Works_with_masking_data .ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Deskripsi Proyek"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Perusahaan asuransi bernama \"Sure Tomorrow\" ingin menyelesaikan beberapa masalah dengan bantuan *machine learning*. Anda pun diminta untuk mengevaluasi kemungkinan tersebut.\n",
    "\n",
    "- Tugas 1: Temukan klien yang mirip dengan kriteria klien tertentu. Tugas ini akan memudahkan perusahaan untuk melakukan pemasaran.\n",
    "- Tugas 2: Prediksi apakah klien baru kemungkinan akan mengambil klaim asuransi. Apakah prediksi model lebih baik daripada prediksi model *dummy*?\n",
    "- Tugas 3: Prediksi besaran klaim asuransi yang mungkin diterima klien baru menggunakan model regresi linear.\n",
    "- Tugas 4: Lindungi data pribadi klien tanpa merusak model dari tugas sebelumnya. Sangatlah penting untuk mengembangkan algoritme transformasi data yang dapat mencegah penyalahgunaan informasi pribadi klien jika data tersebut jatuh ke tangan yang salah. Hal ini disebut penyembunyian data atau pengaburan data. Namun, prosedur perlindungan datanya pun perlu diperhatikan agar kualitas *model machine* learning tidak menurun. Di sini, Anda tidak perlu memilih model yang terbaik, cukup buktikan bahwa algoritme Anda bekerja secara akurat."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Pra-pemrosesan & Eksplorasi Data\n",
    "\n",
    "## Inisialisasi"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "\n",
    "import seaborn as sns,matplotlib.pyplot as plt  \n",
    "\n",
    "from sklearn.metrics import confusion_matrix, f1_score, accuracy_score\n",
    "import sklearn.linear_model\n",
    "from sklearn.linear_model import LinearRegression\n",
    "\n",
    "from sklearn.neighbors import NearestNeighbors, KNeighborsClassifier\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "\n",
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "from IPython.display import display"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Muat Data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Muat data dan lakukan pemeriksaan untuk memastikan data bebas dari permasalahan."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.read_csv('/Users/sultanazhari/Downloads/Visual Studio/Project/Triple10/data/insurance_us.csv')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Kami mengganti nama kolom untuk membuat kode terlihat lebih konsisten dengan gayanya."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = df.rename(columns={'Gender': 'gender', 'Age': 'age', 'Salary': 'income', 'Family members': 'family_members', 'Insurance benefits': 'insurance_benefits'})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>gender</th>\n",
       "      <th>age</th>\n",
       "      <th>income</th>\n",
       "      <th>family_members</th>\n",
       "      <th>insurance_benefits</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>556</th>\n",
       "      <td>1</td>\n",
       "      <td>21.0</td>\n",
       "      <td>55300.0</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3002</th>\n",
       "      <td>1</td>\n",
       "      <td>32.0</td>\n",
       "      <td>47500.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2431</th>\n",
       "      <td>1</td>\n",
       "      <td>27.0</td>\n",
       "      <td>61000.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4068</th>\n",
       "      <td>1</td>\n",
       "      <td>22.0</td>\n",
       "      <td>48600.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>298</th>\n",
       "      <td>0</td>\n",
       "      <td>20.0</td>\n",
       "      <td>48200.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2909</th>\n",
       "      <td>1</td>\n",
       "      <td>36.0</td>\n",
       "      <td>35900.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2443</th>\n",
       "      <td>1</td>\n",
       "      <td>25.0</td>\n",
       "      <td>32700.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2186</th>\n",
       "      <td>1</td>\n",
       "      <td>34.0</td>\n",
       "      <td>33600.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>718</th>\n",
       "      <td>1</td>\n",
       "      <td>22.0</td>\n",
       "      <td>32600.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>218</th>\n",
       "      <td>0</td>\n",
       "      <td>29.0</td>\n",
       "      <td>38500.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      gender   age   income  family_members  insurance_benefits\n",
       "556        1  21.0  55300.0               2                   0\n",
       "3002       1  32.0  47500.0               1                   0\n",
       "2431       1  27.0  61000.0               0                   0\n",
       "4068       1  22.0  48600.0               1                   0\n",
       "298        0  20.0  48200.0               1                   0\n",
       "2909       1  36.0  35900.0               1                   0\n",
       "2443       1  25.0  32700.0               0                   0\n",
       "2186       1  34.0  33600.0               1                   0\n",
       "718        1  22.0  32600.0               1                   0\n",
       "218        0  29.0  38500.0               0                   0"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.sample(10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 5000 entries, 0 to 4999\n",
      "Data columns (total 5 columns):\n",
      " #   Column              Non-Null Count  Dtype  \n",
      "---  ------              --------------  -----  \n",
      " 0   gender              5000 non-null   int64  \n",
      " 1   age                 5000 non-null   float64\n",
      " 2   income              5000 non-null   float64\n",
      " 3   family_members      5000 non-null   int64  \n",
      " 4   insurance_benefits  5000 non-null   int64  \n",
      "dtypes: float64(2), int64(3)\n",
      "memory usage: 195.4 KB\n"
     ]
    }
   ],
   "source": [
    "df.info()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div class=\"alert-info\">\n",
    "    \n",
    "\n",
    "<b>Note</b>\n",
    "\n",
    "<b>gender</b> - Jenis kelamin. - Data OK <br>\n",
    "<b>age</b> - Umur. - Type data Float, sehingga perlu diubah menjadi interger <br>\n",
    "<b>income</b> - Penghasilan - Type data Float, sehingga perlu diubah menjadi interger <br>\n",
    "<b>family_members</b> - Jumlah anggota keluarga. - Data OK<br>\n",
    "<b>insurance_benefits</b> - 1 Menggunakan asuransi, 0 Tidak menggunakan asuransi. - Data OK<br>\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Merubah Type data pada kolom umur dan penghasilan\n",
    "df['age'] = df['age'].astype(int)\n",
    "df['income'] = df['income'].astype(int)"
   ]
  },
  {
   "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>gender</th>\n",
       "      <th>age</th>\n",
       "      <th>income</th>\n",
       "      <th>family_members</th>\n",
       "      <th>insurance_benefits</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>41</td>\n",
       "      <td>49600</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>46</td>\n",
       "      <td>38000</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>29</td>\n",
       "      <td>21000</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "      <td>21</td>\n",
       "      <td>41700</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1</td>\n",
       "      <td>28</td>\n",
       "      <td>26100</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   gender  age  income  family_members  insurance_benefits\n",
       "0       1   41   49600               1                   0\n",
       "1       0   46   38000               1                   1\n",
       "2       0   29   21000               0                   0\n",
       "3       0   21   41700               2                   0\n",
       "4       1   28   26100               0                   0"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# periksa apakah proses pengubahannya sudah berhasil\n",
    "df.head(5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "153"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Melakukan pengecekan data duplikat\n",
    "df.duplicated().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Menghapus data duplikat agar tidak terjadi informasi redundant\n",
    "df = df.drop_duplicates()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Melakukan pengecekan kembali\n",
    "df.duplicated().sum()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## EDA"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Mari kita periksa dengan cepat, apakah ada kelompok pelanggan tertentu dengan melihat plot pasangan?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Melihat distribusi antara jenis kelamin dengan penggunaan asuransi. \n",
    "plt.title('Distribusi jenis kelamin dengan penggunaan asuransi')\n",
    "sns.scatterplot(data=df, x='gender', y='insurance_benefits')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div class=\"alert-info\">\n",
    "    \n",
    "\n",
    "<b>Insight</b>\n",
    "    \n",
    "Tidak terdapat perbedaan signifikan antara jenis kelamin dengan penggunaan asuransi\n",
    "\n",
    "\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Melihat distribusi antara umur dengan penggunaan asuransi. \n",
    "plt.title('Distribusi umur dengan penggunaan asuransi')\n",
    "sns.scatterplot(data=df, x='age', y='insurance_benefits')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div class=\"alert-info\">\n",
    "    \n",
    "\n",
    "<b>Insight</b>\n",
    "    \n",
    "Makin tinggi umur, makin banyak pula yang melakukan klaim asuransi.\n",
    "\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Melihat distribusi antara umur dengan penggunaan asuransi. \n",
    "plt.title('Distribusi penghasilan dengan penggunaan asuransi')\n",
    "sns.scatterplot(data=df, x='income', y='insurance_benefits')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div class=\"alert-info\">\n",
    "    \n",
    "\n",
    "<b>Insight</b>\n",
    "    \n",
    "Tidak terdapat perbedaan signifikan antara income dengan penggunaan asuransi\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Melihat distribusi antara jumlah keluarga dengan penggunaan asuransi. \n",
    "plt.title('Distribusi jumlah keluarga dengan penggunaan asuransi')\n",
    "sns.scatterplot(data=df, x='family_members', y='insurance_benefits')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div class=\"alert-info\">\n",
    "    \n",
    "\n",
    "<b>Insight</b>\n",
    "    \n",
    "Dapat dikatakan makin banyak jumlah keluarga, penggunaan klaim asuransi lebih sedikit\n",
    "agak sulit untuk melihat kelompok (klaster) yang jelas mengingat sulitnya menggabungkan beberapa variabel secara bersamaan (untuk menganalisis distribusi multivariat). "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Tugas 1. Klien yang Mirip"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Dalam bahasa pemrograman ML, penting untuk mengembangkan prosedur yang bisa menentukan *k-nearest neighbors* (objek) pada objek tertentu berdasarkan jarak antar objek.\n",
    "\n",
    "Anda mungkin ingin mengulas kembali pelajaran berikut (bab -> pelajaran)\n",
    "- Jarak Antar Vektor -> Jarak Euklides\n",
    "- Jarak Antar Vektor -> Jarak Manhattan\n",
    "\n",
    "Untuk menyelesaikan tugas ini, kita bisa mencoba beberapa metrik jarak."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Buatlah sebuah fungsi yang menampilkan *k-nearest neighbors* terdekat untuk objek ke-n berdasarkan metrik jarak tertentu.  Besarnya klaim asuransi yang diterima tidak perlu diperhitungkan dalam tugas ini.\n",
    "\n",
    "Anda bisa menggunakan implementasi algoritme kNN yang sudah tersedia di Scikit-learn  (periksa [tautan](https://scikit-learn.org/stable/modules/generated/sklearn.neighbors.NearestNeighbors.html#sklearn.neighbors.NearestNeighbors) ini) atau menggunakan punya Anda sendiri.\n",
    "\n",
    "Uji algoritme tersebut untuk empat kombinasi dari dua kasus\n",
    "- Penskalaan\n",
    "  - data tidak diskalakan\n",
    "  - data diskalakan dengan [MaxAbsScaler](https://scikit-learn.org/stable/modules/generated/sklearn.preprocessing.MaxAbsScaler.html) \n",
    "- Metrik Jarak\n",
    "  - Euklides\n",
    "  - Manhattan\n",
    "\n",
    "Jawablah pertanyaan-pertanyaan berikut:\n",
    "- Apakah data yang tidak berskala memengaruhi algoritme kNN? Jika berpengaruh, lalu bagaimanakah bentuknya?\n",
    "- Seberapa mirip hasil yang Anda dapatkan dari metrik Euklides dengan saat Anda menggunakan metrik jarak Manhattan (abaikan ada atau tidaknya skala)?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "feature_names = ['gender', 'age', 'income', 'family_members']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_knn(df, n, k, metric):\n",
    "    \n",
    "    \"\"\"\n",
    "    Menampilkan k nearest neighbors\n",
    "\n",
    "    :param df: DataFrame Pandas digunakan untuk menemukan objek-objek yang mirip di dalamnya\n",
    "    :param n: nomor objek yang dicari k nearest neighbors\n",
    "    :param k: jumlah k nearest neighbors yang ditampilkan\n",
    "    :param metric: nama metrik jarak\n",
    "    \"\"\"\n",
    "\n",
    "    nbrs = NearestNeighbors(n_neighbors=k, algorithm='brute', metric=metric).fit(df[feature_names])\n",
    "    nbrs_distances, nbrs_indices = nbrs.kneighbors([df.iloc[n][feature_names]], k, return_distance=True)\n",
    "\n",
    "    list_index = df.iloc[nbrs_indices[0]].index\n",
    "\n",
    "    df_res = pd.concat([df.iloc[nbrs_indices[0]].reset_index(drop=True),\n",
    "                        pd.DataFrame(nbrs_distances.T, index=nbrs_indices[0], columns=['distance']).reset_index(drop=True)\n",
    "        ], axis=1).set_index(list_index)\n",
    "\n",
    "    return df_res"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Sekarang, mari kita dapatkan catatan yang mirip untuk setiap kombinasi yang ada"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>gender</th>\n",
       "      <th>age</th>\n",
       "      <th>income</th>\n",
       "      <th>family_members</th>\n",
       "      <th>insurance_benefits</th>\n",
       "      <th>distance</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>100</th>\n",
       "      <td>0</td>\n",
       "      <td>19</td>\n",
       "      <td>37300</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1831</th>\n",
       "      <td>0</td>\n",
       "      <td>20</td>\n",
       "      <td>37300</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>1.414214</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4278</th>\n",
       "      <td>1</td>\n",
       "      <td>19</td>\n",
       "      <td>37300</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>1.414214</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2624</th>\n",
       "      <td>1</td>\n",
       "      <td>19</td>\n",
       "      <td>37300</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1.414214</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1086</th>\n",
       "      <td>0</td>\n",
       "      <td>22</td>\n",
       "      <td>37300</td>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "      <td>3.605551</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      gender  age  income  family_members  insurance_benefits  distance\n",
       "100        0   19   37300               1                   0  0.000000\n",
       "1831       0   20   37300               2                   0  1.414214\n",
       "4278       1   19   37300               2                   0  1.414214\n",
       "2624       1   19   37300               0                   0  1.414214\n",
       "1086       0   22   37300               3                   0  3.605551"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# List orang yang dekat dengan orang yang ada di index 100 dengan menggunakan rumus Euclidean\n",
    "get_knn(df, n=100, k=5, metric='euclidean')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>gender</th>\n",
       "      <th>age</th>\n",
       "      <th>income</th>\n",
       "      <th>family_members</th>\n",
       "      <th>insurance_benefits</th>\n",
       "      <th>distance</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>100</th>\n",
       "      <td>0</td>\n",
       "      <td>19</td>\n",
       "      <td>37300</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4278</th>\n",
       "      <td>1</td>\n",
       "      <td>19</td>\n",
       "      <td>37300</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2624</th>\n",
       "      <td>1</td>\n",
       "      <td>19</td>\n",
       "      <td>37300</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1831</th>\n",
       "      <td>0</td>\n",
       "      <td>20</td>\n",
       "      <td>37300</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1086</th>\n",
       "      <td>0</td>\n",
       "      <td>22</td>\n",
       "      <td>37300</td>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "      <td>5.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      gender  age  income  family_members  insurance_benefits  distance\n",
       "100        0   19   37300               1                   0       0.0\n",
       "4278       1   19   37300               2                   0       2.0\n",
       "2624       1   19   37300               0                   0       2.0\n",
       "1831       0   20   37300               2                   0       2.0\n",
       "1086       0   22   37300               3                   0       5.0"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# List orang yang dekat dengan orang yang ada di index 100 dengan menggunakan rumus Manhattan\n",
    "get_knn(df, n=100, k=5, metric='manhattan')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div class=\"alert-info\">\n",
    "    \n",
    "\n",
    "<b>Note</b>\n",
    "\n",
    "Terlihat pada orang yang terdekat yang muncul memiliki kesamaan income, dimana hal ini disebabkan karena nilai income yang tinggi dan dianggap dominan.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Mengatur skala data."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "transformer_mas = sklearn.preprocessing.MaxAbsScaler().fit(df[feature_names].to_numpy())\n",
    "\n",
    "df_scaled = df.copy()\n",
    "df_scaled.loc[:, feature_names] = transformer_mas.transform(df[feature_names].to_numpy())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>gender</th>\n",
       "      <th>age</th>\n",
       "      <th>income</th>\n",
       "      <th>family_members</th>\n",
       "      <th>insurance_benefits</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2422</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.307692</td>\n",
       "      <td>0.350633</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1389</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.369231</td>\n",
       "      <td>0.464557</td>\n",
       "      <td>0.166667</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2026</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.615385</td>\n",
       "      <td>0.563291</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4819</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.507692</td>\n",
       "      <td>0.496203</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>518</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.507692</td>\n",
       "      <td>0.482278</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      gender       age    income  family_members  insurance_benefits\n",
       "2422     0.0  0.307692  0.350633        0.000000                   0\n",
       "1389     1.0  0.369231  0.464557        0.166667                   0\n",
       "2026     1.0  0.615385  0.563291        0.000000                   0\n",
       "4819     0.0  0.507692  0.496203        0.000000                   0\n",
       "518      1.0  0.507692  0.482278        0.333333                   0"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_scaled.sample(5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>gender</th>\n",
       "      <th>age</th>\n",
       "      <th>income</th>\n",
       "      <th>family_members</th>\n",
       "      <th>insurance_benefits</th>\n",
       "      <th>distance</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>100</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.292308</td>\n",
       "      <td>0.472152</td>\n",
       "      <td>0.166667</td>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>243</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.292308</td>\n",
       "      <td>0.467089</td>\n",
       "      <td>0.166667</td>\n",
       "      <td>0</td>\n",
       "      <td>0.005063</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4915</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.292308</td>\n",
       "      <td>0.477215</td>\n",
       "      <td>0.166667</td>\n",
       "      <td>0</td>\n",
       "      <td>0.005063</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1323</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.292308</td>\n",
       "      <td>0.456962</td>\n",
       "      <td>0.166667</td>\n",
       "      <td>0</td>\n",
       "      <td>0.015190</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>811</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.307692</td>\n",
       "      <td>0.470886</td>\n",
       "      <td>0.166667</td>\n",
       "      <td>0</td>\n",
       "      <td>0.015437</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      gender       age    income  family_members  insurance_benefits  distance\n",
       "100      0.0  0.292308  0.472152        0.166667                   0  0.000000\n",
       "243      0.0  0.292308  0.467089        0.166667                   0  0.005063\n",
       "4915     0.0  0.292308  0.477215        0.166667                   0  0.005063\n",
       "1323     0.0  0.292308  0.456962        0.166667                   0  0.015190\n",
       "811      0.0  0.307692  0.470886        0.166667                   0  0.015437"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# List orang yang dekat dengan orang yang ada di index 100 setelah discaling dan menggunakan rumus euclidean\n",
    "get_knn(df_scaled, n=100, k=5, metric='euclidean')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "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>gender</th>\n",
       "      <th>age</th>\n",
       "      <th>income</th>\n",
       "      <th>family_members</th>\n",
       "      <th>insurance_benefits</th>\n",
       "      <th>distance</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>100</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.292308</td>\n",
       "      <td>0.472152</td>\n",
       "      <td>0.166667</td>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4915</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.292308</td>\n",
       "      <td>0.477215</td>\n",
       "      <td>0.166667</td>\n",
       "      <td>0</td>\n",
       "      <td>0.005063</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>243</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.292308</td>\n",
       "      <td>0.467089</td>\n",
       "      <td>0.166667</td>\n",
       "      <td>0</td>\n",
       "      <td>0.005063</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1323</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.292308</td>\n",
       "      <td>0.456962</td>\n",
       "      <td>0.166667</td>\n",
       "      <td>0</td>\n",
       "      <td>0.015190</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>811</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.307692</td>\n",
       "      <td>0.470886</td>\n",
       "      <td>0.166667</td>\n",
       "      <td>0</td>\n",
       "      <td>0.016650</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      gender       age    income  family_members  insurance_benefits  distance\n",
       "100      0.0  0.292308  0.472152        0.166667                   0  0.000000\n",
       "4915     0.0  0.292308  0.477215        0.166667                   0  0.005063\n",
       "243      0.0  0.292308  0.467089        0.166667                   0  0.005063\n",
       "1323     0.0  0.292308  0.456962        0.166667                   0  0.015190\n",
       "811      0.0  0.307692  0.470886        0.166667                   0  0.016650"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# List orang yang dekat dengan orang yang ada di index 100 setelah discaling dan menggunakan rumus Manhattan\n",
    "get_knn(df_scaled, n=100, k=5, metric='manhattan')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Apakah data yang tidak berskala memengaruhi algoritme kNN? Jika berpengaruh, lalu bagaimanakah bentuknya?** \n",
    "\n",
    "Ya Berpengaruh, Data yang muncul adalah data yang mimiliki kesamaan pada nilai dominan.\n",
    "pada kasus ini nilai tinggi yang dominan adalah income. sehingga pada saat menggunakan algoritma kNN, yang dicari terdekat berdasarkan income atau nilai dominan.\n",
    "\n",
    "\n",
    "sedangkan setelah dilakukan skala, datanya cendrung sesuai."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Seberapa mirip hasil yang Anda dapatkan dari metrik Euklides dengan saat Anda menggunakan metrik jarak Manhattan (abaikan ada atau tidaknya skala)?** \n",
    "\n",
    "Keduanya data yang muncul mirip saat menggunakan metrik euklides dan manhattan."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Tugas 2. Apakah klien kemungkinan menerima klaim asuransi?"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Dalam konteks *machine learning*, tugas ini sama seperti tugas klasifikasi biner."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Dengan target insurance_benefits yang lebih dari nol, coba evaluasi apakah model klasifikasi kNN merupakan pendekatan yang lebih baik daripada model *dummy*.\n",
    "\n",
    "Instruksi:\n",
    "- Buat pengklasifikasi berbasis KNN dan ukur kualitasnya dengan metrik F1 untuk k=1..10 bagi data asli dan data yang diskalakan. Akan menarik untuk melihat bagaimana k dapat memengaruhi metrik evaluasi, dan apakah penskalaan data membuat hasilnya berbeda. Gunakan implementasi algoritme klasifikasi kNN yang tersedia di Scikit-learn (periksa [tautan ini](https://scikit-learn.org/stable/modules/generated/sklearn.neighbors.KNeighborsClassifier.html)) atau gunakan milik Anda sendiri.\n",
    "- Buat model *dummy* yang acak untuk kasus ini. Model tersebut harusnya menampilkan \"1\" dengan beberapa probabilitas. Mari kita uji model dengan empat nilai probabilitas: 0, probabilitas membayar benefit asuransi apa pun; 0,5; 1.\n",
    "\n",
    "Probabilitas membayar klaim asuransi dapat didefinisikan sebagai\n",
    "\n",
    "$$\n",
    "P\\{\\text{klaim asuransi yang diterima}\\}=\\frac{\\text{jumlah klien yang menerima klaim asuransi}}{\\text{jumlah klien secara keseluruhan}}.\n",
    "$$\n",
    "\n",
    "Pisahkan keseluruhan data menjadi 70:30 untuk proporsi *training* dan *test set*."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "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>gender</th>\n",
       "      <th>age</th>\n",
       "      <th>income</th>\n",
       "      <th>family_members</th>\n",
       "      <th>insurance_benefits</th>\n",
       "      <th>insurance_benefits_received</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>41</td>\n",
       "      <td>49600</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>46</td>\n",
       "      <td>38000</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>29</td>\n",
       "      <td>21000</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "      <td>21</td>\n",
       "      <td>41700</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1</td>\n",
       "      <td>28</td>\n",
       "      <td>26100</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   gender  age  income  family_members  insurance_benefits  \\\n",
       "0       1   41   49600               1                   0   \n",
       "1       0   46   38000               1                   1   \n",
       "2       0   29   21000               0                   0   \n",
       "3       0   21   41700               2                   0   \n",
       "4       1   28   26100               0                   0   \n",
       "\n",
       "   insurance_benefits_received  \n",
       "0                            0  \n",
       "1                            1  \n",
       "2                            0  \n",
       "3                            0  \n",
       "4                            0  "
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Bikin kolom baru untuk mengubah orang yang claim dan yang tidak\n",
    "df['insurance_benefits_received'] = (df['insurance_benefits'] > 0).astype('int')\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    4284\n",
       "1     563\n",
       "Name: insurance_benefits_received, dtype: int64"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# periksa ketidakseimbangan kelas dengan value_counts()\n",
    "df['insurance_benefits_received'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "def eval_classifier(y_true, y_pred):\n",
    "    \n",
    "    f1_score = sklearn.metrics.f1_score(y_true, y_pred.round())\n",
    "    acc = sklearn.metrics.accuracy_score(y_true, y_pred.round())\n",
    "    print(f'F1: {f1_score:.2f}')\n",
    "    print(f'Accuraccy: {acc:.2f}')\n",
    "    \n",
    "    \n",
    "# jika Anda memiliki masalah dengan baris berikut, muat ulang kernel dan jalankan notebook kembali\n",
    "    cm = sklearn.metrics.confusion_matrix(y_true, y_pred.round(), normalize='all')\n",
    "    print('Matriks Kesalahan')\n",
    "    print(cm)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "# menghasilkan output dari model acak\n",
    "\n",
    "def rnd_model_predict(P, size, seed=42):\n",
    "\n",
    "    rng = np.random.default_rng(seed=seed)\n",
    "    return rng.binomial(n=1, p=P, size=size)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Probabilitasnya: 0.00\n",
      "F1: 0.00\n",
      "Accuraccy: 0.88\n",
      "Matriks Kesalahan\n",
      "[[0.88384568 0.        ]\n",
      " [0.11615432 0.        ]]\n",
      "\n",
      "Probabilitasnya: 0.12\n",
      "F1: 0.13\n",
      "Accuraccy: 0.80\n",
      "Matriks Kesalahan\n",
      "[[0.78502166 0.09882401]\n",
      " [0.1017124  0.01444192]]\n",
      "\n",
      "Probabilitasnya: 0.50\n",
      "F1: 0.19\n",
      "Accuraccy: 0.51\n",
      "Matriks Kesalahan\n",
      "[[0.44873117 0.4351145 ]\n",
      " [0.05921188 0.05694244]]\n",
      "\n",
      "Probabilitasnya: 1.00\n",
      "F1: 0.21\n",
      "Accuraccy: 0.12\n",
      "Matriks Kesalahan\n",
      "[[0.         0.88384568]\n",
      " [0.         0.11615432]]\n",
      "\n"
     ]
    }
   ],
   "source": [
    "for P in [0, df['insurance_benefits_received'].sum() / len(df), 0.5, 1]:\n",
    "\n",
    "    print(f'Probabilitasnya: {P:.2f}')\n",
    "    y_pred_rnd = rnd_model_predict(P, size=len(df), seed=42)\n",
    "        \n",
    "    eval_classifier(df['insurance_benefits_received'], y_pred_rnd)\n",
    "    \n",
    "    print()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Train Test Split\n",
    "X = df[feature_names]\n",
    "y = df['insurance_benefits_received']\n",
    "\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "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>gender</th>\n",
       "      <th>age</th>\n",
       "      <th>income</th>\n",
       "      <th>family_members</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>-0.999742</td>\n",
       "      <td>-1.186407</td>\n",
       "      <td>-0.224105</td>\n",
       "      <td>1.622794</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>-0.999742</td>\n",
       "      <td>-0.949242</td>\n",
       "      <td>0.170696</td>\n",
       "      <td>-1.095339</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1.000258</td>\n",
       "      <td>-1.304990</td>\n",
       "      <td>1.112145</td>\n",
       "      <td>1.622794</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1.000258</td>\n",
       "      <td>1.896740</td>\n",
       "      <td>-2.147496</td>\n",
       "      <td>-0.189295</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1.000258</td>\n",
       "      <td>-0.119164</td>\n",
       "      <td>0.089711</td>\n",
       "      <td>-1.095339</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     gender       age    income  family_members\n",
       "0 -0.999742 -1.186407 -0.224105        1.622794\n",
       "1 -0.999742 -0.949242  0.170696       -1.095339\n",
       "2  1.000258 -1.304990  1.112145        1.622794\n",
       "3  1.000258  1.896740 -2.147496       -0.189295\n",
       "4  1.000258 -0.119164  0.089711       -1.095339"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Scaling data\n",
    "scaler = StandardScaler()\n",
    "scaler.fit(X_train)\n",
    "\n",
    "X_train = pd.DataFrame(scaler.transform(X_train), columns=feature_names)\n",
    "X_train.head()"
   ]
  },
  {
   "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>gender</th>\n",
       "      <th>age</th>\n",
       "      <th>income</th>\n",
       "      <th>family_members</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1.000258</td>\n",
       "      <td>-0.237746</td>\n",
       "      <td>1.061530</td>\n",
       "      <td>-0.189295</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>-0.999742</td>\n",
       "      <td>-0.119164</td>\n",
       "      <td>-1.428755</td>\n",
       "      <td>-1.095339</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>-0.999742</td>\n",
       "      <td>1.066662</td>\n",
       "      <td>0.150450</td>\n",
       "      <td>1.622794</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1.000258</td>\n",
       "      <td>0.118001</td>\n",
       "      <td>-0.821369</td>\n",
       "      <td>-1.095339</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1.000258</td>\n",
       "      <td>1.659575</td>\n",
       "      <td>-0.426567</td>\n",
       "      <td>-1.095339</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     gender       age    income  family_members\n",
       "0  1.000258 -0.237746  1.061530       -0.189295\n",
       "1 -0.999742 -0.119164 -1.428755       -1.095339\n",
       "2 -0.999742  1.066662  0.150450        1.622794\n",
       "3  1.000258  0.118001 -0.821369       -1.095339\n",
       "4  1.000258  1.659575 -0.426567       -1.095339"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_test = pd.DataFrame(scaler.transform(X_test), columns=feature_names)\n",
    "X_test.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "KNeighborsClassifier(n_neighbors=3)"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Bikin model dengan menggunakan KNN\n",
    "knn = KNeighborsClassifier(n_neighbors=3)\n",
    "knn.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "F1: 0.98\n",
      "Accuraccy: 0.99\n",
      "Matriks Kesalahan\n",
      "[[8.85478463e-01 7.73794171e-04]\n",
      " [4.38483363e-03 1.09362909e-01]]\n"
     ]
    }
   ],
   "source": [
    "# Predict data Training\n",
    "y_pred_train = knn.predict(X_train)\n",
    "eval_classifier(y_train, y_pred_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "F1: 0.94\n",
      "Accuraccy: 0.99\n",
      "Matriks Kesalahan\n",
      "[[0.87216495 0.00206186]\n",
      " [0.01134021 0.11443299]]\n"
     ]
    }
   ],
   "source": [
    "# Predict data Testing\n",
    "y_pred_test = knn.predict(X_test)\n",
    "eval_classifier(y_test, y_pred_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Tugas 3. Regresi (dengan Regresi Linear)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Dengan `insurance_benefit` sebagai target, evaluasi berapa RMSE untuk model Regresi Linear."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LinearRegression()"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "lr = LinearRegression()\n",
    "lr.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "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>gender</th>\n",
       "      <th>age</th>\n",
       "      <th>income</th>\n",
       "      <th>family_members</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>-0.999742</td>\n",
       "      <td>-1.186407</td>\n",
       "      <td>-0.224105</td>\n",
       "      <td>1.622794</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>-0.999742</td>\n",
       "      <td>-0.949242</td>\n",
       "      <td>0.170696</td>\n",
       "      <td>-1.095339</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1.000258</td>\n",
       "      <td>-1.304990</td>\n",
       "      <td>1.112145</td>\n",
       "      <td>1.622794</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1.000258</td>\n",
       "      <td>1.896740</td>\n",
       "      <td>-2.147496</td>\n",
       "      <td>-0.189295</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1.000258</td>\n",
       "      <td>-0.119164</td>\n",
       "      <td>0.089711</td>\n",
       "      <td>-1.095339</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     gender       age    income  family_members\n",
       "0 -0.999742 -1.186407 -0.224105        1.622794\n",
       "1 -0.999742 -0.949242  0.170696       -1.095339\n",
       "2  1.000258 -1.304990  1.112145        1.622794\n",
       "3  1.000258  1.896740 -2.147496       -0.189295\n",
       "4  1.000258 -0.119164  0.089711       -1.095339"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "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>gender</th>\n",
       "      <th>age</th>\n",
       "      <th>income</th>\n",
       "      <th>family_members</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1.000258</td>\n",
       "      <td>-0.237746</td>\n",
       "      <td>1.061530</td>\n",
       "      <td>-0.189295</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>-0.999742</td>\n",
       "      <td>-0.119164</td>\n",
       "      <td>-1.428755</td>\n",
       "      <td>-1.095339</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>-0.999742</td>\n",
       "      <td>1.066662</td>\n",
       "      <td>0.150450</td>\n",
       "      <td>1.622794</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1.000258</td>\n",
       "      <td>0.118001</td>\n",
       "      <td>-0.821369</td>\n",
       "      <td>-1.095339</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1.000258</td>\n",
       "      <td>1.659575</td>\n",
       "      <td>-0.426567</td>\n",
       "      <td>-1.095339</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     gender       age    income  family_members\n",
       "0  1.000258 -0.237746  1.061530       -0.189295\n",
       "1 -0.999742 -0.119164 -1.428755       -1.095339\n",
       "2 -0.999742  1.066662  0.150450        1.622794\n",
       "3  1.000258  0.118001 -0.821369       -1.095339\n",
       "4  1.000258  1.659575 -0.426567       -1.095339"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_test.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div class=\"alert-info\">\n",
    "    \n",
    "\n",
    "<b>Note</b>\n",
    "\n",
    "Data sudah discaling"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "F1: 0.53\n",
      "Accuraccy: 0.93\n",
      "Matriks Kesalahan\n",
      "[[0.88625226 0.        ]\n",
      " [0.07299458 0.04075316]]\n"
     ]
    }
   ],
   "source": [
    "# Predict data Training\n",
    "y_pred_train = lr.predict(X_train)\n",
    "eval_classifier(y_train, y_pred_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "F1: 0.64\n",
      "Accuraccy: 0.93\n",
      "Matriks Kesalahan\n",
      "[[0.8742268  0.        ]\n",
      " [0.06701031 0.05876289]]\n"
     ]
    }
   ],
   "source": [
    "# Predict data Testing\n",
    "y_pred_test = lr.predict(X_test)\n",
    "eval_classifier(y_test, y_pred_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Tugas 4. Pengaburan Data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Cara terbaik untuk mengaburkan data adalah dengan mengalikan fitur-fitur numerik (ingat bahwa fitur-fitur tersebut bisa di lihat di matriks $X$) dengan matriks yang dapat dibalik (*invertible*) $P$. \n",
    "\n",
    "$$\n",
    "X' = X \\times P\n",
    "$$\n",
    "\n",
    "Coba lakukan itu dan periksa bagaimana nilai fitur akan terlihat setelah transformasi. Sifat *invertible* sangat penting pada kasus ini, jadi pastikan $P$ bahwa $P$ bersifat *invertible*.\n",
    "\n",
    "Anda mungkin ingin meninjau kembali pelajaran 'Matriks dan Operasi Matriks -> Perkalian Matriks' untuk mengingat aturan perkalian matriks dan implementasinya dengan NumPy."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {},
   "outputs": [],
   "source": [
    "personal_info_column_list = ['gender', 'age', 'income', 'family_members']\n",
    "df_pn = df[personal_info_column_list]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_masking = df_pn.to_numpy()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Membuat matriks acak $P$."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {},
   "outputs": [],
   "source": [
    "rng = np.random.default_rng(seed=42)\n",
    "P = rng.random(size=(df_masking.shape[1], df_masking.shape[1]))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Memeriksa apakah matriks $P$ *invertible*"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.24339135998015463\n"
     ]
    }
   ],
   "source": [
    "P_det = np.linalg.det(P)\n",
    "\n",
    "# Kalau determinan dari matrix nya adalah 0, maka kita tidak bisa masking datanya\n",
    "print(P_det)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Bisakah Anda menebak usia klien atau pendapatannya setelah dilakukan transformasi?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "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>gender</th>\n",
       "      <th>age</th>\n",
       "      <th>income</th>\n",
       "      <th>family_members</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>6359.715273</td>\n",
       "      <td>22380.404676</td>\n",
       "      <td>18424.090742</td>\n",
       "      <td>46000.696690</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>4873.294065</td>\n",
       "      <td>17160.367030</td>\n",
       "      <td>14125.780761</td>\n",
       "      <td>35253.455773</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2693.117429</td>\n",
       "      <td>9486.397744</td>\n",
       "      <td>7808.831560</td>\n",
       "      <td>19484.860631</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>5345.603937</td>\n",
       "      <td>18803.227203</td>\n",
       "      <td>15479.148373</td>\n",
       "      <td>38663.061863</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>3347.176735</td>\n",
       "      <td>11782.829283</td>\n",
       "      <td>9699.998942</td>\n",
       "      <td>24211.273378</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        gender           age        income  family_members\n",
       "0  6359.715273  22380.404676  18424.090742    46000.696690\n",
       "1  4873.294065  17160.367030  14125.780761    35253.455773\n",
       "2  2693.117429   9486.397744   7808.831560    19484.860631\n",
       "3  5345.603937  18803.227203  15479.148373    38663.061863\n",
       "4  3347.176735  11782.829283   9699.998942    24211.273378"
      ]
     },
     "execution_count": 81,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Masking data dengan matrix yang sudah kita buat\n",
    "df_masking = df_masking @ P\n",
    "df_masking = pd.DataFrame(df_masking, columns=personal_info_column_list)\n",
    "df_masking.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Bisakah Anda memulihkan data asli dari $X'$ jika $P$ sudah diketahui? Coba periksa hal tersebut melalui perhitungan dengan memindahkan $P$ dari sisi kanan rumus di atas ke sisi kiri. Aturan perkalian matriks sangat berguna di sini."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Mengembalikan data yang sudah di masking\n",
    "P_inv = np.linalg.inv(P)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "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>gender</th>\n",
       "      <th>age</th>\n",
       "      <th>income</th>\n",
       "      <th>family_members</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1.000000e+00</td>\n",
       "      <td>41.0</td>\n",
       "      <td>49600.0</td>\n",
       "      <td>1.000000e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>-4.473636e-12</td>\n",
       "      <td>46.0</td>\n",
       "      <td>38000.0</td>\n",
       "      <td>1.000000e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>-2.515869e-12</td>\n",
       "      <td>29.0</td>\n",
       "      <td>21000.0</td>\n",
       "      <td>9.524523e-13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>-4.844982e-12</td>\n",
       "      <td>21.0</td>\n",
       "      <td>41700.0</td>\n",
       "      <td>2.000000e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1.000000e+00</td>\n",
       "      <td>28.0</td>\n",
       "      <td>26100.0</td>\n",
       "      <td>-1.019907e-13</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         gender   age   income  family_members\n",
       "0  1.000000e+00  41.0  49600.0    1.000000e+00\n",
       "1 -4.473636e-12  46.0  38000.0    1.000000e+00\n",
       "2 -2.515869e-12  29.0  21000.0    9.524523e-13\n",
       "3 -4.844982e-12  21.0  41700.0    2.000000e+00\n",
       "4  1.000000e+00  28.0  26100.0   -1.019907e-13"
      ]
     },
     "execution_count": 83,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_return = df_masking @ P_inv\n",
    "df_return.columns = personal_info_column_list\n",
    "df_return.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Tampilkan ketiga kasus untuk beberapa klien\n",
    "- Data asli\n",
    "- Data yang sudah ditransformasikan\n",
    "- Data yang telah dipulihkan ke semula"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "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 tr th {\n",
       "        text-align: left;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th colspan=\"6\" halign=\"left\">Asli</th>\n",
       "      <th colspan=\"4\" halign=\"left\">Masked</th>\n",
       "      <th colspan=\"4\" halign=\"left\">Return</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th>gender</th>\n",
       "      <th>age</th>\n",
       "      <th>income</th>\n",
       "      <th>family_members</th>\n",
       "      <th>insurance_benefits</th>\n",
       "      <th>insurance_benefits_received</th>\n",
       "      <th>gender</th>\n",
       "      <th>age</th>\n",
       "      <th>income</th>\n",
       "      <th>family_members</th>\n",
       "      <th>gender</th>\n",
       "      <th>age</th>\n",
       "      <th>income</th>\n",
       "      <th>family_members</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1.0</td>\n",
       "      <td>41.0</td>\n",
       "      <td>49600.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>6359.715273</td>\n",
       "      <td>22380.404676</td>\n",
       "      <td>18424.090742</td>\n",
       "      <td>46000.696690</td>\n",
       "      <td>1.000000e+00</td>\n",
       "      <td>41.0</td>\n",
       "      <td>49600.0</td>\n",
       "      <td>1.000000e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.0</td>\n",
       "      <td>46.0</td>\n",
       "      <td>38000.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>4873.294065</td>\n",
       "      <td>17160.367030</td>\n",
       "      <td>14125.780761</td>\n",
       "      <td>35253.455773</td>\n",
       "      <td>-4.473636e-12</td>\n",
       "      <td>46.0</td>\n",
       "      <td>38000.0</td>\n",
       "      <td>1.000000e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.0</td>\n",
       "      <td>29.0</td>\n",
       "      <td>21000.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2693.117429</td>\n",
       "      <td>9486.397744</td>\n",
       "      <td>7808.831560</td>\n",
       "      <td>19484.860631</td>\n",
       "      <td>-2.515869e-12</td>\n",
       "      <td>29.0</td>\n",
       "      <td>21000.0</td>\n",
       "      <td>9.524523e-13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.0</td>\n",
       "      <td>21.0</td>\n",
       "      <td>41700.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>5345.603937</td>\n",
       "      <td>18803.227203</td>\n",
       "      <td>15479.148373</td>\n",
       "      <td>38663.061863</td>\n",
       "      <td>-4.844982e-12</td>\n",
       "      <td>21.0</td>\n",
       "      <td>41700.0</td>\n",
       "      <td>2.000000e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1.0</td>\n",
       "      <td>28.0</td>\n",
       "      <td>26100.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>3347.176735</td>\n",
       "      <td>11782.829283</td>\n",
       "      <td>9699.998942</td>\n",
       "      <td>24211.273378</td>\n",
       "      <td>1.000000e+00</td>\n",
       "      <td>28.0</td>\n",
       "      <td>26100.0</td>\n",
       "      <td>-1.019907e-13</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    Asli                                                   \\\n",
       "  gender   age   income family_members insurance_benefits   \n",
       "0    1.0  41.0  49600.0            1.0                0.0   \n",
       "1    0.0  46.0  38000.0            1.0                1.0   \n",
       "2    0.0  29.0  21000.0            0.0                0.0   \n",
       "3    0.0  21.0  41700.0            2.0                0.0   \n",
       "4    1.0  28.0  26100.0            0.0                0.0   \n",
       "\n",
       "                                    Masked                              \\\n",
       "  insurance_benefits_received       gender           age        income   \n",
       "0                         0.0  6359.715273  22380.404676  18424.090742   \n",
       "1                         1.0  4873.294065  17160.367030  14125.780761   \n",
       "2                         0.0  2693.117429   9486.397744   7808.831560   \n",
       "3                         0.0  5345.603937  18803.227203  15479.148373   \n",
       "4                         0.0  3347.176735  11782.829283   9699.998942   \n",
       "\n",
       "                        Return                                \n",
       "  family_members        gender   age   income family_members  \n",
       "0   46000.696690  1.000000e+00  41.0  49600.0   1.000000e+00  \n",
       "1   35253.455773 -4.473636e-12  46.0  38000.0   1.000000e+00  \n",
       "2   19484.860631 -2.515869e-12  29.0  21000.0   9.524523e-13  \n",
       "3   38663.061863 -4.844982e-12  21.0  41700.0   2.000000e+00  \n",
       "4   24211.273378  1.000000e+00  28.0  26100.0  -1.019907e-13  "
      ]
     },
     "execution_count": 84,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.concat([df, df_masking, df_return], axis=1, keys=['Asli', 'Masked', 'Return']).head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##  Buktikan Bahwa Pengaburan Data Bisa Bekerja dengan LR"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Tugas regresi pada proyek ini telah diselesaikan dengan regresi linear. Tugas selanjutnya adalah untuk membuktikan secara analitis bahwa metode pengaburan data tertentu tidak akan memengaruhi prediksi nilai regresi linear - nilai-nilai prediksi tersebut tidak akan berubah dari hasil awalnya. Apakah Anda yakin Anda bisa melakukannya?? Anda tidak harus yakin sekarang, tetapi Anda harus membuktikannya!"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Jadi, data yang ada telah dikaburkan dan sekarang terdapat $X \\times P$ alih-alih hanya $X$. Akibatnya, ada bobot lain $w_P$ sebagai\n",
    "$$\n",
    "w = (X^T X)^{-1} X^T y \\quad \\Rightarrow \\quad w_P = [(XP)^T XP]^{-1} (XP)^T y\n",
    "$$\n",
    "\n",
    "Bagaimana menghubungkan $w$ dan $w_P$ jika Anda menyederhanakan rumus untuk wP di atas?\n",
    "\n",
    "Nilai apa yang akan $w_P$ prediksi? \n",
    "\n",
    "Apa artinya bagi kualitas regresi linier jika Anda mengukurnya dengan RMSE?\n",
    "\n",
    "Periksa lampiran B (Sifat-sifat Matriks) di bagian *notebook*. Tersedia beberapa rumus yang berguna di sana!\n",
    "\n",
    "Tidak ada kode yang begitu penting di sesi ini, hanya penjelasan analitis!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Training dengan data yang dimasking\n",
    "X_train, X_test, y_train, y_test = train_test_split(df_masking, y, test_size=0.2, random_state=10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "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>gender</th>\n",
       "      <th>age</th>\n",
       "      <th>income</th>\n",
       "      <th>family_members</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>-0.224252</td>\n",
       "      <td>-0.226030</td>\n",
       "      <td>-0.226096</td>\n",
       "      <td>-0.224962</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.169185</td>\n",
       "      <td>0.168677</td>\n",
       "      <td>0.168777</td>\n",
       "      <td>0.169944</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1.112581</td>\n",
       "      <td>1.110173</td>\n",
       "      <td>1.110247</td>\n",
       "      <td>1.111305</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>-2.146178</td>\n",
       "      <td>-2.144094</td>\n",
       "      <td>-2.144185</td>\n",
       "      <td>-2.146132</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.089330</td>\n",
       "      <td>0.089321</td>\n",
       "      <td>0.089478</td>\n",
       "      <td>0.089635</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     gender       age    income  family_members\n",
       "0 -0.224252 -0.226030 -0.226096       -0.224962\n",
       "1  0.169185  0.168677  0.168777        0.169944\n",
       "2  1.112581  1.110173  1.110247        1.111305\n",
       "3 -2.146178 -2.144094 -2.144185       -2.146132\n",
       "4  0.089330  0.089321  0.089478        0.089635"
      ]
     },
     "execution_count": 86,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Scaling data\n",
    "scaler = StandardScaler()\n",
    "scaler.fit(X_train)\n",
    "\n",
    "X_train = pd.DataFrame(scaler.transform(X_train), columns=personal_info_column_list)\n",
    "X_train.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "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>gender</th>\n",
       "      <th>age</th>\n",
       "      <th>income</th>\n",
       "      <th>family_members</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1.061615</td>\n",
       "      <td>1.061158</td>\n",
       "      <td>1.061257</td>\n",
       "      <td>1.061411</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>-1.429799</td>\n",
       "      <td>-1.429326</td>\n",
       "      <td>-1.429298</td>\n",
       "      <td>-1.428938</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.151730</td>\n",
       "      <td>0.152712</td>\n",
       "      <td>0.152426</td>\n",
       "      <td>0.151231</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>-0.821631</td>\n",
       "      <td>-0.821369</td>\n",
       "      <td>-0.821231</td>\n",
       "      <td>-0.821291</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>-0.425849</td>\n",
       "      <td>-0.423696</td>\n",
       "      <td>-0.423709</td>\n",
       "      <td>-0.425366</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     gender       age    income  family_members\n",
       "0  1.061615  1.061158  1.061257        1.061411\n",
       "1 -1.429799 -1.429326 -1.429298       -1.428938\n",
       "2  0.151730  0.152712  0.152426        0.151231\n",
       "3 -0.821631 -0.821369 -0.821231       -0.821291\n",
       "4 -0.425849 -0.423696 -0.423709       -0.425366"
      ]
     },
     "execution_count": 87,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_test = pd.DataFrame(scaler.transform(X_test), columns=personal_info_column_list)\n",
    "X_test.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "KNeighborsClassifier(n_neighbors=3)"
      ]
     },
     "execution_count": 88,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Bikin model dengan menggunakan KNN\n",
    "knn = KNeighborsClassifier(n_neighbors=3)\n",
    "knn.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "F1: 0.74\n",
      "Accuraccy: 0.95\n",
      "Matriks Kesalahan\n",
      "[[0.8800619  0.00619035]\n",
      " [0.04384834 0.06989941]]\n"
     ]
    }
   ],
   "source": [
    "# Predict data Training\n",
    "y_pred_train = knn.predict(X_train)\n",
    "eval_classifier(y_train, y_pred_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 90,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "F1: 0.35\n",
      "Accuraccy: 0.89\n",
      "Matriks Kesalahan\n",
      "[[0.85773196 0.01649485]\n",
      " [0.09587629 0.02989691]]\n"
     ]
    }
   ],
   "source": [
    "# Predict data Testing\n",
    "y_pred_test = knn.predict(X_test)\n",
    "eval_classifier(y_test, y_pred_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Uji Regresi Linear dengan Pengaburan Data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Sekarang, mari kita buktikan kalau Regresi Linear dapat bekerja secara komputasi dengan transformasi pengaburan yang dipilih.\n",
    "\n",
    "Buat prosedur atau kelas yang menjalankan regresi linear dengan pengaburan data. Anda bisa menggunakan regresi linear yang tersedia di scikit-learn atau milik Anda sendiri.\n",
    "\n",
    "Jalankan regresi linear terhadap data asli dan data yang disamarkan, bandingkan nilai prediksi dan RMSE, serta nilai metrik $R^2$. Apakah ada perbedaan?"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Prosedur**\n",
    "\n",
    "- Buat matriks persegi $P$ dari angka acak.\n",
    "- Periksa apakah matriks tersebut *invertible*. Jika tidak, ulangi dari langkah pertama sampai kita mendapatkan matriks yang bisa dibalik (*invertible*).\n",
    "- <! ketik komentar Anda di sini !>\n",
    "- Gunakan $XP$ sebagai matriks fitur baru "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 91,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LinearRegression()"
      ]
     },
     "execution_count": 91,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "lr = LinearRegression()\n",
    "lr.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "metadata": {
    "collapsed": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>gender</th>\n",
       "      <th>age</th>\n",
       "      <th>income</th>\n",
       "      <th>family_members</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>-0.224252</td>\n",
       "      <td>-0.226030</td>\n",
       "      <td>-0.226096</td>\n",
       "      <td>-0.224962</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.169185</td>\n",
       "      <td>0.168677</td>\n",
       "      <td>0.168777</td>\n",
       "      <td>0.169944</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1.112581</td>\n",
       "      <td>1.110173</td>\n",
       "      <td>1.110247</td>\n",
       "      <td>1.111305</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>-2.146178</td>\n",
       "      <td>-2.144094</td>\n",
       "      <td>-2.144185</td>\n",
       "      <td>-2.146132</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.089330</td>\n",
       "      <td>0.089321</td>\n",
       "      <td>0.089478</td>\n",
       "      <td>0.089635</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     gender       age    income  family_members\n",
       "0 -0.224252 -0.226030 -0.226096       -0.224962\n",
       "1  0.169185  0.168677  0.168777        0.169944\n",
       "2  1.112581  1.110173  1.110247        1.111305\n",
       "3 -2.146178 -2.144094 -2.144185       -2.146132\n",
       "4  0.089330  0.089321  0.089478        0.089635"
      ]
     },
     "execution_count": 92,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 93,
   "metadata": {
    "collapsed": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>gender</th>\n",
       "      <th>age</th>\n",
       "      <th>income</th>\n",
       "      <th>family_members</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1.061615</td>\n",
       "      <td>1.061158</td>\n",
       "      <td>1.061257</td>\n",
       "      <td>1.061411</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>-1.429799</td>\n",
       "      <td>-1.429326</td>\n",
       "      <td>-1.429298</td>\n",
       "      <td>-1.428938</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.151730</td>\n",
       "      <td>0.152712</td>\n",
       "      <td>0.152426</td>\n",
       "      <td>0.151231</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>-0.821631</td>\n",
       "      <td>-0.821369</td>\n",
       "      <td>-0.821231</td>\n",
       "      <td>-0.821291</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>-0.425849</td>\n",
       "      <td>-0.423696</td>\n",
       "      <td>-0.423709</td>\n",
       "      <td>-0.425366</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     gender       age    income  family_members\n",
       "0  1.061615  1.061158  1.061257        1.061411\n",
       "1 -1.429799 -1.429326 -1.429298       -1.428938\n",
       "2  0.151730  0.152712  0.152426        0.151231\n",
       "3 -0.821631 -0.821369 -0.821231       -0.821291\n",
       "4 -0.425849 -0.423696 -0.423709       -0.425366"
      ]
     },
     "execution_count": 93,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_test.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div class=\"alert-info\">\n",
    "    \n",
    "\n",
    "<b>Note</b>\n",
    "\n",
    "Data sudah discaling"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "metadata": {
    "collapsed": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "F1: 0.53\n",
      "Accuraccy: 0.93\n",
      "Matriks Kesalahan\n",
      "[[0.88625226 0.        ]\n",
      " [0.07299458 0.04075316]]\n"
     ]
    }
   ],
   "source": [
    "# Predict data Training\n",
    "y_pred_train = lr.predict(X_train)\n",
    "eval_classifier(y_train, y_pred_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 96,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "F1: 0.64\n",
      "Accuraccy: 0.93\n",
      "Matriks Kesalahan\n",
      "[[0.8742268  0.        ]\n",
      " [0.06701031 0.05876289]]\n"
     ]
    }
   ],
   "source": [
    "# Predict data Testing\n",
    "y_pred_test = lr.predict(X_test)\n",
    "eval_classifier(y_test, y_pred_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Kesimpulan"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div class=\"alert-info\">\n",
    "    \n",
    "\n",
    "<b>Kesimpulan</b>\n",
    "\n",
    "- Hasil dari prediksi Data masking maupun data asli keduanya sama"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Apendiks \n",
    "\n",
    "## Apendiks A: Menulis Rumus di Jupyter Notebooks"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Anda dapat menulis rumus di Jupyter Notebook dalam bahasa *markup* yang disediakan oleh sistem penerbitan berkualitas tinggi bernama $\\LaTeX$ (diucapkan \"Lah-tech\"), dan rumus tersebut akan terlihat seperti rumus yang ada di buku teks.\n",
    "\n",
    "Untuk memasukkan rumus ke dalam teks, letakkan tanda dolar (\\\\$) sebelum dan sesudah teks rumus. $\\frac{1}{2} \\times \\frac{3}{2} = \\frac{3}{4}$ or $y = x^2, x \\ge 1$.\n",
    "\n",
    "Jika rumus harus ditulis dalam satu paragraf tersendiri, letakkan dua tanda dolar (\\\\$\\\\$) sebelum dan sesudah teks.\n",
    "\n",
    "$$\n",
    "\\bar{x} = \\frac{1}{n}\\sum_{i=1}^{n} x_i.\n",
    "$$\n",
    "\n",
    "Bahasa markup [LaTeX](https://en.wikipedia.org/wiki/LaTeX) sangat populer di kalangan orang-orang yang menggunakan rumus dalam artikel, buku, dan teks. Bahasa tesebut memang tampak rumit, tetapi dasarnya sebenarnya mudah. Periksa dua halaman [cheatsheet](http://tug.ctan.org/info/undergradmath/undergradmath.pdf) berikut untuk mempelajari cara membuat rumus yang paling umum."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Apendiks B: Properti Matriks"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Matriks memiliki banyak properti di Aljabar Linear. Beberapa di antaranya ada di daftar berikut yang dapat membantu pembuktian analitis dalam proyek ini."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<table>\n",
    "<tr>\n",
    "<td>Pendistribusian</td><td>$A(B+C)=AB+AC$</td>\n",
    "</tr>\n",
    "<tr>\n",
    "<td>Non-komutativitas</td><td>$AB \\neq BA$</td>\n",
    "</tr>\n",
    "<tr>\n",
    "<td>Properti asosiatif dari perkalian</td><td>$(AB)C = A(BC)$</td>\n",
    "</tr>\n",
    "<tr>\n",
    "<td>Properti identitas perkalian</td><td>$IA = AI = A$</td>\n",
    "</tr>\n",
    "<tr>\n",
    "<td></td><td>$A^{-1}A = AA^{-1} = I$\n",
    "</td>\n",
    "</tr>    \n",
    "<tr>\n",
    "<td></td><td>$(AB)^{-1} = B^{-1}A^{-1}$</td>\n",
    "</tr>    \n",
    "<tr>\n",
    "<td>Reversibilitas dari transposisi hasil kali matriks,</td><td>$(AB)^T = B^TA^T$</td>\n",
    "</tr>    \n",
    "</table>"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.12.3"
  },
  "toc": {
   "base_numbering": 1,
   "nav_menu": {},
   "number_sections": true,
   "sideBar": true,
   "skip_h1_title": true,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": false,
   "toc_position": {},
   "toc_section_display": true,
   "toc_window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}