen, maka skornya semakin besar\n",
    "- IDF -> semakin sering sebuah kata muncul dalam banyak dokumen (katanya semakin pasar), skornya semakin kecil"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [],
   "source": [
    "# define vectorizer\n",
    "vect_tfidf = TfidfVectorizer(norm=None, ngram_range=(1,1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "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>lemmatize</th>\n",
       "      <th>airline_sentiment</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>said</td>\n",
       "      <td>positive</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>plus added commercial experience tacky</td>\n",
       "      <td>positive</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>today must mean need take another trip</td>\n",
       "      <td>positive</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>really aggressive blast obnoxious entertainmen...</td>\n",
       "      <td>negative</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>really big bad thing</td>\n",
       "      <td>negative</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                           lemmatize airline_sentiment\n",
       "0                                               said          positive\n",
       "1             plus added commercial experience tacky          positive\n",
       "2             today must mean need take another trip          positive\n",
       "3  really aggressive blast obnoxious entertainmen...          negative\n",
       "4                               really big bad thing          negative"
      ]
     },
     "execution_count": 48,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Load Dataset\n",
    "df_clean = df[['lemmatize', 'airline_sentiment']].copy()\n",
    "df_clean.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "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>lemmatize</th>\n",
       "      <th>airline_sentiment</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>said</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>plus added commercial experience tacky</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>today must mean need take another trip</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>really aggressive blast obnoxious entertainmen...</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>really big bad thing</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                           lemmatize  airline_sentiment\n",
       "0                                               said                  0\n",
       "1             plus added commercial experience tacky                  0\n",
       "2             today must mean need take another trip                  0\n",
       "3  really aggressive blast obnoxious entertainmen...                  1\n",
       "4                               really big bad thing                  1"
      ]
     },
     "execution_count": 49,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_clean['airline_sentiment'] = np.where(df_clean['airline_sentiment'] == 'negative', 1, 0)\n",
    "\n",
    "df_clean.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [],
   "source": [
    "X = df_clean['lemmatize']\n",
    "y = df_clean['airline_sentiment']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "airline_sentiment\n",
       "1    63.898713\n",
       "0    36.101287\n",
       "Name: count, dtype: float64"
      ]
     },
     "execution_count": 51,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y.value_counts()/len(y) * 100 # imbalance"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Train Test Split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Splitting data training dan test dengan proporsi 70:30\n",
    "X_train, X_test, y_train, y_test= train_test_split(\n",
    "    X,\n",
    "    y,  \n",
    "    stratify=y,\n",
    "    test_size=0.2,\n",
    "    random_state=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [],
   "source": [
    "# algorithm\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.neighbors import KNeighborsClassifier\n",
    "from sklearn.tree import DecisionTreeClassifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [],
   "source": [
    "from imblearn.pipeline import Pipeline\n",
    "from imblearn.under_sampling import RandomUnderSampler, NearMiss\n",
    "from imblearn.over_sampling import RandomOverSampler, SMOTE\n",
    "from sklearn.tree import DecisionTreeClassifier\n",
    "from sklearn.model_selection import cross_val_score, RandomizedSearchCV, TimeSeriesSplit"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.ensemble import VotingClassifier, StackingClassifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [],
   "source": [
    "# algo\n",
    "logreg = LogisticRegression(solver=\"liblinear\", random_state=0)\n",
    "knn = KNeighborsClassifier(n_neighbors=5)\n",
    "tree = DecisionTreeClassifier(max_depth=5, random_state=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Voting Classifier\n",
    "voting = VotingClassifier(estimators=[\n",
    "    ('clf1', logreg),\n",
    "    ('clf2', knn),\n",
    "    ('clf3', tree)\n",
    "])\n",
    "\n",
    "# Stacking Classifier\n",
    "stacking = StackingClassifier(estimators=[\n",
    "    ('clf1', logreg),\n",
    "    ('clf2', knn),\n",
    "    ('clf3', tree)\n",
    "], \n",
    "final_estimator=logreg # digunakan untuk meta learner\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [],
   "source": [
    "# cross validation memilih model terbaik berdasarkan mae\n",
    "models = [logreg, knn, tree, voting, stacking]\n",
    "\n",
    "list_all = []\n",
    "list_mean = []\n",
    "list_std = []\n",
    "\n",
    "# mencari model terbaik berdasarkan MAE terkecil\n",
    "for i in models:\n",
    "\n",
    "    # Pipeline berisi 3 tahap, untuk diisi di pada estimator di cross val\n",
    "    pipe_model = Pipeline([\n",
    "                        ('preprocess', vect_tf),\n",
    "                        ('model', i)\n",
    "                    ])\n",
    "\n",
    "    # cross val\n",
    "    # cross_validate untuk menghitung skor untuk beberapa metrik\n",
    "    cv_score = cross_val_score(\n",
    "        estimator = pipe_model, \n",
    "        X = X_train, # menggunakan X_train awal (karena transform dan scaling dilakukan di Pipeline)\n",
    "        y = y_train, \n",
    "        cv = 5,\n",
    "        scoring = 'recall',\n",
    "        error_score = \"raise\",\n",
    "        n_jobs = -1\n",
    "    )\n",
    "    \n",
    "    list_all.append(cv_score.round(4))\n",
    "    list_mean.append(cv_score.mean())\n",
    "    list_std.append(cv_score.std())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "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>algo</th>\n",
       "      <th>all</th>\n",
       "      <th>mean</th>\n",
       "      <th>std</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>LogisticRegression(random_state=0, solver='lib...</td>\n",
       "      <td>[0.8491, 0.8533, 0.8685, 0.888, 0.8534]</td>\n",
       "      <td>0.862458</td>\n",
       "      <td>0.014361</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>KNeighborsClassifier()</td>\n",
       "      <td>[0.4547, 0.4457, 0.3986, 0.4246, 0.3658]</td>\n",
       "      <td>0.417884</td>\n",
       "      <td>0.032407</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>DecisionTreeClassifier(max_depth=5, random_sta...</td>\n",
       "      <td>[0.9522, 0.964, 0.9654, 0.971, 0.9599]</td>\n",
       "      <td>0.962501</td>\n",
       "      <td>0.006230</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>VotingClassifier(estimators=[('clf1',\\n       ...</td>\n",
       "      <td>[0.8533, 0.8491, 0.8616, 0.8817, 0.8485]</td>\n",
       "      <td>0.858861</td>\n",
       "      <td>0.012354</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>StackingClassifier(estimators=[('clf1',\\n     ...</td>\n",
       "      <td>[0.8588, 0.8554, 0.8754, 0.8949, 0.8686]</td>\n",
       "      <td>0.870621</td>\n",
       "      <td>0.014056</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                algo  \\\n",
       "0  LogisticRegression(random_state=0, solver='lib...   \n",
       "1                             KNeighborsClassifier()   \n",
       "2  DecisionTreeClassifier(max_depth=5, random_sta...   \n",
       "3  VotingClassifier(estimators=[('clf1',\\n       ...   \n",
       "4  StackingClassifier(estimators=[('clf1',\\n     ...   \n",
       "\n",
       "                                        all      mean       std  \n",
       "0   [0.8491, 0.8533, 0.8685, 0.888, 0.8534]  0.862458  0.014361  \n",
       "1  [0.4547, 0.4457, 0.3986, 0.4246, 0.3658]  0.417884  0.032407  \n",
       "2    [0.9522, 0.964, 0.9654, 0.971, 0.9599]  0.962501  0.006230  \n",
       "3  [0.8533, 0.8491, 0.8616, 0.8817, 0.8485]  0.858861  0.012354  \n",
       "4  [0.8588, 0.8554, 0.8754, 0.8949, 0.8686]  0.870621  0.014056  "
      ]
     },
     "execution_count": 60,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_cv = pd.DataFrame()\n",
    "df_cv['algo'] = models\n",
    "df_cv['all']= list_all\n",
    "df_cv['mean'] = list_mean\n",
    "df_cv['std'] = list_std\n",
    "\n",
    "df_cv"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Hyperparameter Tuning"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "airline_sentiment\n",
       "1    1807\n",
       "0    1021\n",
       "Name: count, dtype: int64"
      ]
     },
     "execution_count": 61,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_test.value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [],
   "source": [
    "# oversampling\n",
    "smote = SMOTE(random_state=0, sampling_strategy= \"minority\" )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [],
   "source": [
    "# undersampling\n",
    "nearmiss = NearMiss(sampling_strategy=\"majority\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [],
   "source": [
    "vect_tf2 = CountVectorizer(ngram_range=(2,2))\n",
    "vect_tfidf2 = TfidfVectorizer(ngram_range=(2,2), norm=None)\n",
    "# param_grid\n",
    "hyperparam = {\n",
    "    \"modeling__max_depth\": range(2, 31, 1),\n",
    "    \"modeling__criterion\": ['gini', 'entropy', 'log_loss'],\n",
    "    \"modeling__min_samples_split\": range(2, 50, 1),\n",
    "    \"modeling__min_samples_leaf\": range(2, 50, 1),\n",
    "    \"resampling\": [smote, nearmiss],\n",
    "    # \"vect__ngram_range\" : [(1,1),(2,2)],\n",
    "    \"vect\": [vect_tf, vect_tf2, vect_tfidf, vect_tfidf2]\n",
    "}\n",
    "\n",
    "# pipeline\n",
    "pipe_model = Pipeline([\n",
    "    (\"vect\", None),\n",
    "    ('resampling', None),\n",
    "    (\"modeling\", tree)\n",
    "])\n",
    "\n",
    "# tuning -> gridsearchcv\n",
    "randomsearch = RandomizedSearchCV(\n",
    "    estimator=pipe_model,\n",
    "    param_distributions =hyperparam,\n",
    "    cv=5,\n",
    "    scoring=\"recall\",\n",
    "    n_jobs=-1,\n",
    "    n_iter=100,\n",
    "    random_state=0\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Library/Frameworks/Python.framework/Versions/3.11/lib/python3.11/site-packages/joblib/externals/loky/process_executor.py:752: UserWarning: A worker stopped while some jobs were given to the executor. This can be caused by a too short worker timeout or by a memory leak.\n",
      "  warnings.warn(\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<style>#sk-container-id-1 {color: black;}#sk-container-id-1 pre{padding: 0;}#sk-container-id-1 div.sk-toggleable {background-color: white;}#sk-container-id-1 label.sk-toggleable__label {cursor: pointer;display: block;width: 100%;margin-bottom: 0;padding: 0.3em;box-sizing: border-box;text-align: center;}#sk-container-id-1 label.sk-toggleable__label-arrow:before {content: \"▸\";float: left;margin-right: 0.25em;color: #696969;}#sk-container-id-1 label.sk-toggleable__label-arrow:hover:before {color: black;}#sk-container-id-1 div.sk-estimator:hover label.sk-toggleable__label-arrow:before {color: black;}#sk-container-id-1 div.sk-toggleable__content {max-height: 0;max-width: 0;overflow: hidden;text-align: left;background-color: #f0f8ff;}#sk-container-id-1 div.sk-toggleable__content pre {margin: 0.2em;color: black;border-radius: 0.25em;background-color: #f0f8ff;}#sk-container-id-1 input.sk-toggleable__control:checked~div.sk-toggleable__content {max-height: 200px;max-width: 100%;overflow: auto;}#sk-container-id-1 input.sk-toggleable__control:checked~label.sk-toggleable__label-arrow:before {content: \"▾\";}#sk-container-id-1 div.sk-estimator input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-1 div.sk-label input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-1 input.sk-hidden--visually {border: 0;clip: rect(1px 1px 1px 1px);clip: rect(1px, 1px, 1px, 1px);height: 1px;margin: -1px;overflow: hidden;padding: 0;position: absolute;width: 1px;}#sk-container-id-1 div.sk-estimator {font-family: monospace;background-color: #f0f8ff;border: 1px dotted black;border-radius: 0.25em;box-sizing: border-box;margin-bottom: 0.5em;}#sk-container-id-1 div.sk-estimator:hover {background-color: #d4ebff;}#sk-container-id-1 div.sk-parallel-item::after {content: \"\";width: 100%;border-bottom: 1px solid gray;flex-grow: 1;}#sk-container-id-1 div.sk-label:hover label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-1 div.sk-serial::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 0;bottom: 0;left: 50%;z-index: 0;}#sk-container-id-1 div.sk-serial {display: flex;flex-direction: column;align-items: center;background-color: white;padding-right: 0.2em;padding-left: 0.2em;position: relative;}#sk-container-id-1 div.sk-item {position: relative;z-index: 1;}#sk-container-id-1 div.sk-parallel {display: flex;align-items: stretch;justify-content: center;background-color: white;position: relative;}#sk-container-id-1 div.sk-item::before, #sk-container-id-1 div.sk-parallel-item::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 0;bottom: 0;left: 50%;z-index: -1;}#sk-container-id-1 div.sk-parallel-item {display: flex;flex-direction: column;z-index: 1;position: relative;background-color: white;}#sk-container-id-1 div.sk-parallel-item:first-child::after {align-self: flex-end;width: 50%;}#sk-container-id-1 div.sk-parallel-item:last-child::after {align-self: flex-start;width: 50%;}#sk-container-id-1 div.sk-parallel-item:only-child::after {width: 0;}#sk-container-id-1 div.sk-dashed-wrapped {border: 1px dashed gray;margin: 0 0.4em 0.5em 0.4em;box-sizing: border-box;padding-bottom: 0.4em;background-color: white;}#sk-container-id-1 div.sk-label label {font-family: monospace;font-weight: bold;display: inline-block;line-height: 1.2em;}#sk-container-id-1 div.sk-label-container {text-align: center;}#sk-container-id-1 div.sk-container {/* jupyter's `normalize.less` sets `[hidden] { display: none; }` but bootstrap.min.css set `[hidden] { display: none !important; }` so we also need the `!important` here to be able to override the default hidden behavior on the sphinx rendered scikit-learn.org. See: https://github.com/scikit-learn/scikit-learn/issues/21755 */display: inline-block !important;position: relative;}#sk-container-id-1 div.sk-text-repr-fallback {display: none;}</style><div id=\"sk-container-id-1\" class=\"sk-top-container\"><div class=\"sk-text-repr-fallback\"><pre>RandomizedSearchCV(cv=5,\n",
       "                   estimator=Pipeline(steps=[(&#x27;vect&#x27;, None),\n",
       "                                             (&#x27;resampling&#x27;, None),\n",
       "                                             (&#x27;modeling&#x27;,\n",
       "                                              DecisionTreeClassifier(max_depth=5,\n",
       "                                                                     random_state=0))]),\n",
       "                   n_iter=100, n_jobs=-1,\n",
       "                   param_distributions={&#x27;modeling__criterion&#x27;: [&#x27;gini&#x27;,\n",
       "                                                                &#x27;entropy&#x27;,\n",
       "                                                                &#x27;log_loss&#x27;],\n",
       "                                        &#x27;modeling__max_depth&#x27;: range(2, 31),\n",
       "                                        &#x27;modeling__min_samples_leaf&#x27;: range(2, 50),\n",
       "                                        &#x27;modeling__min_samples_split&#x27;: range(2, 50),\n",
       "                                        &#x27;resampling&#x27;: [SMOTE(random_state=0,\n",
       "                                                             sampling_strategy=&#x27;minority&#x27;),\n",
       "                                                       NearMiss(sampling_strategy=&#x27;majority&#x27;)],\n",
       "                                        &#x27;vect&#x27;: [CountVectorizer(),\n",
       "                                                 CountVectorizer(ngram_range=(2,\n",
       "                                                                              2)),\n",
       "                                                 TfidfVectorizer(norm=None),\n",
       "                                                 TfidfVectorizer(ngram_range=(2,\n",
       "                                                                              2),\n",
       "                                                                 norm=None)]},\n",
       "                   random_state=0, scoring=&#x27;recall&#x27;)</pre><b>In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook. <br />On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.</b></div><div class=\"sk-container\" hidden><div class=\"sk-item sk-dashed-wrapped\"><div class=\"sk-label-container\"><div class=\"sk-label sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-1\" type=\"checkbox\" ><label for=\"sk-estimator-id-1\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">RandomizedSearchCV</label><div class=\"sk-toggleable__content\"><pre>RandomizedSearchCV(cv=5,\n",
       "                   estimator=Pipeline(steps=[(&#x27;vect&#x27;, None),\n",
       "                                             (&#x27;resampling&#x27;, None),\n",
       "                                             (&#x27;modeling&#x27;,\n",
       "                                              DecisionTreeClassifier(max_depth=5,\n",
       "                                                                     random_state=0))]),\n",
       "                   n_iter=100, n_jobs=-1,\n",
       "                   param_distributions={&#x27;modeling__criterion&#x27;: [&#x27;gini&#x27;,\n",
       "                                                                &#x27;entropy&#x27;,\n",
       "                                                                &#x27;log_loss&#x27;],\n",
       "                                        &#x27;modeling__max_depth&#x27;: range(2, 31),\n",
       "                                        &#x27;modeling__min_samples_leaf&#x27;: range(2, 50),\n",
       "                                        &#x27;modeling__min_samples_split&#x27;: range(2, 50),\n",
       "                                        &#x27;resampling&#x27;: [SMOTE(random_state=0,\n",
       "                                                             sampling_strategy=&#x27;minority&#x27;),\n",
       "                                                       NearMiss(sampling_strategy=&#x27;majority&#x27;)],\n",
       "                                        &#x27;vect&#x27;: [CountVectorizer(),\n",
       "                                                 CountVectorizer(ngram_range=(2,\n",
       "                                                                              2)),\n",
       "                                                 TfidfVectorizer(norm=None),\n",
       "                                                 TfidfVectorizer(ngram_range=(2,\n",
       "                                                                              2),\n",
       "                                                                 norm=None)]},\n",
       "                   random_state=0, scoring=&#x27;recall&#x27;)</pre></div></div></div><div class=\"sk-parallel\"><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-2\" type=\"checkbox\" ><label for=\"sk-estimator-id-2\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">estimator: Pipeline</label><div class=\"sk-toggleable__content\"><pre>Pipeline(steps=[(&#x27;vect&#x27;, None), (&#x27;resampling&#x27;, None),\n",
       "                (&#x27;modeling&#x27;,\n",
       "                 DecisionTreeClassifier(max_depth=5, random_state=0))])</pre></div></div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-3\" type=\"checkbox\" ><label for=\"sk-estimator-id-3\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">None</label><div class=\"sk-toggleable__content\"><pre>None</pre></div></div></div><div class=\"sk-item\"><div class=\"sk-estimator sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-4\" type=\"checkbox\" ><label for=\"sk-estimator-id-4\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">None</label><div class=\"sk-toggleable__content\"><pre>None</pre></div></div></div><div class=\"sk-item\"><div class=\"sk-estimator sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-5\" type=\"checkbox\" ><label for=\"sk-estimator-id-5\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">DecisionTreeClassifier</label><div class=\"sk-toggleable__content\"><pre>DecisionTreeClassifier(max_depth=5, random_state=0)</pre></div></div></div></div></div></div></div></div></div></div></div></div>"
      ],
      "text/plain": [
       "RandomizedSearchCV(cv=5,\n",
       "                   estimator=Pipeline(steps=[('vect', None),\n",
       "                                             ('resampling', None),\n",
       "                                             ('modeling',\n",
       "                                              DecisionTreeClassifier(max_depth=5,\n",
       "                                                                     random_state=0))]),\n",
       "                   n_iter=100, n_jobs=-1,\n",
       "                   param_distributions={'modeling__criterion': ['gini',\n",
       "                                                                'entropy',\n",
       "                                                                'log_loss'],\n",
       "                                        'modeling__max_depth': range(2, 31),\n",
       "                                        'modeling__min_samples_leaf': range(2, 50),\n",
       "                                        'modeling__min_samples_split': range(2, 50),\n",
       "                                        'resampling': [SMOTE(random_state=0,\n",
       "                                                             sampling_strategy='minority'),\n",
       "                                                       NearMiss(sampling_strategy='majority')],\n",
       "                                        'vect': [CountVectorizer(),\n",
       "                                                 CountVectorizer(ngram_range=(2,\n",
       "                                                                              2)),\n",
       "                                                 TfidfVectorizer(norm=None),\n",
       "                                                 TfidfVectorizer(ngram_range=(2,\n",
       "                                                                              2),\n",
       "                                                                 norm=None)]},\n",
       "                   random_state=0, scoring='recall')"
      ]
     },
     "execution_count": 65,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "randomsearch.fit(X_train, y_train)"
   ]
  },
  {
   "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>mean_fit_time</th>\n",
       "      <th>std_fit_time</th>\n",
       "      <th>mean_score_time</th>\n",
       "      <th>std_score_time</th>\n",
       "      <th>param_vect</th>\n",
       "      <th>param_resampling</th>\n",
       "      <th>param_modeling__min_samples_split</th>\n",
       "      <th>param_modeling__min_samples_leaf</th>\n",
       "      <th>param_modeling__max_depth</th>\n",
       "      <th>param_modeling__criterion</th>\n",
       "      <th>params</th>\n",
       "      <th>split0_test_score</th>\n",
       "      <th>split1_test_score</th>\n",
       "      <th>split2_test_score</th>\n",
       "      <th>split3_test_score</th>\n",
       "      <th>split4_test_score</th>\n",
       "      <th>mean_test_score</th>\n",
       "      <th>std_test_score</th>\n",
       "      <th>rank_test_score</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>80</th>\n",
       "      <td>0.750471</td>\n",
       "      <td>0.102376</td>\n",
       "      <td>0.013160</td>\n",
       "      <td>0.003598</td>\n",
       "      <td>CountVectorizer()</td>\n",
       "      <td>NearMiss(sampling_strategy='majority')</td>\n",
       "      <td>28</td>\n",
       "      <td>7</td>\n",
       "      <td>2</td>\n",
       "      <td>entropy</td>\n",
       "      <td>{'vect': CountVectorizer(), 'resampling': Near...</td>\n",
       "      <td>0.959862</td>\n",
       "      <td>0.971626</td>\n",
       "      <td>0.971626</td>\n",
       "      <td>0.967497</td>\n",
       "      <td>0.966805</td>\n",
       "      <td>0.967483</td>\n",
       "      <td>0.004310</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>0.573973</td>\n",
       "      <td>0.008334</td>\n",
       "      <td>0.015691</td>\n",
       "      <td>0.005228</td>\n",
       "      <td>TfidfVectorizer(norm=None)</td>\n",
       "      <td>SMOTE(random_state=0, sampling_strategy='minor...</td>\n",
       "      <td>9</td>\n",
       "      <td>40</td>\n",
       "      <td>5</td>\n",
       "      <td>entropy</td>\n",
       "      <td>{'vect': TfidfVectorizer(norm=None), 'resampli...</td>\n",
       "      <td>0.951557</td>\n",
       "      <td>0.970242</td>\n",
       "      <td>0.965398</td>\n",
       "      <td>0.963347</td>\n",
       "      <td>0.960581</td>\n",
       "      <td>0.962225</td>\n",
       "      <td>0.006199</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>59</th>\n",
       "      <td>0.939578</td>\n",
       "      <td>0.095348</td>\n",
       "      <td>0.011759</td>\n",
       "      <td>0.001518</td>\n",
       "      <td>CountVectorizer()</td>\n",
       "      <td>NearMiss(sampling_strategy='majority')</td>\n",
       "      <td>2</td>\n",
       "      <td>19</td>\n",
       "      <td>3</td>\n",
       "      <td>log_loss</td>\n",
       "      <td>{'vect': CountVectorizer(), 'resampling': Near...</td>\n",
       "      <td>0.952249</td>\n",
       "      <td>0.966090</td>\n",
       "      <td>0.962630</td>\n",
       "      <td>0.963347</td>\n",
       "      <td>0.955740</td>\n",
       "      <td>0.960011</td>\n",
       "      <td>0.005166</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>62</th>\n",
       "      <td>0.536093</td>\n",
       "      <td>0.024337</td>\n",
       "      <td>0.014927</td>\n",
       "      <td>0.003321</td>\n",
       "      <td>TfidfVectorizer(norm=None)</td>\n",
       "      <td>SMOTE(random_state=0, sampling_strategy='minor...</td>\n",
       "      <td>25</td>\n",
       "      <td>48</td>\n",
       "      <td>3</td>\n",
       "      <td>entropy</td>\n",
       "      <td>{'vect': TfidfVectorizer(norm=None), 'resampli...</td>\n",
       "      <td>0.947405</td>\n",
       "      <td>0.960554</td>\n",
       "      <td>0.957093</td>\n",
       "      <td>0.963347</td>\n",
       "      <td>0.955740</td>\n",
       "      <td>0.956828</td>\n",
       "      <td>0.005412</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>87</th>\n",
       "      <td>0.344816</td>\n",
       "      <td>0.023308</td>\n",
       "      <td>0.012829</td>\n",
       "      <td>0.002858</td>\n",
       "      <td>CountVectorizer()</td>\n",
       "      <td>SMOTE(random_state=0, sampling_strategy='minor...</td>\n",
       "      <td>47</td>\n",
       "      <td>17</td>\n",
       "      <td>28</td>\n",
       "      <td>log_loss</td>\n",
       "      <td>{'vect': CountVectorizer(), 'resampling': SMOT...</td>\n",
       "      <td>0.677509</td>\n",
       "      <td>0.686505</td>\n",
       "      <td>0.703114</td>\n",
       "      <td>0.705394</td>\n",
       "      <td>0.654910</td>\n",
       "      <td>0.685486</td>\n",
       "      <td>0.018475</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>35</th>\n",
       "      <td>0.356033</td>\n",
       "      <td>0.032776</td>\n",
       "      <td>0.025463</td>\n",
       "      <td>0.007204</td>\n",
       "      <td>CountVectorizer(ngram_range=(2, 2))</td>\n",
       "      <td>SMOTE(random_state=0, sampling_strategy='minor...</td>\n",
       "      <td>24</td>\n",
       "      <td>10</td>\n",
       "      <td>3</td>\n",
       "      <td>entropy</td>\n",
       "      <td>{'vect': CountVectorizer(ngram_range=(2, 2)), ...</td>\n",
       "      <td>0.142561</td>\n",
       "      <td>0.132872</td>\n",
       "      <td>0.153633</td>\n",
       "      <td>0.150069</td>\n",
       "      <td>0.134163</td>\n",
       "      <td>0.142660</td>\n",
       "      <td>0.008286</td>\n",
       "      <td>96</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>65</th>\n",
       "      <td>0.260275</td>\n",
       "      <td>0.022035</td>\n",
       "      <td>0.020876</td>\n",
       "      <td>0.004961</td>\n",
       "      <td>CountVectorizer(ngram_range=(2, 2))</td>\n",
       "      <td>SMOTE(random_state=0, sampling_strategy='minor...</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>3</td>\n",
       "      <td>gini</td>\n",
       "      <td>{'vect': CountVectorizer(ngram_range=(2, 2)), ...</td>\n",
       "      <td>0.142561</td>\n",
       "      <td>0.132872</td>\n",
       "      <td>0.153633</td>\n",
       "      <td>0.150069</td>\n",
       "      <td>0.134163</td>\n",
       "      <td>0.142660</td>\n",
       "      <td>0.008286</td>\n",
       "      <td>96</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>0.276917</td>\n",
       "      <td>0.015044</td>\n",
       "      <td>0.020389</td>\n",
       "      <td>0.003677</td>\n",
       "      <td>CountVectorizer(ngram_range=(2, 2))</td>\n",
       "      <td>SMOTE(random_state=0, sampling_strategy='minor...</td>\n",
       "      <td>42</td>\n",
       "      <td>19</td>\n",
       "      <td>2</td>\n",
       "      <td>entropy</td>\n",
       "      <td>{'vect': CountVectorizer(ngram_range=(2, 2)), ...</td>\n",
       "      <td>0.103114</td>\n",
       "      <td>0.091349</td>\n",
       "      <td>0.099654</td>\n",
       "      <td>0.101660</td>\n",
       "      <td>0.097510</td>\n",
       "      <td>0.098658</td>\n",
       "      <td>0.004113</td>\n",
       "      <td>98</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>54</th>\n",
       "      <td>1.077123</td>\n",
       "      <td>0.022111</td>\n",
       "      <td>0.025241</td>\n",
       "      <td>0.003827</td>\n",
       "      <td>CountVectorizer(ngram_range=(2, 2))</td>\n",
       "      <td>NearMiss(sampling_strategy='majority')</td>\n",
       "      <td>2</td>\n",
       "      <td>39</td>\n",
       "      <td>4</td>\n",
       "      <td>log_loss</td>\n",
       "      <td>{'vect': CountVectorizer(ngram_range=(2, 2)), ...</td>\n",
       "      <td>0.108651</td>\n",
       "      <td>0.062284</td>\n",
       "      <td>0.072664</td>\n",
       "      <td>0.074689</td>\n",
       "      <td>0.079530</td>\n",
       "      <td>0.079563</td>\n",
       "      <td>0.015595</td>\n",
       "      <td>99</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>89</th>\n",
       "      <td>1.444719</td>\n",
       "      <td>0.056866</td>\n",
       "      <td>0.026205</td>\n",
       "      <td>0.006711</td>\n",
       "      <td>TfidfVectorizer(ngram_range=(2, 2), norm=None)</td>\n",
       "      <td>NearMiss(sampling_strategy='majority')</td>\n",
       "      <td>19</td>\n",
       "      <td>26</td>\n",
       "      <td>2</td>\n",
       "      <td>gini</td>\n",
       "      <td>{'vect': TfidfVectorizer(ngram_range=(2, 2), n...</td>\n",
       "      <td>0.060208</td>\n",
       "      <td>0.050519</td>\n",
       "      <td>0.058131</td>\n",
       "      <td>0.056708</td>\n",
       "      <td>0.060858</td>\n",
       "      <td>0.057285</td>\n",
       "      <td>0.003691</td>\n",
       "      <td>100</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>100 rows × 19 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "    mean_fit_time  std_fit_time  mean_score_time  std_score_time  \\\n",
       "80       0.750471      0.102376         0.013160        0.003598   \n",
       "15       0.573973      0.008334         0.015691        0.005228   \n",
       "59       0.939578      0.095348         0.011759        0.001518   \n",
       "62       0.536093      0.024337         0.014927        0.003321   \n",
       "87       0.344816      0.023308         0.012829        0.002858   \n",
       "..            ...           ...              ...             ...   \n",
       "35       0.356033      0.032776         0.025463        0.007204   \n",
       "65       0.260275      0.022035         0.020876        0.004961   \n",
       "16       0.276917      0.015044         0.020389        0.003677   \n",
       "54       1.077123      0.022111         0.025241        0.003827   \n",
       "89       1.444719      0.056866         0.026205        0.006711   \n",
       "\n",
       "                                        param_vect  \\\n",
       "80                               CountVectorizer()   \n",
       "15                      TfidfVectorizer(norm=None)   \n",
       "59                               CountVectorizer()   \n",
       "62                      TfidfVectorizer(norm=None)   \n",
       "87                               CountVectorizer()   \n",
       "..                                             ...   \n",
       "35             CountVectorizer(ngram_range=(2, 2))   \n",
       "65             CountVectorizer(ngram_range=(2, 2))   \n",
       "16             CountVectorizer(ngram_range=(2, 2))   \n",
       "54             CountVectorizer(ngram_range=(2, 2))   \n",
       "89  TfidfVectorizer(ngram_range=(2, 2), norm=None)   \n",
       "\n",
       "                                     param_resampling  \\\n",
       "80             NearMiss(sampling_strategy='majority')   \n",
       "15  SMOTE(random_state=0, sampling_strategy='minor...   \n",
       "59             NearMiss(sampling_strategy='majority')   \n",
       "62  SMOTE(random_state=0, sampling_strategy='minor...   \n",
       "87  SMOTE(random_state=0, sampling_strategy='minor...   \n",
       "..                                                ...   \n",
       "35  SMOTE(random_state=0, sampling_strategy='minor...   \n",
       "65  SMOTE(random_state=0, sampling_strategy='minor...   \n",
       "16  SMOTE(random_state=0, sampling_strategy='minor...   \n",
       "54             NearMiss(sampling_strategy='majority')   \n",
       "89             NearMiss(sampling_strategy='majority')   \n",
       "\n",
       "   param_modeling__min_samples_split param_modeling__min_samples_leaf  \\\n",
       "80                                28                                7   \n",
       "15                                 9                               40   \n",
       "59                                 2                               19   \n",
       "62                                25                               48   \n",
       "87                                47                               17   \n",
       "..                               ...                              ...   \n",
       "35                                24                               10   \n",
       "65                                 2                                8   \n",
       "16                                42                               19   \n",
       "54                                 2                               39   \n",
       "89                                19                               26   \n",
       "\n",
       "   param_modeling__max_depth param_modeling__criterion  \\\n",
       "80                         2                   entropy   \n",
       "15                         5                   entropy   \n",
       "59                         3                  log_loss   \n",
       "62                         3                   entropy   \n",
       "87                        28                  log_loss   \n",
       "..                       ...                       ...   \n",
       "35                         3                   entropy   \n",
       "65                         3                      gini   \n",
       "16                         2                   entropy   \n",
       "54                         4                  log_loss   \n",
       "89                         2                      gini   \n",
       "\n",
       "                                               params  split0_test_score  \\\n",
       "80  {'vect': CountVectorizer(), 'resampling': Near...           0.959862   \n",
       "15  {'vect': TfidfVectorizer(norm=None), 'resampli...           0.951557   \n",
       "59  {'vect': CountVectorizer(), 'resampling': Near...           0.952249   \n",
       "62  {'vect': TfidfVectorizer(norm=None), 'resampli...           0.947405   \n",
       "87  {'vect': CountVectorizer(), 'resampling': SMOT...           0.677509   \n",
       "..                                                ...                ...   \n",
       "35  {'vect': CountVectorizer(ngram_range=(2, 2)), ...           0.142561   \n",
       "65  {'vect': CountVectorizer(ngram_range=(2, 2)), ...           0.142561   \n",
       "16  {'vect': CountVectorizer(ngram_range=(2, 2)), ...           0.103114   \n",
       "54  {'vect': CountVectorizer(ngram_range=(2, 2)), ...           0.108651   \n",
       "89  {'vect': TfidfVectorizer(ngram_range=(2, 2), n...           0.060208   \n",
       "\n",
       "    split1_test_score  split2_test_score  split3_test_score  \\\n",
       "80           0.971626           0.971626           0.967497   \n",
       "15           0.970242           0.965398           0.963347   \n",
       "59           0.966090           0.962630           0.963347   \n",
       "62           0.960554           0.957093           0.963347   \n",
       "87           0.686505           0.703114           0.705394   \n",
       "..                ...                ...                ...   \n",
       "35           0.132872           0.153633           0.150069   \n",
       "65           0.132872           0.153633           0.150069   \n",
       "16           0.091349           0.099654           0.101660   \n",
       "54           0.062284           0.072664           0.074689   \n",
       "89           0.050519           0.058131           0.056708   \n",
       "\n",
       "    split4_test_score  mean_test_score  std_test_score  rank_test_score  \n",
       "80           0.966805         0.967483        0.004310                1  \n",
       "15           0.960581         0.962225        0.006199                2  \n",
       "59           0.955740         0.960011        0.005166                3  \n",
       "62           0.955740         0.956828        0.005412                4  \n",
       "87           0.654910         0.685486        0.018475                5  \n",
       "..                ...              ...             ...              ...  \n",
       "35           0.134163         0.142660        0.008286               96  \n",
       "65           0.134163         0.142660        0.008286               96  \n",
       "16           0.097510         0.098658        0.004113               98  \n",
       "54           0.079530         0.079563        0.015595               99  \n",
       "89           0.060858         0.057285        0.003691              100  \n",
       "\n",
       "[100 rows x 19 columns]"
      ]
     },
     "execution_count": 76,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# mendapatkan hasil parameter yang terbaik\n",
    "pd.DataFrame(randomsearch.cv_results_).sort_values('rank_test_score')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'vect': CountVectorizer(),\n",
       " 'resampling': NearMiss(sampling_strategy='majority'),\n",
       " 'modeling__min_samples_split': 28,\n",
       " 'modeling__min_samples_leaf': 7,\n",
       " 'modeling__max_depth': 2,\n",
       " 'modeling__criterion': 'entropy'}"
      ]
     },
     "execution_count": 77,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# nilai terbaik dari parameter\n",
    "randomsearch.best_params_ "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<style>#sk-container-id-3 {color: black;}#sk-container-id-3 pre{padding: 0;}#sk-container-id-3 div.sk-toggleable {background-color: white;}#sk-container-id-3 label.sk-toggleable__label {cursor: pointer;display: block;width: 100%;margin-bottom: 0;padding: 0.3em;box-sizing: border-box;text-align: center;}#sk-container-id-3 label.sk-toggleable__label-arrow:before {content: \"▸\";float: left;margin-right: 0.25em;color: #696969;}#sk-container-id-3 label.sk-toggleable__label-arrow:hover:before {color: black;}#sk-container-id-3 div.sk-estimator:hover label.sk-toggleable__label-arrow:before {color: black;}#sk-container-id-3 div.sk-toggleable__content {max-height: 0;max-width: 0;overflow: hidden;text-align: left;background-color: #f0f8ff;}#sk-container-id-3 div.sk-toggleable__content pre {margin: 0.2em;color: black;border-radius: 0.25em;background-color: #f0f8ff;}#sk-container-id-3 input.sk-toggleable__control:checked~div.sk-toggleable__content {max-height: 200px;max-width: 100%;overflow: auto;}#sk-container-id-3 input.sk-toggleable__control:checked~label.sk-toggleable__label-arrow:before {content: \"▾\";}#sk-container-id-3 div.sk-estimator input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-3 div.sk-label input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-3 input.sk-hidden--visually {border: 0;clip: rect(1px 1px 1px 1px);clip: rect(1px, 1px, 1px, 1px);height: 1px;margin: -1px;overflow: hidden;padding: 0;position: absolute;width: 1px;}#sk-container-id-3 div.sk-estimator {font-family: monospace;background-color: #f0f8ff;border: 1px dotted black;border-radius: 0.25em;box-sizing: border-box;margin-bottom: 0.5em;}#sk-container-id-3 div.sk-estimator:hover {background-color: #d4ebff;}#sk-container-id-3 div.sk-parallel-item::after {content: \"\";width: 100%;border-bottom: 1px solid gray;flex-grow: 1;}#sk-container-id-3 div.sk-label:hover label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-3 div.sk-serial::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 0;bottom: 0;left: 50%;z-index: 0;}#sk-container-id-3 div.sk-serial {display: flex;flex-direction: column;align-items: center;background-color: white;padding-right: 0.2em;padding-left: 0.2em;position: relative;}#sk-container-id-3 div.sk-item {position: relative;z-index: 1;}#sk-container-id-3 div.sk-parallel {display: flex;align-items: stretch;justify-content: center;background-color: white;position: relative;}#sk-container-id-3 div.sk-item::before, #sk-container-id-3 div.sk-parallel-item::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 0;bottom: 0;left: 50%;z-index: -1;}#sk-container-id-3 div.sk-parallel-item {display: flex;flex-direction: column;z-index: 1;position: relative;background-color: white;}#sk-container-id-3 div.sk-parallel-item:first-child::after {align-self: flex-end;width: 50%;}#sk-container-id-3 div.sk-parallel-item:last-child::after {align-self: flex-start;width: 50%;}#sk-container-id-3 div.sk-parallel-item:only-child::after {width: 0;}#sk-container-id-3 div.sk-dashed-wrapped {border: 1px dashed gray;margin: 0 0.4em 0.5em 0.4em;box-sizing: border-box;padding-bottom: 0.4em;background-color: white;}#sk-container-id-3 div.sk-label label {font-family: monospace;font-weight: bold;display: inline-block;line-height: 1.2em;}#sk-container-id-3 div.sk-label-container {text-align: center;}#sk-container-id-3 div.sk-container {/* jupyter's `normalize.less` sets `[hidden] { display: none; }` but bootstrap.min.css set `[hidden] { display: none !important; }` so we also need the `!important` here to be able to override the default hidden behavior on the sphinx rendered scikit-learn.org. See: https://github.com/scikit-learn/scikit-learn/issues/21755 */display: inline-block !important;position: relative;}#sk-container-id-3 div.sk-text-repr-fallback {display: none;}</style><div id=\"sk-container-id-3\" class=\"sk-top-container\"><div class=\"sk-text-repr-fallback\"><pre>Pipeline(steps=[(&#x27;vect&#x27;, CountVectorizer()),\n",
       "                (&#x27;resampling&#x27;, NearMiss(sampling_strategy=&#x27;majority&#x27;)),\n",
       "                (&#x27;modeling&#x27;,\n",
       "                 DecisionTreeClassifier(criterion=&#x27;entropy&#x27;, max_depth=2,\n",
       "                                        min_samples_leaf=7,\n",
       "                                        min_samples_split=28,\n",
       "                                        random_state=0))])</pre><b>In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook. <br />On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.</b></div><div class=\"sk-container\" hidden><div class=\"sk-item sk-dashed-wrapped\"><div class=\"sk-label-container\"><div class=\"sk-label sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-10\" type=\"checkbox\" ><label for=\"sk-estimator-id-10\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">Pipeline</label><div class=\"sk-toggleable__content\"><pre>Pipeline(steps=[(&#x27;vect&#x27;, CountVectorizer()),\n",
       "                (&#x27;resampling&#x27;, NearMiss(sampling_strategy=&#x27;majority&#x27;)),\n",
       "                (&#x27;modeling&#x27;,\n",
       "                 DecisionTreeClassifier(criterion=&#x27;entropy&#x27;, max_depth=2,\n",
       "                                        min_samples_leaf=7,\n",
       "                                        min_samples_split=28,\n",
       "                                        random_state=0))])</pre></div></div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-11\" type=\"checkbox\" ><label for=\"sk-estimator-id-11\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">CountVectorizer</label><div class=\"sk-toggleable__content\"><pre>CountVectorizer()</pre></div></div></div><div class=\"sk-item\"><div class=\"sk-estimator sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-12\" type=\"checkbox\" ><label for=\"sk-estimator-id-12\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">NearMiss</label><div class=\"sk-toggleable__content\"><pre>NearMiss(sampling_strategy=&#x27;majority&#x27;)</pre></div></div></div><div class=\"sk-item\"><div class=\"sk-estimator sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-13\" type=\"checkbox\" ><label for=\"sk-estimator-id-13\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">DecisionTreeClassifier</label><div class=\"sk-toggleable__content\"><pre>DecisionTreeClassifier(criterion=&#x27;entropy&#x27;, max_depth=2, min_samples_leaf=7,\n",
       "                       min_samples_split=28, random_state=0)</pre></div></div></div></div></div></div></div>"
      ],
      "text/plain": [
       "Pipeline(steps=[('vect', CountVectorizer()),\n",
       "                ('resampling', NearMiss(sampling_strategy='majority')),\n",
       "                ('modeling',\n",
       "                 DecisionTreeClassifier(criterion='entropy', max_depth=2,\n",
       "                                        min_samples_leaf=7,\n",
       "                                        min_samples_split=28,\n",
       "                                        random_state=0))])"
      ]
     },
     "execution_count": 78,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# best estimator\n",
    "randomsearch.best_estimator_"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Predict to test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {},
   "outputs": [],
   "source": [
    "# metric\n",
    "from sklearn.metrics import accuracy_score, precision_score, recall_score, f1_score "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Before tuning"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9623685666851135"
      ]
     },
     "execution_count": 80,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model = Pipeline([\n",
    "    (\"vect\", vect_tf),\n",
    "    (\"modeling\", tree)\n",
    "])\n",
    "model.fit(X_train, y_train)\n",
    "y_pred_before = model.predict(X_test)\n",
    "recall_score(y_test, y_pred_before)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "After tuning"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9690094078583287"
      ]
     },
     "execution_count": 81,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model_tuning = randomsearch.best_estimator_\n",
    "model_tuning.fit(X_train, y_train)\n",
    "y_pred_after = model_tuning.predict(X_test)\n",
    "recall_score(y_test, y_pred_after)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.metrics import classification_report, confusion_matrix, ConfusionMatrixDisplay"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[ 213,  808],\n",
       "       [  68, 1739]])"
      ]
     },
     "execution_count": 89,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cf_before=confusion_matrix(y_test, y_pred_before)\n",
    "cf_before"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 90,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[ 132,  889],\n",
       "       [  56, 1751]])"
      ]
     },
     "execution_count": 90,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cf_after = confusion_matrix(y_test, y_pred_after)\n",
    "cf_after"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.76      0.21      0.33      1021\n",
      "           1       0.68      0.96      0.80      1807\n",
      "\n",
      "    accuracy                           0.69      2828\n",
      "   macro avg       0.72      0.59      0.56      2828\n",
      "weighted avg       0.71      0.69      0.63      2828\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(classification_report(y_test, y_pred_before))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.70      0.13      0.22      1021\n",
      "           1       0.66      0.97      0.79      1807\n",
      "\n",
      "    accuracy                           0.67      2828\n",
      "   macro avg       0.68      0.55      0.50      2828\n",
      "weighted avg       0.68      0.67      0.58      2828\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(classification_report(y_test, y_pred_after))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 91,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Display confusion matrix without numeric values\n",
    "disp = ConfusionMatrixDisplay(confusion_matrix=cf_before, display_labels=['Negative', 'Positive'])\n",
    "disp.plot(cmap='Blues', values_format='')\n",
    "\n",
    "# Add abbreviations above the cells with higher positioning\n",
    "for i in range(2):\n",
    "    for j in range(2):\n",
    "        cell_value = disp.confusion_matrix[i, j]\n",
    "        if i == 0 and j == 0:\n",
    "            plt.text(j, i - 0.2, 'TN', ha='center', va='center', color='skyblue', fontweight='bold')\n",
    "        elif i == 0 and j == 1:\n",
    "            plt.text(j, i - 0.2, 'FP', ha='center', va='center', color='skyblue', fontweight='bold')\n",
    "        elif i == 1 and j == 0:\n",
    "            plt.text(j, i - 0.2, 'FN', ha='center', va='center', color='skyblue', fontweight='bold')\n",
    "        elif i == 1 and j == 1:\n",
    "            plt.text(j, i - 0.2, 'TP', ha='center', va='center', color='skyblue', fontweight='bold')\n",
    "\n",
    "# Show the plot\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Display confusion matrix without numeric values\n",
    "disp = ConfusionMatrixDisplay(confusion_matrix=cf_after, display_labels=['Negative', 'Positive'])\n",
    "disp.plot(cmap='Blues', values_format='')\n",
    "\n",
    "# Add abbreviations above the cells with higher positioning\n",
    "for i in range(2):\n",
    "    for j in range(2):\n",
    "        cell_value = disp.confusion_matrix[i, j]\n",
    "        if i == 0 and j == 0:\n",
    "            plt.text(j, i - 0.2, 'TN', ha='center', va='center', color='skyblue', fontweight='bold')\n",
    "        elif i == 0 and j == 1:\n",
    "            plt.text(j, i - 0.2, 'FP', ha='center', va='center', color='skyblue', fontweight='bold')\n",
    "        elif i == 1 and j == 0:\n",
    "            plt.text(j, i - 0.2, 'FN', ha='center', va='center', color='skyblue', fontweight='bold')\n",
    "        elif i == 1 and j == 1:\n",
    "            plt.text(j, i - 0.2, 'TP', ha='center', va='center', color='skyblue', fontweight='bold')\n",
    "\n",
    "# Show the plot\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "FP: Diprediksi tweet negatif -> alarm nyala -> actualnya tweet positive -> cost : waktu untuk cek tweet\n",
    "\n",
    "FN: Diprediksi tweet positif -> alarm tidak menyala -> actualnya tweet negatif -> cost : tweet negatif tidak di handle (merusak reputasi)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Conclusion"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Berdasarkan kedua confusion matrix (before dan after), kita berhasil mengurangi resiko terjadinya salah memprediksi negatif yang tidak dihandle. Dari 68 menjadi 56, sehingga setidaknya kita terlalu merusak reputasi"
   ]
  },
  {
   "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": "3.11.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}