ML-Sec.5.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Loading Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 145,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "from keras import optimizers\n",
    "from sklearn.preprocessing import MinMaxScaler\n",
    "from sklearn.metrics import accuracy_score\n",
    "from sklearn.metrics import precision_score\n",
    "from sklearn.metrics import recall_score\n",
    "from sklearn.metrics import f1_score\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "import math\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Dense, Dropout, Activation, Flatten\n",
    "from keras.utils import np_utils"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 94,
   "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>A1</th>\n",
       "      <th>A2</th>\n",
       "      <th>A3</th>\n",
       "      <th>A4</th>\n",
       "      <th>A5</th>\n",
       "      <th>A6</th>\n",
       "      <th>A7</th>\n",
       "      <th>A8</th>\n",
       "      <th>A9</th>\n",
       "      <th>A10</th>\n",
       "      <th>A11</th>\n",
       "      <th>A12</th>\n",
       "      <th>A13</th>\n",
       "      <th>A14</th>\n",
       "      <th>A15</th>\n",
       "      <th>A16</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>b</td>\n",
       "      <td>30.83</td>\n",
       "      <td>0</td>\n",
       "      <td>u</td>\n",
       "      <td>g</td>\n",
       "      <td>w</td>\n",
       "      <td>v</td>\n",
       "      <td>1.25</td>\n",
       "      <td>t</td>\n",
       "      <td>t</td>\n",
       "      <td>01</td>\n",
       "      <td>f</td>\n",
       "      <td>g</td>\n",
       "      <td>00202</td>\n",
       "      <td>0</td>\n",
       "      <td>+</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>a</td>\n",
       "      <td>58.67</td>\n",
       "      <td>4.46</td>\n",
       "      <td>u</td>\n",
       "      <td>g</td>\n",
       "      <td>q</td>\n",
       "      <td>h</td>\n",
       "      <td>3.04</td>\n",
       "      <td>t</td>\n",
       "      <td>t</td>\n",
       "      <td>06</td>\n",
       "      <td>f</td>\n",
       "      <td>g</td>\n",
       "      <td>00043</td>\n",
       "      <td>560</td>\n",
       "      <td>+</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>a</td>\n",
       "      <td>24.50</td>\n",
       "      <td>0.5</td>\n",
       "      <td>u</td>\n",
       "      <td>g</td>\n",
       "      <td>q</td>\n",
       "      <td>h</td>\n",
       "      <td>1.5</td>\n",
       "      <td>t</td>\n",
       "      <td>f</td>\n",
       "      <td>0</td>\n",
       "      <td>f</td>\n",
       "      <td>g</td>\n",
       "      <td>00280</td>\n",
       "      <td>824</td>\n",
       "      <td>+</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>b</td>\n",
       "      <td>27.83</td>\n",
       "      <td>1.54</td>\n",
       "      <td>u</td>\n",
       "      <td>g</td>\n",
       "      <td>w</td>\n",
       "      <td>v</td>\n",
       "      <td>3.75</td>\n",
       "      <td>t</td>\n",
       "      <td>t</td>\n",
       "      <td>05</td>\n",
       "      <td>t</td>\n",
       "      <td>g</td>\n",
       "      <td>00100</td>\n",
       "      <td>3</td>\n",
       "      <td>+</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>b</td>\n",
       "      <td>20.17</td>\n",
       "      <td>5.625</td>\n",
       "      <td>u</td>\n",
       "      <td>g</td>\n",
       "      <td>w</td>\n",
       "      <td>v</td>\n",
       "      <td>1.71</td>\n",
       "      <td>t</td>\n",
       "      <td>f</td>\n",
       "      <td>0</td>\n",
       "      <td>f</td>\n",
       "      <td>s</td>\n",
       "      <td>00120</td>\n",
       "      <td>0</td>\n",
       "      <td>+</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>685</th>\n",
       "      <td>b</td>\n",
       "      <td>21.08</td>\n",
       "      <td>10.085</td>\n",
       "      <td>y</td>\n",
       "      <td>p</td>\n",
       "      <td>e</td>\n",
       "      <td>h</td>\n",
       "      <td>1.25</td>\n",
       "      <td>f</td>\n",
       "      <td>f</td>\n",
       "      <td>0</td>\n",
       "      <td>f</td>\n",
       "      <td>g</td>\n",
       "      <td>00260</td>\n",
       "      <td>0</td>\n",
       "      <td>-</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>686</th>\n",
       "      <td>a</td>\n",
       "      <td>22.67</td>\n",
       "      <td>0.75</td>\n",
       "      <td>u</td>\n",
       "      <td>g</td>\n",
       "      <td>c</td>\n",
       "      <td>v</td>\n",
       "      <td>2</td>\n",
       "      <td>f</td>\n",
       "      <td>t</td>\n",
       "      <td>02</td>\n",
       "      <td>t</td>\n",
       "      <td>g</td>\n",
       "      <td>00200</td>\n",
       "      <td>394</td>\n",
       "      <td>-</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>687</th>\n",
       "      <td>a</td>\n",
       "      <td>25.25</td>\n",
       "      <td>13.5</td>\n",
       "      <td>y</td>\n",
       "      <td>p</td>\n",
       "      <td>ff</td>\n",
       "      <td>ff</td>\n",
       "      <td>2</td>\n",
       "      <td>f</td>\n",
       "      <td>t</td>\n",
       "      <td>01</td>\n",
       "      <td>t</td>\n",
       "      <td>g</td>\n",
       "      <td>00200</td>\n",
       "      <td>1</td>\n",
       "      <td>-</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>688</th>\n",
       "      <td>b</td>\n",
       "      <td>17.92</td>\n",
       "      <td>0.205</td>\n",
       "      <td>u</td>\n",
       "      <td>g</td>\n",
       "      <td>aa</td>\n",
       "      <td>v</td>\n",
       "      <td>0.04</td>\n",
       "      <td>f</td>\n",
       "      <td>f</td>\n",
       "      <td>0</td>\n",
       "      <td>f</td>\n",
       "      <td>g</td>\n",
       "      <td>00280</td>\n",
       "      <td>750</td>\n",
       "      <td>-</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>689</th>\n",
       "      <td>b</td>\n",
       "      <td>35.00</td>\n",
       "      <td>3.375</td>\n",
       "      <td>u</td>\n",
       "      <td>g</td>\n",
       "      <td>c</td>\n",
       "      <td>h</td>\n",
       "      <td>8.29</td>\n",
       "      <td>f</td>\n",
       "      <td>f</td>\n",
       "      <td>0</td>\n",
       "      <td>t</td>\n",
       "      <td>g</td>\n",
       "      <td>00000</td>\n",
       "      <td>0</td>\n",
       "      <td>-</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>690 rows × 16 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "    A1     A2      A3 A4 A5  A6  A7    A8 A9 A10 A11 A12 A13    A14  A15 A16\n",
       "0    b  30.83       0  u  g   w   v  1.25  t   t  01   f   g  00202    0   +\n",
       "1    a  58.67    4.46  u  g   q   h  3.04  t   t  06   f   g  00043  560   +\n",
       "2    a  24.50     0.5  u  g   q   h   1.5  t   f   0   f   g  00280  824   +\n",
       "3    b  27.83    1.54  u  g   w   v  3.75  t   t  05   t   g  00100    3   +\n",
       "4    b  20.17   5.625  u  g   w   v  1.71  t   f   0   f   s  00120    0   +\n",
       "..  ..    ...     ... .. ..  ..  ..   ... ..  ..  ..  ..  ..    ...  ...  ..\n",
       "685  b  21.08  10.085  y  p   e   h  1.25  f   f   0   f   g  00260    0   -\n",
       "686  a  22.67    0.75  u  g   c   v     2  f   t  02   t   g  00200  394   -\n",
       "687  a  25.25    13.5  y  p  ff  ff     2  f   t  01   t   g  00200    1   -\n",
       "688  b  17.92   0.205  u  g  aa   v  0.04  f   f   0   f   g  00280  750   -\n",
       "689  b  35.00   3.375  u  g   c   h  8.29  f   f   0   t   g  00000    0   -\n",
       "\n",
       "[690 rows x 16 columns]"
      ]
     },
     "execution_count": 94,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "path = r\"C:\\Users\\wayiqin\\Downloads\\machine_learning\\Assignment5\\crx.csv\" #change path for linux\n",
    "dataset = pd.read_csv(path,header=None,sep='\\t', delimiter=',',names=[\"data\"])\n",
    "dataset.head(10)\n",
    "splitdata = dataset[\"data\"].str.split(\",\")\n",
    "data = splitdata.to_list()\n",
    "names = [\"A1\", \"A2\", \"A3\", \"A4\", \"A5\", \"A6\", \"A7\", \"A8\", \"A9\", \"A10\", \"A11\", \"A12\", \"A13\", \"A14\", \"A15\", \"A16\"]\n",
    "new_df = pd.DataFrame(data, columns=names)\n",
    "new_df"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Checking for null values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "A1     False\n",
      "A2     False\n",
      "A3     False\n",
      "A4     False\n",
      "A5     False\n",
      "A6     False\n",
      "A7     False\n",
      "A8     False\n",
      "A9     False\n",
      "A10    False\n",
      "A11    False\n",
      "A12    False\n",
      "A13    False\n",
      "A14    False\n",
      "A15    False\n",
      "A16    False\n",
      "dtype: bool\n"
     ]
    }
   ],
   "source": [
    "print(new_df.isnull().any())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Checking for meaningless values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "b    468\n",
       "a    210\n",
       "?     12\n",
       "Name: A1, dtype: int64"
      ]
     },
     "execution_count": 78,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "new_df.A1.value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "metadata": {},
   "outputs": [],
   "source": [
    "new_df = new_df[new_df[new_df.columns] != '?']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "b    468\n",
       "a    210\n",
       "Name: A1, dtype: int64"
      ]
     },
     "execution_count": 80,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "new_df.A1.value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 96,
   "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>A1</th>\n",
       "      <th>A2</th>\n",
       "      <th>A3</th>\n",
       "      <th>A4</th>\n",
       "      <th>A5</th>\n",
       "      <th>A6</th>\n",
       "      <th>A7</th>\n",
       "      <th>A8</th>\n",
       "      <th>A9</th>\n",
       "      <th>A10</th>\n",
       "      <th>A11</th>\n",
       "      <th>A12</th>\n",
       "      <th>A13</th>\n",
       "      <th>A14</th>\n",
       "      <th>A15</th>\n",
       "      <th>A16</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>b</td>\n",
       "      <td>30.83</td>\n",
       "      <td>0</td>\n",
       "      <td>u</td>\n",
       "      <td>g</td>\n",
       "      <td>w</td>\n",
       "      <td>v</td>\n",
       "      <td>1.25</td>\n",
       "      <td>t</td>\n",
       "      <td>t</td>\n",
       "      <td>01</td>\n",
       "      <td>f</td>\n",
       "      <td>g</td>\n",
       "      <td>00202</td>\n",
       "      <td>0</td>\n",
       "      <td>+</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>a</td>\n",
       "      <td>58.67</td>\n",
       "      <td>4.46</td>\n",
       "      <td>u</td>\n",
       "      <td>g</td>\n",
       "      <td>q</td>\n",
       "      <td>h</td>\n",
       "      <td>3.04</td>\n",
       "      <td>t</td>\n",
       "      <td>t</td>\n",
       "      <td>06</td>\n",
       "      <td>f</td>\n",
       "      <td>g</td>\n",
       "      <td>00043</td>\n",
       "      <td>560</td>\n",
       "      <td>+</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>a</td>\n",
       "      <td>24.50</td>\n",
       "      <td>0.5</td>\n",
       "      <td>u</td>\n",
       "      <td>g</td>\n",
       "      <td>q</td>\n",
       "      <td>h</td>\n",
       "      <td>1.5</td>\n",
       "      <td>t</td>\n",
       "      <td>f</td>\n",
       "      <td>0</td>\n",
       "      <td>f</td>\n",
       "      <td>g</td>\n",
       "      <td>00280</td>\n",
       "      <td>824</td>\n",
       "      <td>+</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>b</td>\n",
       "      <td>27.83</td>\n",
       "      <td>1.54</td>\n",
       "      <td>u</td>\n",
       "      <td>g</td>\n",
       "      <td>w</td>\n",
       "      <td>v</td>\n",
       "      <td>3.75</td>\n",
       "      <td>t</td>\n",
       "      <td>t</td>\n",
       "      <td>05</td>\n",
       "      <td>t</td>\n",
       "      <td>g</td>\n",
       "      <td>00100</td>\n",
       "      <td>3</td>\n",
       "      <td>+</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>b</td>\n",
       "      <td>20.17</td>\n",
       "      <td>5.625</td>\n",
       "      <td>u</td>\n",
       "      <td>g</td>\n",
       "      <td>w</td>\n",
       "      <td>v</td>\n",
       "      <td>1.71</td>\n",
       "      <td>t</td>\n",
       "      <td>f</td>\n",
       "      <td>0</td>\n",
       "      <td>f</td>\n",
       "      <td>s</td>\n",
       "      <td>00120</td>\n",
       "      <td>0</td>\n",
       "      <td>+</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>685</th>\n",
       "      <td>b</td>\n",
       "      <td>21.08</td>\n",
       "      <td>10.085</td>\n",
       "      <td>y</td>\n",
       "      <td>p</td>\n",
       "      <td>e</td>\n",
       "      <td>h</td>\n",
       "      <td>1.25</td>\n",
       "      <td>f</td>\n",
       "      <td>f</td>\n",
       "      <td>0</td>\n",
       "      <td>f</td>\n",
       "      <td>g</td>\n",
       "      <td>00260</td>\n",
       "      <td>0</td>\n",
       "      <td>-</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>686</th>\n",
       "      <td>a</td>\n",
       "      <td>22.67</td>\n",
       "      <td>0.75</td>\n",
       "      <td>u</td>\n",
       "      <td>g</td>\n",
       "      <td>c</td>\n",
       "      <td>v</td>\n",
       "      <td>2</td>\n",
       "      <td>f</td>\n",
       "      <td>t</td>\n",
       "      <td>02</td>\n",
       "      <td>t</td>\n",
       "      <td>g</td>\n",
       "      <td>00200</td>\n",
       "      <td>394</td>\n",
       "      <td>-</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>687</th>\n",
       "      <td>a</td>\n",
       "      <td>25.25</td>\n",
       "      <td>13.5</td>\n",
       "      <td>y</td>\n",
       "      <td>p</td>\n",
       "      <td>ff</td>\n",
       "      <td>ff</td>\n",
       "      <td>2</td>\n",
       "      <td>f</td>\n",
       "      <td>t</td>\n",
       "      <td>01</td>\n",
       "      <td>t</td>\n",
       "      <td>g</td>\n",
       "      <td>00200</td>\n",
       "      <td>1</td>\n",
       "      <td>-</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>688</th>\n",
       "      <td>b</td>\n",
       "      <td>17.92</td>\n",
       "      <td>0.205</td>\n",
       "      <td>u</td>\n",
       "      <td>g</td>\n",
       "      <td>aa</td>\n",
       "      <td>v</td>\n",
       "      <td>0.04</td>\n",
       "      <td>f</td>\n",
       "      <td>f</td>\n",
       "      <td>0</td>\n",
       "      <td>f</td>\n",
       "      <td>g</td>\n",
       "      <td>00280</td>\n",
       "      <td>750</td>\n",
       "      <td>-</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>689</th>\n",
       "      <td>b</td>\n",
       "      <td>35.00</td>\n",
       "      <td>3.375</td>\n",
       "      <td>u</td>\n",
       "      <td>g</td>\n",
       "      <td>c</td>\n",
       "      <td>h</td>\n",
       "      <td>8.29</td>\n",
       "      <td>f</td>\n",
       "      <td>f</td>\n",
       "      <td>0</td>\n",
       "      <td>t</td>\n",
       "      <td>g</td>\n",
       "      <td>00000</td>\n",
       "      <td>0</td>\n",
       "      <td>-</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>653 rows × 16 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "    A1     A2      A3 A4 A5  A6  A7    A8 A9 A10 A11 A12 A13    A14  A15 A16\n",
       "0    b  30.83       0  u  g   w   v  1.25  t   t  01   f   g  00202    0   +\n",
       "1    a  58.67    4.46  u  g   q   h  3.04  t   t  06   f   g  00043  560   +\n",
       "2    a  24.50     0.5  u  g   q   h   1.5  t   f   0   f   g  00280  824   +\n",
       "3    b  27.83    1.54  u  g   w   v  3.75  t   t  05   t   g  00100    3   +\n",
       "4    b  20.17   5.625  u  g   w   v  1.71  t   f   0   f   s  00120    0   +\n",
       "..  ..    ...     ... .. ..  ..  ..   ... ..  ..  ..  ..  ..    ...  ...  ..\n",
       "685  b  21.08  10.085  y  p   e   h  1.25  f   f   0   f   g  00260    0   -\n",
       "686  a  22.67    0.75  u  g   c   v     2  f   t  02   t   g  00200  394   -\n",
       "687  a  25.25    13.5  y  p  ff  ff     2  f   t  01   t   g  00200    1   -\n",
       "688  b  17.92   0.205  u  g  aa   v  0.04  f   f   0   f   g  00280  750   -\n",
       "689  b  35.00   3.375  u  g   c   h  8.29  f   f   0   t   g  00000    0   -\n",
       "\n",
       "[653 rows x 16 columns]"
      ]
     },
     "execution_count": 96,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "new_df_nonan = new_df.dropna(how = 'any') \n",
    "new_df_nonan"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 97,
   "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>A1</th>\n",
       "      <th>A2</th>\n",
       "      <th>A3</th>\n",
       "      <th>A4</th>\n",
       "      <th>A5</th>\n",
       "      <th>A6</th>\n",
       "      <th>A7</th>\n",
       "      <th>A8</th>\n",
       "      <th>A9</th>\n",
       "      <th>A10</th>\n",
       "      <th>A11</th>\n",
       "      <th>A12</th>\n",
       "      <th>A13</th>\n",
       "      <th>A14</th>\n",
       "      <th>A15</th>\n",
       "      <th>A16</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>b</td>\n",
       "      <td>30.83</td>\n",
       "      <td>0</td>\n",
       "      <td>u</td>\n",
       "      <td>g</td>\n",
       "      <td>w</td>\n",
       "      <td>v</td>\n",
       "      <td>1.25</td>\n",
       "      <td>t</td>\n",
       "      <td>t</td>\n",
       "      <td>01</td>\n",
       "      <td>f</td>\n",
       "      <td>g</td>\n",
       "      <td>00202</td>\n",
       "      <td>0</td>\n",
       "      <td>+</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>a</td>\n",
       "      <td>58.67</td>\n",
       "      <td>4.46</td>\n",
       "      <td>u</td>\n",
       "      <td>g</td>\n",
       "      <td>q</td>\n",
       "      <td>h</td>\n",
       "      <td>3.04</td>\n",
       "      <td>t</td>\n",
       "      <td>t</td>\n",
       "      <td>06</td>\n",
       "      <td>f</td>\n",
       "      <td>g</td>\n",
       "      <td>00043</td>\n",
       "      <td>560</td>\n",
       "      <td>+</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>a</td>\n",
       "      <td>24.50</td>\n",
       "      <td>0.5</td>\n",
       "      <td>u</td>\n",
       "      <td>g</td>\n",
       "      <td>q</td>\n",
       "      <td>h</td>\n",
       "      <td>1.5</td>\n",
       "      <td>t</td>\n",
       "      <td>f</td>\n",
       "      <td>0</td>\n",
       "      <td>f</td>\n",
       "      <td>g</td>\n",
       "      <td>00280</td>\n",
       "      <td>824</td>\n",
       "      <td>+</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>b</td>\n",
       "      <td>27.83</td>\n",
       "      <td>1.54</td>\n",
       "      <td>u</td>\n",
       "      <td>g</td>\n",
       "      <td>w</td>\n",
       "      <td>v</td>\n",
       "      <td>3.75</td>\n",
       "      <td>t</td>\n",
       "      <td>t</td>\n",
       "      <td>05</td>\n",
       "      <td>t</td>\n",
       "      <td>g</td>\n",
       "      <td>00100</td>\n",
       "      <td>3</td>\n",
       "      <td>+</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>b</td>\n",
       "      <td>20.17</td>\n",
       "      <td>5.625</td>\n",
       "      <td>u</td>\n",
       "      <td>g</td>\n",
       "      <td>w</td>\n",
       "      <td>v</td>\n",
       "      <td>1.71</td>\n",
       "      <td>t</td>\n",
       "      <td>f</td>\n",
       "      <td>0</td>\n",
       "      <td>f</td>\n",
       "      <td>s</td>\n",
       "      <td>00120</td>\n",
       "      <td>0</td>\n",
       "      <td>+</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>648</th>\n",
       "      <td>b</td>\n",
       "      <td>21.08</td>\n",
       "      <td>10.085</td>\n",
       "      <td>y</td>\n",
       "      <td>p</td>\n",
       "      <td>e</td>\n",
       "      <td>h</td>\n",
       "      <td>1.25</td>\n",
       "      <td>f</td>\n",
       "      <td>f</td>\n",
       "      <td>0</td>\n",
       "      <td>f</td>\n",
       "      <td>g</td>\n",
       "      <td>00260</td>\n",
       "      <td>0</td>\n",
       "      <td>-</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>649</th>\n",
       "      <td>a</td>\n",
       "      <td>22.67</td>\n",
       "      <td>0.75</td>\n",
       "      <td>u</td>\n",
       "      <td>g</td>\n",
       "      <td>c</td>\n",
       "      <td>v</td>\n",
       "      <td>2</td>\n",
       "      <td>f</td>\n",
       "      <td>t</td>\n",
       "      <td>02</td>\n",
       "      <td>t</td>\n",
       "      <td>g</td>\n",
       "      <td>00200</td>\n",
       "      <td>394</td>\n",
       "      <td>-</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>650</th>\n",
       "      <td>a</td>\n",
       "      <td>25.25</td>\n",
       "      <td>13.5</td>\n",
       "      <td>y</td>\n",
       "      <td>p</td>\n",
       "      <td>ff</td>\n",
       "      <td>ff</td>\n",
       "      <td>2</td>\n",
       "      <td>f</td>\n",
       "      <td>t</td>\n",
       "      <td>01</td>\n",
       "      <td>t</td>\n",
       "      <td>g</td>\n",
       "      <td>00200</td>\n",
       "      <td>1</td>\n",
       "      <td>-</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>651</th>\n",
       "      <td>b</td>\n",
       "      <td>17.92</td>\n",
       "      <td>0.205</td>\n",
       "      <td>u</td>\n",
       "      <td>g</td>\n",
       "      <td>aa</td>\n",
       "      <td>v</td>\n",
       "      <td>0.04</td>\n",
       "      <td>f</td>\n",
       "      <td>f</td>\n",
       "      <td>0</td>\n",
       "      <td>f</td>\n",
       "      <td>g</td>\n",
       "      <td>00280</td>\n",
       "      <td>750</td>\n",
       "      <td>-</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>652</th>\n",
       "      <td>b</td>\n",
       "      <td>35.00</td>\n",
       "      <td>3.375</td>\n",
       "      <td>u</td>\n",
       "      <td>g</td>\n",
       "      <td>c</td>\n",
       "      <td>h</td>\n",
       "      <td>8.29</td>\n",
       "      <td>f</td>\n",
       "      <td>f</td>\n",
       "      <td>0</td>\n",
       "      <td>t</td>\n",
       "      <td>g</td>\n",
       "      <td>00000</td>\n",
       "      <td>0</td>\n",
       "      <td>-</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>653 rows × 16 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "    A1     A2      A3 A4 A5  A6  A7    A8 A9 A10 A11 A12 A13    A14  A15 A16\n",
       "0    b  30.83       0  u  g   w   v  1.25  t   t  01   f   g  00202    0   +\n",
       "1    a  58.67    4.46  u  g   q   h  3.04  t   t  06   f   g  00043  560   +\n",
       "2    a  24.50     0.5  u  g   q   h   1.5  t   f   0   f   g  00280  824   +\n",
       "3    b  27.83    1.54  u  g   w   v  3.75  t   t  05   t   g  00100    3   +\n",
       "4    b  20.17   5.625  u  g   w   v  1.71  t   f   0   f   s  00120    0   +\n",
       "..  ..    ...     ... .. ..  ..  ..   ... ..  ..  ..  ..  ..    ...  ...  ..\n",
       "648  b  21.08  10.085  y  p   e   h  1.25  f   f   0   f   g  00260    0   -\n",
       "649  a  22.67    0.75  u  g   c   v     2  f   t  02   t   g  00200  394   -\n",
       "650  a  25.25    13.5  y  p  ff  ff     2  f   t  01   t   g  00200    1   -\n",
       "651  b  17.92   0.205  u  g  aa   v  0.04  f   f   0   f   g  00280  750   -\n",
       "652  b  35.00   3.375  u  g   c   h  8.29  f   f   0   t   g  00000    0   -\n",
       "\n",
       "[653 rows x 16 columns]"
      ]
     },
     "execution_count": 97,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "new_df_nonan.reset_index(inplace=True,drop=True)\n",
    "new_df_nonan"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Convert categorical variables into numeric variables"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 99,
   "metadata": {},
   "outputs": [],
   "source": [
    "def A1_to_numeric(x):\n",
    "    if x=='a':\n",
    "        return 1\n",
    "    if x=='b':\n",
    "        return 2\n",
    "def A4_to_numeric(x):\n",
    "    if x=='u':\n",
    "        return 1\n",
    "    if x=='y':\n",
    "        return 2\n",
    "    if x=='l':\n",
    "        return 3\n",
    "    if x=='t':\n",
    "        return 4\n",
    "def A5_to_numeric(x):\n",
    "    if x=='g':\n",
    "        return 1\n",
    "    if x=='p':\n",
    "        return 2\n",
    "    if x=='gg':\n",
    "        return 3\n",
    "def A6_to_numeric(x):\n",
    "    if x=='c':\n",
    "        return 1\n",
    "    if x=='d':\n",
    "        return 2\n",
    "    if x=='cc':\n",
    "        return 3\n",
    "    if x=='i':\n",
    "        return 4\n",
    "    if x=='j':\n",
    "        return 5\n",
    "    if x=='k':\n",
    "        return 6\n",
    "    if x=='m':\n",
    "        return 7\n",
    "    if x=='r':\n",
    "        return 8\n",
    "    if x=='q':\n",
    "        return 9\n",
    "    if x=='w':\n",
    "        return 10\n",
    "    if x=='x':\n",
    "        return 11\n",
    "    if x=='e':\n",
    "        return 12\n",
    "    if x=='aa':\n",
    "        return 13\n",
    "    if x=='ff':\n",
    "        return 14\n",
    "def A7_to_numeric(x):\n",
    "    if x=='v':\n",
    "        return 1\n",
    "    if x=='h':\n",
    "        return 2\n",
    "    if x=='bb':\n",
    "        return 3\n",
    "    if x=='j':\n",
    "        return 4\n",
    "    if x=='n':\n",
    "        return 5\n",
    "    if x=='z':\n",
    "        return 6\n",
    "    if x=='dd':\n",
    "        return 7\n",
    "    if x=='ff':\n",
    "        return 8\n",
    "    if x=='o':\n",
    "        return 9\n",
    "def A9_10_12_to_numeric(x):\n",
    "    if x=='t':\n",
    "        return 1\n",
    "    if x=='f':\n",
    "        return 2\n",
    "def A13_to_numeric(x):\n",
    "    if x=='g':\n",
    "        return 1\n",
    "    if x=='p':\n",
    "        return 2\n",
    "    if x=='s':\n",
    "        return 3"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 100,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wayiqin\\Documents\\Anaconda\\envs\\pythonenv\\lib\\site-packages\\pandas\\core\\indexing.py:1047: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  self.obj[item_labels[indexer[info_axis]]] = value\n",
      "C:\\Users\\wayiqin\\Documents\\Anaconda\\envs\\pythonenv\\lib\\site-packages\\pandas\\core\\indexing.py:965: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  self.obj[item] = s\n"
     ]
    },
    {
     "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>A1</th>\n",
       "      <th>A2</th>\n",
       "      <th>A3</th>\n",
       "      <th>A4</th>\n",
       "      <th>A5</th>\n",
       "      <th>A6</th>\n",
       "      <th>A7</th>\n",
       "      <th>A8</th>\n",
       "      <th>A9</th>\n",
       "      <th>A10</th>\n",
       "      <th>A11</th>\n",
       "      <th>A12</th>\n",
       "      <th>A13</th>\n",
       "      <th>A14</th>\n",
       "      <th>A15</th>\n",
       "      <th>A16</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2</td>\n",
       "      <td>30.83</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>10</td>\n",
       "      <td>1</td>\n",
       "      <td>1.25</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>01</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>00202</td>\n",
       "      <td>0</td>\n",
       "      <td>+</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>58.67</td>\n",
       "      <td>4.46</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>9</td>\n",
       "      <td>2</td>\n",
       "      <td>3.04</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>06</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>00043</td>\n",
       "      <td>560</td>\n",
       "      <td>+</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>24.50</td>\n",
       "      <td>0.5</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>9</td>\n",
       "      <td>2</td>\n",
       "      <td>1.5</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>00280</td>\n",
       "      <td>824</td>\n",
       "      <td>+</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2</td>\n",
       "      <td>27.83</td>\n",
       "      <td>1.54</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>10</td>\n",
       "      <td>1</td>\n",
       "      <td>3.75</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>05</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>00100</td>\n",
       "      <td>3</td>\n",
       "      <td>+</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2</td>\n",
       "      <td>20.17</td>\n",
       "      <td>5.625</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>10</td>\n",
       "      <td>1</td>\n",
       "      <td>1.71</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>3</td>\n",
       "      <td>00120</td>\n",
       "      <td>0</td>\n",
       "      <td>+</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>648</th>\n",
       "      <td>2</td>\n",
       "      <td>21.08</td>\n",
       "      <td>10.085</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>12</td>\n",
       "      <td>2</td>\n",
       "      <td>1.25</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>00260</td>\n",
       "      <td>0</td>\n",
       "      <td>-</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>649</th>\n",
       "      <td>1</td>\n",
       "      <td>22.67</td>\n",
       "      <td>0.75</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>02</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>00200</td>\n",
       "      <td>394</td>\n",
       "      <td>-</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>650</th>\n",
       "      <td>1</td>\n",
       "      <td>25.25</td>\n",
       "      <td>13.5</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>14</td>\n",
       "      <td>8</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>01</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>00200</td>\n",
       "      <td>1</td>\n",
       "      <td>-</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>651</th>\n",
       "      <td>2</td>\n",
       "      <td>17.92</td>\n",
       "      <td>0.205</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>13</td>\n",
       "      <td>1</td>\n",
       "      <td>0.04</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>00280</td>\n",
       "      <td>750</td>\n",
       "      <td>-</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>652</th>\n",
       "      <td>2</td>\n",
       "      <td>35.00</td>\n",
       "      <td>3.375</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>8.29</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>00000</td>\n",
       "      <td>0</td>\n",
       "      <td>-</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>653 rows × 16 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "     A1     A2      A3  A4  A5  A6  A7    A8  A9  A10 A11  A12  A13    A14  \\\n",
       "0     2  30.83       0   1   1  10   1  1.25   1    1  01    2    1  00202   \n",
       "1     1  58.67    4.46   1   1   9   2  3.04   1    1  06    2    1  00043   \n",
       "2     1  24.50     0.5   1   1   9   2   1.5   1    2   0    2    1  00280   \n",
       "3     2  27.83    1.54   1   1  10   1  3.75   1    1  05    1    1  00100   \n",
       "4     2  20.17   5.625   1   1  10   1  1.71   1    2   0    2    3  00120   \n",
       "..   ..    ...     ...  ..  ..  ..  ..   ...  ..  ...  ..  ...  ...    ...   \n",
       "648   2  21.08  10.085   2   2  12   2  1.25   2    2   0    2    1  00260   \n",
       "649   1  22.67    0.75   1   1   1   1     2   2    1  02    1    1  00200   \n",
       "650   1  25.25    13.5   2   2  14   8     2   2    1  01    1    1  00200   \n",
       "651   2  17.92   0.205   1   1  13   1  0.04   2    2   0    2    1  00280   \n",
       "652   2  35.00   3.375   1   1   1   2  8.29   2    2   0    1    1  00000   \n",
       "\n",
       "     A15 A16  \n",
       "0      0   +  \n",
       "1    560   +  \n",
       "2    824   +  \n",
       "3      3   +  \n",
       "4      0   +  \n",
       "..   ...  ..  \n",
       "648    0   -  \n",
       "649  394   -  \n",
       "650    1   -  \n",
       "651  750   -  \n",
       "652    0   -  \n",
       "\n",
       "[653 rows x 16 columns]"
      ]
     },
     "execution_count": 100,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "new_df_nonan.loc[:,'A1']=new_df_nonan.loc[:,'A1'].apply(A1_to_numeric)\n",
    "new_df_nonan.loc[:,'A4']=new_df_nonan.loc[:,'A4'].apply(A4_to_numeric)\n",
    "new_df_nonan.loc[:,'A5']=new_df_nonan.loc[:,'A5'].apply(A5_to_numeric)\n",
    "new_df_nonan.loc[:,'A6']=new_df_nonan.loc[:,'A6'].apply(A6_to_numeric)\n",
    "new_df_nonan.loc[:,'A7']=new_df_nonan.loc[:,'A7'].apply(A7_to_numeric)\n",
    "new_df_nonan.loc[:,'A9']=new_df_nonan.loc[:,'A9'].apply(A9_10_12_to_numeric)\n",
    "new_df_nonan.loc[:,'A10']=new_df_nonan.loc[:,'A10'].apply(A9_10_12_to_numeric)\n",
    "new_df_nonan.loc[:,'A12']=new_df_nonan.loc[:,'A12'].apply(A9_10_12_to_numeric)\n",
    "new_df_nonan.loc[:,'A13']=new_df_nonan.loc[:,'A13'].apply(A13_to_numeric)\n",
    "new_df_nonan"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Define ANN Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 150,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_9\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "dense_25 (Dense)             (None, 30)                480       \n",
      "_________________________________________________________________\n",
      "dense_26 (Dense)             (None, 3)                 93        \n",
      "_________________________________________________________________\n",
      "dense_27 (Dense)             (None, 1)                 4         \n",
      "=================================================================\n",
      "Total params: 577\n",
      "Trainable params: 577\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "#Define the ANN\n",
    "ANNModel = Sequential() #Sequential ANN\n",
    "\n",
    "#Layers of the ANN (Dense in this case)\n",
    "ANNModel.add(Dense(30, input_dim=15, activation='relu'))\n",
    "ANNModel.add(Dense(3, input_dim=30, activation='relu'))\n",
    "ANNModel.add(Dense(1, input_dim=3, activation='sigmoid'))\n",
    "adam = optimizers.Adam(lr=0.01, beta_1=0.9, beta_2=0.999, amsgrad=False)\n",
    "ANNModel.compile(loss='binary_crossentropy',\n",
    "              optimizer= adam,\n",
    "              metrics=['accuracy'])\n",
    "ANNModel.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 116,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0      0\n",
      "1      0\n",
      "2      0\n",
      "3      0\n",
      "4      0\n",
      "      ..\n",
      "648    1\n",
      "649    1\n",
      "650    1\n",
      "651    1\n",
      "652    1\n",
      "Name: A16, Length: 653, dtype: int64\n"
     ]
    }
   ],
   "source": [
    "## split dataset into features and target\n",
    "variables = new_df_nonan.iloc[:,0:15]\n",
    "target = new_df_nonan.iloc[:,15]\n",
    "\n",
    "scalar = MinMaxScaler()\n",
    "scalar.fit(variables)\n",
    "variables = scalar.transform(variables)\n",
    "\n",
    "def target_to_numeric(x):\n",
    "    if x=='+':\n",
    "        return 0\n",
    "    if x=='-':\n",
    "        return 1\n",
    "\n",
    "## convert target into numeric\n",
    "target = target.apply(target_to_numeric)\n",
    "print(target)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 117,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(variables, target, test_size=0.33, random_state=42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 151,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/200\n",
      "437/437 [==============================] - 0s 208us/step - loss: 0.6125 - accuracy: 0.6133\n",
      "Epoch 2/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.4794 - accuracy: 0.7963\n",
      "Epoch 3/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.3766 - accuracy: 0.8764\n",
      "Epoch 4/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.3357 - accuracy: 0.8787\n",
      "Epoch 5/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.3407 - accuracy: 0.8719\n",
      "Epoch 6/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.3213 - accuracy: 0.8787\n",
      "Epoch 7/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.3049 - accuracy: 0.8902\n",
      "Epoch 8/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.3001 - accuracy: 0.8810\n",
      "Epoch 9/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2959 - accuracy: 0.8764\n",
      "Epoch 10/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2936 - accuracy: 0.8902\n",
      "Epoch 11/200\n",
      "437/437 [==============================] - 0s 43us/step - loss: 0.2970 - accuracy: 0.8856\n",
      "Epoch 12/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.3029 - accuracy: 0.8787\n",
      "Epoch 13/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2986 - accuracy: 0.8879\n",
      "Epoch 14/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2825 - accuracy: 0.8902\n",
      "Epoch 15/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2809 - accuracy: 0.8856\n",
      "Epoch 16/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2783 - accuracy: 0.8924\n",
      "Epoch 17/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2794 - accuracy: 0.8833\n",
      "Epoch 18/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2724 - accuracy: 0.8924\n",
      "Epoch 19/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2626 - accuracy: 0.8993\n",
      "Epoch 20/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2705 - accuracy: 0.8993\n",
      "Epoch 21/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2709 - accuracy: 0.8810\n",
      "Epoch 22/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2873 - accuracy: 0.8856\n",
      "Epoch 23/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2566 - accuracy: 0.9130\n",
      "Epoch 24/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2699 - accuracy: 0.8970\n",
      "Epoch 25/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2519 - accuracy: 0.9016\n",
      "Epoch 26/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2483 - accuracy: 0.9039\n",
      "Epoch 27/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2390 - accuracy: 0.9039\n",
      "Epoch 28/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2446 - accuracy: 0.9039\n",
      "Epoch 29/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2310 - accuracy: 0.9016\n",
      "Epoch 30/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2359 - accuracy: 0.9039\n",
      "Epoch 31/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2295 - accuracy: 0.8993\n",
      "Epoch 32/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2222 - accuracy: 0.9039\n",
      "Epoch 33/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2206 - accuracy: 0.9108\n",
      "Epoch 34/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2341 - accuracy: 0.9039\n",
      "Epoch 35/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2323 - accuracy: 0.9039\n",
      "Epoch 36/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2116 - accuracy: 0.9176\n",
      "Epoch 37/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2260 - accuracy: 0.9130\n",
      "Epoch 38/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2111 - accuracy: 0.9176\n",
      "Epoch 39/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2082 - accuracy: 0.9176\n",
      "Epoch 40/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2137 - accuracy: 0.9108\n",
      "Epoch 41/200\n",
      "437/437 [==============================] - 0s 37us/step - loss: 0.2148 - accuracy: 0.9130\n",
      "Epoch 42/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2043 - accuracy: 0.9108\n",
      "Epoch 43/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.2030 - accuracy: 0.9085\n",
      "Epoch 44/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2060 - accuracy: 0.9108\n",
      "Epoch 45/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2039 - accuracy: 0.9130\n",
      "Epoch 46/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.1937 - accuracy: 0.9199\n",
      "Epoch 47/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2060 - accuracy: 0.9153\n",
      "Epoch 48/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.1976 - accuracy: 0.9130\n",
      "Epoch 49/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1957 - accuracy: 0.9222\n",
      "Epoch 50/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.1847 - accuracy: 0.9130\n",
      "Epoch 51/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.1835 - accuracy: 0.9199\n",
      "Epoch 52/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.1867 - accuracy: 0.9291\n",
      "Epoch 53/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.1794 - accuracy: 0.9268\n",
      "Epoch 54/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.1848 - accuracy: 0.9222\n",
      "Epoch 55/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.1901 - accuracy: 0.9199\n",
      "Epoch 56/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.1821 - accuracy: 0.9222\n",
      "Epoch 57/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1901 - accuracy: 0.9314\n",
      "Epoch 58/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1977 - accuracy: 0.9245\n",
      "Epoch 59/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.1708 - accuracy: 0.9176\n",
      "Epoch 60/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1841 - accuracy: 0.9268\n",
      "Epoch 61/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.1804 - accuracy: 0.9291\n",
      "Epoch 62/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1630 - accuracy: 0.9336\n",
      "Epoch 63/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1568 - accuracy: 0.9382\n",
      "Epoch 64/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.1585 - accuracy: 0.9268\n",
      "Epoch 65/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.1594 - accuracy: 0.9291\n",
      "Epoch 66/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.1673 - accuracy: 0.9268\n",
      "Epoch 67/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.1611 - accuracy: 0.9405\n",
      "Epoch 68/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.1777 - accuracy: 0.9314\n",
      "Epoch 69/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.1695 - accuracy: 0.9336\n",
      "Epoch 70/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.1555 - accuracy: 0.9382\n",
      "Epoch 71/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.1656 - accuracy: 0.9336\n",
      "Epoch 72/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.1890 - accuracy: 0.9222\n",
      "Epoch 73/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.1613 - accuracy: 0.9382\n",
      "Epoch 74/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.1512 - accuracy: 0.9428\n",
      "Epoch 75/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.1490 - accuracy: 0.9382\n",
      "Epoch 76/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1463 - accuracy: 0.9428\n",
      "Epoch 77/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.1524 - accuracy: 0.9428\n",
      "Epoch 78/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1356 - accuracy: 0.9451\n",
      "Epoch 79/200\n",
      "437/437 [==============================] - 0s 46us/step - loss: 0.1343 - accuracy: 0.9451\n",
      "Epoch 80/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.1473 - accuracy: 0.9382\n",
      "Epoch 81/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.1400 - accuracy: 0.9451\n",
      "Epoch 82/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.1284 - accuracy: 0.9497\n",
      "Epoch 83/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1764 - accuracy: 0.9291\n",
      "Epoch 84/200\n",
      "437/437 [==============================] - ETA: 0s - loss: 0.2178 - accuracy: 0.90 - 0s 27us/step - loss: 0.1474 - accuracy: 0.9474\n",
      "Epoch 85/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.1506 - accuracy: 0.9497\n",
      "Epoch 86/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.1387 - accuracy: 0.9405\n",
      "Epoch 87/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.1323 - accuracy: 0.9451\n",
      "Epoch 88/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.1427 - accuracy: 0.9428\n",
      "Epoch 89/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.1237 - accuracy: 0.9634\n",
      "Epoch 90/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.1418 - accuracy: 0.9382\n",
      "Epoch 91/200\n",
      "437/437 [==============================] - 0s 23us/step - loss: 0.1289 - accuracy: 0.9474\n",
      "Epoch 92/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.1102 - accuracy: 0.9703\n",
      "Epoch 93/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1149 - accuracy: 0.9611\n",
      "Epoch 94/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.1243 - accuracy: 0.9519\n",
      "Epoch 95/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.1298 - accuracy: 0.9497\n",
      "Epoch 96/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.1221 - accuracy: 0.9542\n",
      "Epoch 97/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1169 - accuracy: 0.9565\n",
      "Epoch 98/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1131 - accuracy: 0.9588\n",
      "Epoch 99/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.1070 - accuracy: 0.9565\n",
      "Epoch 100/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.1093 - accuracy: 0.9565\n",
      "Epoch 101/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.1068 - accuracy: 0.9611\n",
      "Epoch 102/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0973 - accuracy: 0.9657\n",
      "Epoch 103/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1029 - accuracy: 0.9588\n",
      "Epoch 104/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0995 - accuracy: 0.9680\n",
      "Epoch 105/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.1110 - accuracy: 0.9611\n",
      "Epoch 106/200\n",
      "437/437 [==============================] - 0s 23us/step - loss: 0.1044 - accuracy: 0.9634\n",
      "Epoch 107/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.1005 - accuracy: 0.9680\n",
      "Epoch 108/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.1040 - accuracy: 0.9634\n",
      "Epoch 109/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.1000 - accuracy: 0.9634\n",
      "Epoch 110/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1045 - accuracy: 0.9474\n",
      "Epoch 111/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.0982 - accuracy: 0.9565\n",
      "Epoch 112/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.1382 - accuracy: 0.9542\n",
      "Epoch 113/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1053 - accuracy: 0.9657\n",
      "Epoch 114/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1066 - accuracy: 0.9611\n",
      "Epoch 115/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.1023 - accuracy: 0.9611\n",
      "Epoch 116/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.0965 - accuracy: 0.9588\n",
      "Epoch 117/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.1027 - accuracy: 0.9588\n",
      "Epoch 118/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0990 - accuracy: 0.9611\n",
      "Epoch 119/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0937 - accuracy: 0.9634\n",
      "Epoch 120/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0831 - accuracy: 0.9657\n",
      "Epoch 121/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0841 - accuracy: 0.9657\n",
      "Epoch 122/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.0866 - accuracy: 0.9680\n",
      "Epoch 123/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.1227 - accuracy: 0.9474\n",
      "Epoch 124/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2329 - accuracy: 0.9130\n",
      "Epoch 125/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1223 - accuracy: 0.9519\n",
      "Epoch 126/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.1027 - accuracy: 0.9519\n",
      "Epoch 127/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0917 - accuracy: 0.9657\n",
      "Epoch 128/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.0937 - accuracy: 0.9565\n",
      "Epoch 129/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0927 - accuracy: 0.9657\n",
      "Epoch 130/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0965 - accuracy: 0.9634\n",
      "Epoch 131/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0786 - accuracy: 0.9771\n",
      "Epoch 132/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0738 - accuracy: 0.9725\n",
      "Epoch 133/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.0710 - accuracy: 0.9840\n",
      "Epoch 134/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0748 - accuracy: 0.9703\n",
      "Epoch 135/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0673 - accuracy: 0.9794\n",
      "Epoch 136/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0759 - accuracy: 0.9771\n",
      "Epoch 137/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0692 - accuracy: 0.9680\n",
      "Epoch 138/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.0665 - accuracy: 0.9840\n",
      "Epoch 139/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0623 - accuracy: 0.9817\n",
      "Epoch 140/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0631 - accuracy: 0.9817\n",
      "Epoch 141/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0615 - accuracy: 0.9817\n",
      "Epoch 142/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0901 - accuracy: 0.9680\n",
      "Epoch 143/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1093 - accuracy: 0.9611\n",
      "Epoch 144/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.1405 - accuracy: 0.9474\n",
      "Epoch 145/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.1871 - accuracy: 0.9039\n",
      "Epoch 146/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1102 - accuracy: 0.9634\n",
      "Epoch 147/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0978 - accuracy: 0.9611\n",
      "Epoch 148/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0927 - accuracy: 0.9680\n",
      "Epoch 149/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0761 - accuracy: 0.9703\n",
      "Epoch 150/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0734 - accuracy: 0.9703\n",
      "Epoch 151/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0626 - accuracy: 0.9794\n",
      "Epoch 152/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0668 - accuracy: 0.9794\n",
      "Epoch 153/200\n",
      "437/437 [==============================] - 0s 43us/step - loss: 0.0624 - accuracy: 0.9817\n",
      "Epoch 154/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.0564 - accuracy: 0.9817\n",
      "Epoch 155/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0558 - accuracy: 0.9817\n",
      "Epoch 156/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0507 - accuracy: 0.9863\n",
      "Epoch 157/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0494 - accuracy: 0.9817\n",
      "Epoch 158/200\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "437/437 [==============================] - 0s 30us/step - loss: 0.0501 - accuracy: 0.9863\n",
      "Epoch 159/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0494 - accuracy: 0.9908\n",
      "Epoch 160/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0478 - accuracy: 0.9908\n",
      "Epoch 161/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0465 - accuracy: 0.9863\n",
      "Epoch 162/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0486 - accuracy: 0.9863\n",
      "Epoch 163/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.0501 - accuracy: 0.9931\n",
      "Epoch 164/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0473 - accuracy: 0.9863\n",
      "Epoch 165/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0614 - accuracy: 0.9794\n",
      "Epoch 166/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0619 - accuracy: 0.9794\n",
      "Epoch 167/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0501 - accuracy: 0.9886\n",
      "Epoch 168/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0441 - accuracy: 0.9886\n",
      "Epoch 169/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0434 - accuracy: 0.9908\n",
      "Epoch 170/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0419 - accuracy: 0.9931\n",
      "Epoch 171/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0454 - accuracy: 0.9840\n",
      "Epoch 172/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0424 - accuracy: 0.9863\n",
      "Epoch 173/200\n",
      "437/437 [==============================] - 0s 21us/step - loss: 0.0403 - accuracy: 0.9886\n",
      "Epoch 174/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.0430 - accuracy: 0.9886\n",
      "Epoch 175/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0483 - accuracy: 0.9886\n",
      "Epoch 176/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0422 - accuracy: 0.9863\n",
      "Epoch 177/200\n",
      "437/437 [==============================] - 0s 21us/step - loss: 0.0418 - accuracy: 0.9863\n",
      "Epoch 178/200\n",
      "437/437 [==============================] - 0s 23us/step - loss: 0.0405 - accuracy: 0.9908\n",
      "Epoch 179/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0494 - accuracy: 0.9817\n",
      "Epoch 180/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.0432 - accuracy: 0.9863\n",
      "Epoch 181/200\n",
      "437/437 [==============================] - 0s 23us/step - loss: 0.0582 - accuracy: 0.9840\n",
      "Epoch 182/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.0501 - accuracy: 0.9840\n",
      "Epoch 183/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0484 - accuracy: 0.9863\n",
      "Epoch 184/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0438 - accuracy: 0.9908\n",
      "Epoch 185/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.0530 - accuracy: 0.9817\n",
      "Epoch 186/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0504 - accuracy: 0.9886\n",
      "Epoch 187/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0506 - accuracy: 0.9817\n",
      "Epoch 188/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0553 - accuracy: 0.9771\n",
      "Epoch 189/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.0371 - accuracy: 0.9886\n",
      "Epoch 190/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.0332 - accuracy: 0.9931\n",
      "Epoch 191/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0392 - accuracy: 0.9908\n",
      "Epoch 192/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0344 - accuracy: 0.9931\n",
      "Epoch 193/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.0345 - accuracy: 0.9931\n",
      "Epoch 194/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0377 - accuracy: 0.9908\n",
      "Epoch 195/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0413 - accuracy: 0.9908\n",
      "Epoch 196/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0295 - accuracy: 0.9931\n",
      "Epoch 197/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0288 - accuracy: 0.9886\n",
      "Epoch 198/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0388 - accuracy: 0.9840\n",
      "Epoch 199/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.0429 - accuracy: 0.9840\n",
      "Epoch 200/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.0795 - accuracy: 0.9703\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.callbacks.callbacks.History at 0x135b2d45548>"
      ]
     },
     "execution_count": 151,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "nb_epoch = 200 \n",
    "batch_size = 32\n",
    "\n",
    "#Get the ANN parameters                   \n",
    "ANNModel.fit(X_train, y_train,\n",
    "              batch_size = batch_size,\n",
    "              epochs = nb_epoch)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 152,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Truth= 1 Predicted= 0\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 1 Predicted= 0\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 1 Predicted= 0\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 0 Predicted= 1\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 0 Predicted= 1\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 0 Predicted= 1\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 1 Predicted= 0\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 1 Predicted= 0\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 1 Predicted= 0\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 0 Predicted= 1\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 0 Predicted= 1\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 0 Predicted= 1\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 0 Predicted= 1\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 1 Predicted= 0\n",
      "Truth= 0 Predicted= 1\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 0 Predicted= 1\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 0 Predicted= 1\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 1 Predicted= 0\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 1 Predicted= 0\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 0 Predicted= 1\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 1 Predicted= 0\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 1 Predicted= 0\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 0 Predicted= 1\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 1 Predicted= 0\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 0 Predicted= 1\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 1 Predicted= 0\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 1 Predicted= 0\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 1 Predicted= 0\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 0 Predicted= 1\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 0 Predicted= 1\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 1 Predicted= 0\n",
      "Truth= 1 Predicted= 1\n",
      "Truth= 0 Predicted= 0\n",
      "Truth= 0 Predicted= 1\n",
      "Truth= 1 Predicted= 0\n",
      "Truth= 1 Predicted= 0\n",
      "Truth= 0 Predicted= 0\n"
     ]
    }
   ],
   "source": [
    "y_pred = ANNModel.predict_classes(X_test)\n",
    "y_test.reset_index(inplace=True,drop=True)\n",
    "# print(y_pred)\n",
    "# print(y_test)\n",
    "# show the inputs and predicted outputs\n",
    "predy = []\n",
    "testy = []\n",
    "for i in range(len(y_pred)):\n",
    "    predy.append(y_pred[i][0])\n",
    "    testy.append(y_test.loc[i])\n",
    "    print(\"Truth=\", y_test.loc[i], \"Predicted=\", y_pred[i][0])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Evaluate the result"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 153,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.842593\n",
      "Precision: 0.865546\n",
      "Recall: 0.851240\n",
      "F1 score: 0.858333\n"
     ]
    }
   ],
   "source": [
    "accuracy = accuracy_score(testy, predy)\n",
    "print('Accuracy: %f' % accuracy)\n",
    "# precision tp / (tp + fp)\n",
    "precision = precision_score(testy, predy)\n",
    "print('Precision: %f' % precision)\n",
    "# recall: tp / (tp + fn)\n",
    "recall = recall_score(testy, predy)\n",
    "print('Recall: %f' % recall)\n",
    "# f1: 2 tp / (2 tp + fp + fn)\n",
    "f1 = f1_score(testy, predy)\n",
    "print('F1 score: %f' % f1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Increase the Number of epochs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 154,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0652 - accuracy: 0.9748\n",
      "Epoch 2/500\n",
      "437/437 [==============================] - 0s 48us/step - loss: 0.0934 - accuracy: 0.9703\n",
      "Epoch 3/500\n",
      "437/437 [==============================] - 0s 39us/step - loss: 0.0495 - accuracy: 0.9794\n",
      "Epoch 4/500\n",
      "437/437 [==============================] - 0s 39us/step - loss: 0.0443 - accuracy: 0.9863\n",
      "Epoch 5/500\n",
      "437/437 [==============================] - 0s 39us/step - loss: 0.0370 - accuracy: 0.9840\n",
      "Epoch 6/500\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.0353 - accuracy: 0.9908\n",
      "Epoch 7/500\n",
      "437/437 [==============================] - 0s 39us/step - loss: 0.0474 - accuracy: 0.9771\n",
      "Epoch 8/500\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.0415 - accuracy: 0.9840\n",
      "Epoch 9/500\n",
      "437/437 [==============================] - 0s 37us/step - loss: 0.0492 - accuracy: 0.9840\n",
      "Epoch 10/500\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.0383 - accuracy: 0.9908\n",
      "Epoch 11/500\n",
      "437/437 [==============================] - 0s 37us/step - loss: 0.0309 - accuracy: 0.9908\n",
      "Epoch 12/500\n",
      "437/437 [==============================] - 0s 39us/step - loss: 0.0285 - accuracy: 0.9954\n",
      "Epoch 13/500\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.0255 - accuracy: 0.9954\n",
      "Epoch 14/500\n",
      "437/437 [==============================] - 0s 39us/step - loss: 0.0273 - accuracy: 0.9931\n",
      "Epoch 15/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0281 - accuracy: 0.9931\n",
      "Epoch 16/500\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.0290 - accuracy: 0.9908\n",
      "Epoch 17/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0243 - accuracy: 0.9954\n",
      "Epoch 18/500\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.0237 - accuracy: 0.9954\n",
      "Epoch 19/500\n",
      "437/437 [==============================] - 0s 37us/step - loss: 0.0231 - accuracy: 0.9931\n",
      "Epoch 20/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0270 - accuracy: 0.9908\n",
      "Epoch 21/500\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.0266 - accuracy: 0.9931\n",
      "Epoch 22/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0224 - accuracy: 0.9954\n",
      "Epoch 23/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0243 - accuracy: 0.9931\n",
      "Epoch 24/500\n",
      "437/437 [==============================] - 0s 37us/step - loss: 0.0266 - accuracy: 0.9908\n",
      "Epoch 25/500\n",
      "437/437 [==============================] - 0s 36us/step - loss: 0.0264 - accuracy: 0.9908\n",
      "Epoch 26/500\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.0270 - accuracy: 0.9908\n",
      "Epoch 27/500\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.0276 - accuracy: 0.9886\n",
      "Epoch 28/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0380 - accuracy: 0.9840\n",
      "Epoch 29/500\n",
      "437/437 [==============================] - 0s 37us/step - loss: 0.0335 - accuracy: 0.9931\n",
      "Epoch 30/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0412 - accuracy: 0.9840\n",
      "Epoch 31/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0428 - accuracy: 0.9817\n",
      "Epoch 32/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0330 - accuracy: 0.9886\n",
      "Epoch 33/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0232 - accuracy: 0.9931\n",
      "Epoch 34/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0278 - accuracy: 0.9908\n",
      "Epoch 35/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0349 - accuracy: 0.9863\n",
      "Epoch 36/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0560 - accuracy: 0.9794\n",
      "Epoch 37/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0758 - accuracy: 0.9725\n",
      "Epoch 38/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0541 - accuracy: 0.9840\n",
      "Epoch 39/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0285 - accuracy: 0.9908\n",
      "Epoch 40/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0338 - accuracy: 0.9863\n",
      "Epoch 41/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0281 - accuracy: 0.9908\n",
      "Epoch 42/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0257 - accuracy: 0.9863\n",
      "Epoch 43/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0220 - accuracy: 0.9931\n",
      "Epoch 44/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0199 - accuracy: 0.9954\n",
      "Epoch 45/500\n",
      "437/437 [==============================] - ETA: 0s - loss: 0.0169 - accuracy: 1.00 - 0s 27us/step - loss: 0.0188 - accuracy: 0.9954\n",
      "Epoch 46/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0190 - accuracy: 0.9954\n",
      "Epoch 47/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0182 - accuracy: 0.9954\n",
      "Epoch 48/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0182 - accuracy: 0.9954\n",
      "Epoch 49/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0191 - accuracy: 0.9931\n",
      "Epoch 50/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0187 - accuracy: 0.9954\n",
      "Epoch 51/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0165 - accuracy: 0.9954\n",
      "Epoch 52/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0208 - accuracy: 0.9886\n",
      "Epoch 53/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0225 - accuracy: 0.9908\n",
      "Epoch 54/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0177 - accuracy: 0.9954\n",
      "Epoch 55/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0182 - accuracy: 0.9954\n",
      "Epoch 56/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0179 - accuracy: 0.9954\n",
      "Epoch 57/500\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.0171 - accuracy: 0.9954\n",
      "Epoch 58/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0172 - accuracy: 0.9954\n",
      "Epoch 59/500\n",
      "437/437 [==============================] - 0s 39us/step - loss: 0.0161 - accuracy: 0.9954\n",
      "Epoch 60/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0182 - accuracy: 0.9954\n",
      "Epoch 61/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0187 - accuracy: 0.9931\n",
      "Epoch 62/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0169 - accuracy: 0.9931\n",
      "Epoch 63/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0165 - accuracy: 0.9954\n",
      "Epoch 64/500\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.0162 - accuracy: 0.9954\n",
      "Epoch 65/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0190 - accuracy: 0.9931\n",
      "Epoch 66/500\n",
      "437/437 [==============================] - 0s 50us/step - loss: 0.0150 - accuracy: 0.9954\n",
      "Epoch 67/500\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.0172 - accuracy: 0.9931\n",
      "Epoch 68/500\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.0178 - accuracy: 0.9931\n",
      "Epoch 69/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0168 - accuracy: 0.9954\n",
      "Epoch 70/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0200 - accuracy: 0.9908\n",
      "Epoch 71/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0196 - accuracy: 0.9954\n",
      "Epoch 72/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0223 - accuracy: 0.9908\n",
      "Epoch 73/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0371 - accuracy: 0.9863\n",
      "Epoch 74/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0365 - accuracy: 0.9886\n",
      "Epoch 75/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0254 - accuracy: 0.9931\n",
      "Epoch 76/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0164 - accuracy: 0.9954\n",
      "Epoch 77/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0170 - accuracy: 0.9954\n",
      "Epoch 78/500\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.0163 - accuracy: 0.9954\n",
      "Epoch 79/500\n",
      "437/437 [==============================] - 0s 36us/step - loss: 0.0173 - accuracy: 0.9954\n",
      "Epoch 80/500\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.0154 - accuracy: 0.9931\n",
      "Epoch 81/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0218 - accuracy: 0.9908\n",
      "Epoch 82/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0181 - accuracy: 0.9931\n",
      "Epoch 83/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0222 - accuracy: 0.9886\n",
      "Epoch 84/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0217 - accuracy: 0.9931\n",
      "Epoch 85/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0210 - accuracy: 0.9908\n",
      "Epoch 86/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0231 - accuracy: 0.9931\n",
      "Epoch 87/500\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.0187 - accuracy: 0.9954\n",
      "Epoch 88/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0148 - accuracy: 0.9954\n",
      "Epoch 89/500\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.0145 - accuracy: 0.9954\n",
      "Epoch 90/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0122 - accuracy: 0.9954\n",
      "Epoch 91/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0168 - accuracy: 0.9931\n",
      "Epoch 92/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0211 - accuracy: 0.9886\n",
      "Epoch 93/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0190 - accuracy: 0.9931\n",
      "Epoch 94/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0226 - accuracy: 0.9886\n",
      "Epoch 95/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0177 - accuracy: 0.9931\n",
      "Epoch 96/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0133 - accuracy: 0.9954\n",
      "Epoch 97/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0167 - accuracy: 0.9908\n",
      "Epoch 98/500\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.0174 - accuracy: 0.9931\n",
      "Epoch 99/500\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.0153 - accuracy: 0.9931\n",
      "Epoch 100/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0177 - accuracy: 0.9931\n",
      "Epoch 101/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0168 - accuracy: 0.9908\n",
      "Epoch 102/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0342 - accuracy: 0.9863\n",
      "Epoch 103/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0229 - accuracy: 0.9908\n",
      "Epoch 104/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0133 - accuracy: 0.9954\n",
      "Epoch 105/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0147 - accuracy: 0.9931\n",
      "Epoch 106/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0151 - accuracy: 0.9954\n",
      "Epoch 107/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0131 - accuracy: 0.9954\n",
      "Epoch 108/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0145 - accuracy: 0.9931\n",
      "Epoch 109/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0157 - accuracy: 0.9931\n",
      "Epoch 110/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0135 - accuracy: 0.9954\n",
      "Epoch 111/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0146 - accuracy: 0.9954\n",
      "Epoch 112/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0134 - accuracy: 0.9931\n",
      "Epoch 113/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0156 - accuracy: 0.9931\n",
      "Epoch 114/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0118 - accuracy: 0.9931\n",
      "Epoch 115/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0116 - accuracy: 0.9954\n",
      "Epoch 116/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0135 - accuracy: 0.9954\n",
      "Epoch 117/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0114 - accuracy: 0.9954\n",
      "Epoch 118/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0170 - accuracy: 0.9908\n",
      "Epoch 119/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0176 - accuracy: 0.9931\n",
      "Epoch 120/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0172 - accuracy: 0.9908\n",
      "Epoch 121/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0205 - accuracy: 0.9908\n",
      "Epoch 122/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0118 - accuracy: 0.9954\n",
      "Epoch 123/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0122 - accuracy: 0.9931\n",
      "Epoch 124/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0133 - accuracy: 0.9954\n",
      "Epoch 125/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0109 - accuracy: 0.9931\n",
      "Epoch 126/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0128 - accuracy: 0.9931\n",
      "Epoch 127/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0101 - accuracy: 0.9954\n",
      "Epoch 128/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0119 - accuracy: 0.9954\n",
      "Epoch 129/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0112 - accuracy: 0.9954\n",
      "Epoch 130/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0126 - accuracy: 0.9954\n",
      "Epoch 131/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0137 - accuracy: 0.9954\n",
      "Epoch 132/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0141 - accuracy: 0.9931\n",
      "Epoch 133/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0175 - accuracy: 0.9908\n",
      "Epoch 134/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0285 - accuracy: 0.9886\n",
      "Epoch 135/500\n",
      "437/437 [==============================] - 0s 39us/step - loss: 0.1015 - accuracy: 0.9725\n",
      "Epoch 136/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2722 - accuracy: 0.9405\n",
      "Epoch 137/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2267 - accuracy: 0.9405\n",
      "Epoch 138/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1122 - accuracy: 0.9588\n",
      "Epoch 139/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0680 - accuracy: 0.9725\n",
      "Epoch 140/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0854 - accuracy: 0.9794\n",
      "Epoch 141/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0636 - accuracy: 0.9748\n",
      "Epoch 142/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0534 - accuracy: 0.9863\n",
      "Epoch 143/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0259 - accuracy: 0.9886\n",
      "Epoch 144/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0163 - accuracy: 0.9954\n",
      "Epoch 145/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0159 - accuracy: 0.9954\n",
      "Epoch 146/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0142 - accuracy: 0.9954\n",
      "Epoch 147/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0129 - accuracy: 0.9954\n",
      "Epoch 148/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0123 - accuracy: 0.9931\n",
      "Epoch 149/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0144 - accuracy: 0.9977\n",
      "Epoch 150/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0146 - accuracy: 0.9931\n",
      "Epoch 151/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0122 - accuracy: 0.9977\n",
      "Epoch 152/500\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.0120 - accuracy: 0.9954\n",
      "Epoch 153/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0108 - accuracy: 0.9954\n",
      "Epoch 154/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0110 - accuracy: 0.9977\n",
      "Epoch 155/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0110 - accuracy: 0.9954\n",
      "Epoch 156/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0131 - accuracy: 0.9954\n",
      "Epoch 157/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0114 - accuracy: 0.9977\n",
      "Epoch 158/500\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.0099 - accuracy: 0.9977\n",
      "Epoch 159/500\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "437/437 [==============================] - 0s 32us/step - loss: 0.0115 - accuracy: 0.9954\n",
      "Epoch 160/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0110 - accuracy: 0.9954\n",
      "Epoch 161/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0112 - accuracy: 0.9954\n",
      "Epoch 162/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0099 - accuracy: 0.9954\n",
      "Epoch 163/500\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.0116 - accuracy: 0.9931\n",
      "Epoch 164/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0134 - accuracy: 0.9977\n",
      "Epoch 165/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0113 - accuracy: 0.9954\n",
      "Epoch 166/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0114 - accuracy: 0.9977\n",
      "Epoch 167/500\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.0115 - accuracy: 0.9931\n",
      "Epoch 168/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0125 - accuracy: 0.9954\n",
      "Epoch 169/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0101 - accuracy: 0.9954\n",
      "Epoch 170/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0106 - accuracy: 0.9954\n",
      "Epoch 171/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0101 - accuracy: 0.9954\n",
      "Epoch 172/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0099 - accuracy: 0.9977\n",
      "Epoch 173/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0093 - accuracy: 0.9954\n",
      "Epoch 174/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0096 - accuracy: 0.9977\n",
      "Epoch 175/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0095 - accuracy: 0.9954\n",
      "Epoch 176/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0091 - accuracy: 0.9954\n",
      "Epoch 177/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0136 - accuracy: 0.9931\n",
      "Epoch 178/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0120 - accuracy: 0.9931\n",
      "Epoch 179/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0145 - accuracy: 0.9954\n",
      "Epoch 180/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0120 - accuracy: 0.9977\n",
      "Epoch 181/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0092 - accuracy: 0.9977\n",
      "Epoch 182/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0128 - accuracy: 0.9954\n",
      "Epoch 183/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0123 - accuracy: 0.9977\n",
      "Epoch 184/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0092 - accuracy: 0.9954\n",
      "Epoch 185/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0101 - accuracy: 0.9977\n",
      "Epoch 186/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0086 - accuracy: 0.9977\n",
      "Epoch 187/500\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.0096 - accuracy: 0.9977\n",
      "Epoch 188/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0103 - accuracy: 0.9954\n",
      "Epoch 189/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0109 - accuracy: 0.9977\n",
      "Epoch 190/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0094 - accuracy: 0.9977\n",
      "Epoch 191/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0084 - accuracy: 0.9977\n",
      "Epoch 192/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0112 - accuracy: 0.9954\n",
      "Epoch 193/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0096 - accuracy: 0.9954\n",
      "Epoch 194/500\n",
      "437/437 [==============================] - 0s 37us/step - loss: 0.0096 - accuracy: 0.9977\n",
      "Epoch 195/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0091 - accuracy: 0.9954\n",
      "Epoch 196/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0088 - accuracy: 0.9977\n",
      "Epoch 197/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0092 - accuracy: 0.9954\n",
      "Epoch 198/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0134 - accuracy: 0.9977\n",
      "Epoch 199/500\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.0086 - accuracy: 0.9954\n",
      "Epoch 200/500\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.0083 - accuracy: 0.9977\n",
      "Epoch 201/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0090 - accuracy: 0.9977\n",
      "Epoch 202/500\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.0090 - accuracy: 0.9977\n",
      "Epoch 203/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0106 - accuracy: 0.9931\n",
      "Epoch 204/500\n",
      "437/437 [==============================] - ETA: 0s - loss: 7.1081e-04 - accuracy: 1.00 - 0s 32us/step - loss: 0.0098 - accuracy: 0.9977\n",
      "Epoch 205/500\n",
      "437/437 [==============================] - 0s 43us/step - loss: 0.0092 - accuracy: 0.9954\n",
      "Epoch 206/500\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.0102 - accuracy: 0.9977\n",
      "Epoch 207/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0100 - accuracy: 0.9954\n",
      "Epoch 208/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0090 - accuracy: 0.9954\n",
      "Epoch 209/500\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.0097 - accuracy: 0.9977\n",
      "Epoch 210/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0158 - accuracy: 0.9954\n",
      "Epoch 211/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0206 - accuracy: 0.9931\n",
      "Epoch 212/500\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.0172 - accuracy: 0.9931\n",
      "Epoch 213/500\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.0439 - accuracy: 0.9863\n",
      "Epoch 214/500\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.0602 - accuracy: 0.9817\n",
      "Epoch 215/500\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.0386 - accuracy: 0.9794\n",
      "Epoch 216/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.1072 - accuracy: 0.9748\n",
      "Epoch 217/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.1132 - accuracy: 0.9725\n",
      "Epoch 218/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1457 - accuracy: 0.9542\n",
      "Epoch 219/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.1738 - accuracy: 0.9474\n",
      "Epoch 220/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.1816 - accuracy: 0.9565\n",
      "Epoch 221/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.1086 - accuracy: 0.9634\n",
      "Epoch 222/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0702 - accuracy: 0.9725\n",
      "Epoch 223/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0674 - accuracy: 0.9748\n",
      "Epoch 224/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0340 - accuracy: 0.9817\n",
      "Epoch 225/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0292 - accuracy: 0.9840\n",
      "Epoch 226/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0287 - accuracy: 0.9863\n",
      "Epoch 227/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0324 - accuracy: 0.9840\n",
      "Epoch 228/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0160 - accuracy: 0.9931\n",
      "Epoch 229/500\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.0127 - accuracy: 0.9954\n",
      "Epoch 230/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0111 - accuracy: 0.9977\n",
      "Epoch 231/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0111 - accuracy: 0.9977\n",
      "Epoch 232/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0108 - accuracy: 0.9931\n",
      "Epoch 233/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0105 - accuracy: 0.9954\n",
      "Epoch 234/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0099 - accuracy: 0.9977\n",
      "Epoch 235/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0094 - accuracy: 0.9977\n",
      "Epoch 236/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0093 - accuracy: 0.9977\n",
      "Epoch 237/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0101 - accuracy: 0.9931\n",
      "Epoch 238/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0121 - accuracy: 0.9977\n",
      "Epoch 239/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0097 - accuracy: 0.9954\n",
      "Epoch 240/500\n",
      "437/437 [==============================] - 0s 23us/step - loss: 0.0092 - accuracy: 0.9954\n",
      "Epoch 241/500\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.0090 - accuracy: 0.9977\n",
      "Epoch 242/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0092 - accuracy: 0.9954\n",
      "Epoch 243/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0094 - accuracy: 0.9977\n",
      "Epoch 244/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0094 - accuracy: 0.9977\n",
      "Epoch 245/500\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.0090 - accuracy: 0.9977\n",
      "Epoch 246/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0102 - accuracy: 0.9954\n",
      "Epoch 247/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0077 - accuracy: 0.9977\n",
      "Epoch 248/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0100 - accuracy: 0.9977\n",
      "Epoch 249/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0081 - accuracy: 0.9977\n",
      "Epoch 250/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0086 - accuracy: 0.9977\n",
      "Epoch 251/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0114 - accuracy: 0.9954\n",
      "Epoch 252/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0108 - accuracy: 0.9954\n",
      "Epoch 253/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0086 - accuracy: 0.9977\n",
      "Epoch 254/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0091 - accuracy: 0.9977\n",
      "Epoch 255/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0084 - accuracy: 0.9977\n",
      "Epoch 256/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0084 - accuracy: 0.9977\n",
      "Epoch 257/500\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.0096 - accuracy: 0.9977\n",
      "Epoch 258/500\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.0087 - accuracy: 0.9977\n",
      "Epoch 259/500\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.0099 - accuracy: 0.9977\n",
      "Epoch 260/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0074 - accuracy: 0.9977\n",
      "Epoch 261/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0090 - accuracy: 0.9954\n",
      "Epoch 262/500\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.0080 - accuracy: 0.9977\n",
      "Epoch 263/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0086 - accuracy: 0.9977\n",
      "Epoch 264/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0087 - accuracy: 0.9977\n",
      "Epoch 265/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0073 - accuracy: 0.9977\n",
      "Epoch 266/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0074 - accuracy: 0.9977\n",
      "Epoch 267/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0079 - accuracy: 0.9954\n",
      "Epoch 268/500\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.0076 - accuracy: 0.9977\n",
      "Epoch 269/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0098 - accuracy: 0.9977\n",
      "Epoch 270/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0071 - accuracy: 0.9977\n",
      "Epoch 271/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0074 - accuracy: 0.9977\n",
      "Epoch 272/500\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.0083 - accuracy: 0.9977\n",
      "Epoch 273/500\n",
      "437/437 [==============================] - 0s 36us/step - loss: 0.0087 - accuracy: 0.9954\n",
      "Epoch 274/500\n",
      "437/437 [==============================] - 0s 37us/step - loss: 0.0078 - accuracy: 0.9977\n",
      "Epoch 275/500\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.0091 - accuracy: 0.9977\n",
      "Epoch 276/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0110 - accuracy: 0.9954\n",
      "Epoch 277/500\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.0094 - accuracy: 0.9977\n",
      "Epoch 278/500\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.0070 - accuracy: 0.9977\n",
      "Epoch 279/500\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.0083 - accuracy: 0.9977\n",
      "Epoch 280/500\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.0078 - accuracy: 0.9977\n",
      "Epoch 281/500\n",
      "437/437 [==============================] - 0s 36us/step - loss: 0.0081 - accuracy: 0.9954\n",
      "Epoch 282/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0098 - accuracy: 0.9977\n",
      "Epoch 283/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0096 - accuracy: 0.9954\n",
      "Epoch 284/500\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.0075 - accuracy: 0.9954\n",
      "Epoch 285/500\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.0090 - accuracy: 0.9977\n",
      "Epoch 286/500\n",
      "437/437 [==============================] - 0s 37us/step - loss: 0.0072 - accuracy: 0.9977\n",
      "Epoch 287/500\n",
      "437/437 [==============================] - 0s 37us/step - loss: 0.0108 - accuracy: 0.9954\n",
      "Epoch 288/500\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.0069 - accuracy: 0.9977\n",
      "Epoch 289/500\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.0075 - accuracy: 0.9954\n",
      "Epoch 290/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0090 - accuracy: 0.9954\n",
      "Epoch 291/500\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.0085 - accuracy: 0.9977\n",
      "Epoch 292/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0150 - accuracy: 0.9931\n",
      "Epoch 293/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0109 - accuracy: 0.9977\n",
      "Epoch 294/500\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.0078 - accuracy: 0.9977\n",
      "Epoch 295/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0081 - accuracy: 0.9977\n",
      "Epoch 296/500\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.0083 - accuracy: 0.9954\n",
      "Epoch 297/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0072 - accuracy: 0.9977\n",
      "Epoch 298/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0077 - accuracy: 0.9977\n",
      "Epoch 299/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0106 - accuracy: 0.9977\n",
      "Epoch 300/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0107 - accuracy: 0.9954\n",
      "Epoch 301/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0086 - accuracy: 0.9977\n",
      "Epoch 302/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0105 - accuracy: 0.9977\n",
      "Epoch 303/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0113 - accuracy: 0.9977\n",
      "Epoch 304/500\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.0092 - accuracy: 0.9954\n",
      "Epoch 305/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0073 - accuracy: 0.9977\n",
      "Epoch 306/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0084 - accuracy: 0.9954\n",
      "Epoch 307/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0062 - accuracy: 0.9977\n",
      "Epoch 308/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0079 - accuracy: 0.9977\n",
      "Epoch 309/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0068 - accuracy: 0.9977\n",
      "Epoch 310/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0055 - accuracy: 0.9977\n",
      "Epoch 311/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0089 - accuracy: 0.9977\n",
      "Epoch 312/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0067 - accuracy: 0.9977\n",
      "Epoch 313/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0063 - accuracy: 0.9977\n",
      "Epoch 314/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0073 - accuracy: 0.9977\n",
      "Epoch 315/500\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "437/437 [==============================] - 0s 30us/step - loss: 0.0072 - accuracy: 0.9977\n",
      "Epoch 316/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0071 - accuracy: 0.9977\n",
      "Epoch 317/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0065 - accuracy: 1.0000\n",
      "Epoch 318/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0076 - accuracy: 0.9977\n",
      "Epoch 319/500\n",
      "437/437 [==============================] - ETA: 0s - loss: 0.0393 - accuracy: 0.96 - 0s 30us/step - loss: 0.0124 - accuracy: 0.9908\n",
      "Epoch 320/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0094 - accuracy: 0.9977\n",
      "Epoch 321/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0073 - accuracy: 0.9977\n",
      "Epoch 322/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0082 - accuracy: 0.9954\n",
      "Epoch 323/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0062 - accuracy: 0.9977\n",
      "Epoch 324/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0080 - accuracy: 0.9977\n",
      "Epoch 325/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0072 - accuracy: 0.9977\n",
      "Epoch 326/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0070 - accuracy: 0.9977\n",
      "Epoch 327/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0065 - accuracy: 0.9977\n",
      "Epoch 328/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0093 - accuracy: 0.9977\n",
      "Epoch 329/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0118 - accuracy: 0.9954\n",
      "Epoch 330/500\n",
      "437/437 [==============================] - ETA: 0s - loss: 0.0722 - accuracy: 0.96 - 0s 27us/step - loss: 0.0174 - accuracy: 0.9954\n",
      "Epoch 331/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0072 - accuracy: 0.9977\n",
      "Epoch 332/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0123 - accuracy: 0.9931\n",
      "Epoch 333/500\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.0140 - accuracy: 0.9931\n",
      "Epoch 334/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0183 - accuracy: 0.9954\n",
      "Epoch 335/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0144 - accuracy: 0.9931\n",
      "Epoch 336/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0080 - accuracy: 0.9977\n",
      "Epoch 337/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0107 - accuracy: 0.9954\n",
      "Epoch 338/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0557 - accuracy: 0.9863\n",
      "Epoch 339/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.1635 - accuracy: 0.9588\n",
      "Epoch 340/500\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.2019 - accuracy: 0.9565\n",
      "Epoch 341/500\n",
      "437/437 [==============================] - 0s 41us/step - loss: 0.1393 - accuracy: 0.9519\n",
      "Epoch 342/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.1501 - accuracy: 0.9634\n",
      "Epoch 343/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0884 - accuracy: 0.9588\n",
      "Epoch 344/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0701 - accuracy: 0.9771\n",
      "Epoch 345/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0378 - accuracy: 0.9931\n",
      "Epoch 346/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0312 - accuracy: 0.9931\n",
      "Epoch 347/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0240 - accuracy: 0.9908\n",
      "Epoch 348/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0152 - accuracy: 0.9954\n",
      "Epoch 349/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0147 - accuracy: 0.9954\n",
      "Epoch 350/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0117 - accuracy: 0.9977\n",
      "Epoch 351/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0093 - accuracy: 0.9977\n",
      "Epoch 352/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0087 - accuracy: 0.9977\n",
      "Epoch 353/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0095 - accuracy: 0.9977\n",
      "Epoch 354/500\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.0101 - accuracy: 0.9954\n",
      "Epoch 355/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0084 - accuracy: 0.9954\n",
      "Epoch 356/500\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.0083 - accuracy: 0.9977\n",
      "Epoch 357/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0076 - accuracy: 0.9977\n",
      "Epoch 358/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0075 - accuracy: 0.9977\n",
      "Epoch 359/500\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.0075 - accuracy: 0.9977\n",
      "Epoch 360/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0084 - accuracy: 0.9954\n",
      "Epoch 361/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0076 - accuracy: 0.9977\n",
      "Epoch 362/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0070 - accuracy: 0.9977\n",
      "Epoch 363/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0076 - accuracy: 0.9977\n",
      "Epoch 364/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0066 - accuracy: 0.9977\n",
      "Epoch 365/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0084 - accuracy: 0.9977\n",
      "Epoch 366/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0070 - accuracy: 0.9954\n",
      "Epoch 367/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0067 - accuracy: 0.9977\n",
      "Epoch 368/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0072 - accuracy: 0.9977\n",
      "Epoch 369/500\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.0109 - accuracy: 0.9954\n",
      "Epoch 370/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0083 - accuracy: 0.9977\n",
      "Epoch 371/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0071 - accuracy: 0.9977\n",
      "Epoch 372/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0055 - accuracy: 1.0000\n",
      "Epoch 373/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0078 - accuracy: 0.9977\n",
      "Epoch 374/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0059 - accuracy: 1.0000\n",
      "Epoch 375/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0065 - accuracy: 0.9954\n",
      "Epoch 376/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0066 - accuracy: 0.9977\n",
      "Epoch 377/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0081 - accuracy: 0.9954\n",
      "Epoch 378/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0075 - accuracy: 0.9977\n",
      "Epoch 379/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0067 - accuracy: 0.9977\n",
      "Epoch 380/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0072 - accuracy: 0.9977\n",
      "Epoch 381/500\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.0080 - accuracy: 0.9977\n",
      "Epoch 382/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0064 - accuracy: 1.0000\n",
      "Epoch 383/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0105 - accuracy: 0.9977\n",
      "Epoch 384/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0091 - accuracy: 0.9977\n",
      "Epoch 385/500\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.0069 - accuracy: 0.9977\n",
      "Epoch 386/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0063 - accuracy: 0.9977\n",
      "Epoch 387/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0068 - accuracy: 0.9954\n",
      "Epoch 388/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0070 - accuracy: 0.9977\n",
      "Epoch 389/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0057 - accuracy: 0.9977\n",
      "Epoch 390/500\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.0060 - accuracy: 0.9977\n",
      "Epoch 391/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0073 - accuracy: 0.9977\n",
      "Epoch 392/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0058 - accuracy: 0.9977\n",
      "Epoch 393/500\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "437/437 [==============================] - 0s 27us/step - loss: 0.0056 - accuracy: 0.9977\n",
      "Epoch 394/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0075 - accuracy: 0.9977\n",
      "Epoch 395/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0070 - accuracy: 0.9977\n",
      "Epoch 396/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0061 - accuracy: 0.9977\n",
      "Epoch 397/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0071 - accuracy: 0.9954\n",
      "Epoch 398/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0118 - accuracy: 0.9977\n",
      "Epoch 399/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0157 - accuracy: 0.9954\n",
      "Epoch 400/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0059 - accuracy: 1.0000\n",
      "Epoch 401/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0070 - accuracy: 0.9977\n",
      "Epoch 402/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0061 - accuracy: 0.9977\n",
      "Epoch 403/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0067 - accuracy: 0.9977\n",
      "Epoch 404/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0070 - accuracy: 0.9977\n",
      "Epoch 405/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0066 - accuracy: 0.9977\n",
      "Epoch 406/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0063 - accuracy: 0.9977\n",
      "Epoch 407/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0056 - accuracy: 0.9977\n",
      "Epoch 408/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0075 - accuracy: 0.9954\n",
      "Epoch 409/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0071 - accuracy: 0.9954\n",
      "Epoch 410/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0060 - accuracy: 0.9977\n",
      "Epoch 411/500\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.0065 - accuracy: 0.9977\n",
      "Epoch 412/500\n",
      "437/437 [==============================] - 0s 43us/step - loss: 0.0082 - accuracy: 0.9977\n",
      "Epoch 413/500\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.0073 - accuracy: 0.9977\n",
      "Epoch 414/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0058 - accuracy: 0.9977\n",
      "Epoch 415/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0071 - accuracy: 0.9977\n",
      "Epoch 416/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0069 - accuracy: 0.9977\n",
      "Epoch 417/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0075 - accuracy: 0.9977\n",
      "Epoch 418/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0054 - accuracy: 0.9977\n",
      "Epoch 419/500\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.0069 - accuracy: 0.9977\n",
      "Epoch 420/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0060 - accuracy: 0.9977\n",
      "Epoch 421/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0052 - accuracy: 1.0000\n",
      "Epoch 422/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0053 - accuracy: 0.9977\n",
      "Epoch 423/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0065 - accuracy: 0.9977\n",
      "Epoch 424/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0072 - accuracy: 0.9977\n",
      "Epoch 425/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0060 - accuracy: 0.9977\n",
      "Epoch 426/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0075 - accuracy: 0.9954\n",
      "Epoch 427/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0071 - accuracy: 0.9977\n",
      "Epoch 428/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0057 - accuracy: 0.9977\n",
      "Epoch 429/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0078 - accuracy: 0.9977\n",
      "Epoch 430/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0052 - accuracy: 0.9977\n",
      "Epoch 431/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0067 - accuracy: 0.9977\n",
      "Epoch 432/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0051 - accuracy: 0.9977\n",
      "Epoch 433/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0045 - accuracy: 1.0000\n",
      "Epoch 434/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0051 - accuracy: 0.9977\n",
      "Epoch 435/500\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.0054 - accuracy: 0.9954\n",
      "Epoch 436/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0058 - accuracy: 0.9954\n",
      "Epoch 437/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0061 - accuracy: 0.9977\n",
      "Epoch 438/500\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.0059 - accuracy: 0.9977\n",
      "Epoch 439/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0053 - accuracy: 0.9977\n",
      "Epoch 440/500\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.0073 - accuracy: 0.9977\n",
      "Epoch 441/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0045 - accuracy: 1.0000\n",
      "Epoch 442/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0064 - accuracy: 0.9977\n",
      "Epoch 443/500\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.0059 - accuracy: 0.9977\n",
      "Epoch 444/500\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.0065 - accuracy: 0.9954\n",
      "Epoch 445/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0051 - accuracy: 0.9977\n",
      "Epoch 446/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0065 - accuracy: 0.9954\n",
      "Epoch 447/500\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.0050 - accuracy: 0.9977\n",
      "Epoch 448/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0050 - accuracy: 0.9977\n",
      "Epoch 449/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0102 - accuracy: 0.9954\n",
      "Epoch 450/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0887 - accuracy: 0.9863\n",
      "Epoch 451/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1983 - accuracy: 0.9565\n",
      "Epoch 452/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2221 - accuracy: 0.9565\n",
      "Epoch 453/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1387 - accuracy: 0.9451\n",
      "Epoch 454/500\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.2028 - accuracy: 0.9497\n",
      "Epoch 455/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1222 - accuracy: 0.9680\n",
      "Epoch 456/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0440 - accuracy: 0.9886\n",
      "Epoch 457/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0250 - accuracy: 0.9886\n",
      "Epoch 458/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0179 - accuracy: 0.9954\n",
      "Epoch 459/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0129 - accuracy: 0.9977\n",
      "Epoch 460/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0102 - accuracy: 0.9977\n",
      "Epoch 461/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0077 - accuracy: 0.9977\n",
      "Epoch 462/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0089 - accuracy: 0.9977\n",
      "Epoch 463/500\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.0068 - accuracy: 1.0000\n",
      "Epoch 464/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0100 - accuracy: 0.9977\n",
      "Epoch 465/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0069 - accuracy: 0.9977\n",
      "Epoch 466/500\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.0073 - accuracy: 1.0000\n",
      "Epoch 467/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0067 - accuracy: 1.0000\n",
      "Epoch 468/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0068 - accuracy: 0.9977\n",
      "Epoch 469/500\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.0059 - accuracy: 0.9977\n",
      "Epoch 470/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0058 - accuracy: 1.0000\n",
      "Epoch 471/500\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.0060 - accuracy: 0.9977\n",
      "Epoch 472/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0075 - accuracy: 0.9977\n",
      "Epoch 473/500\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.0066 - accuracy: 0.9977\n",
      "Epoch 474/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0078 - accuracy: 0.9977\n",
      "Epoch 475/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0063 - accuracy: 0.9977\n",
      "Epoch 476/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0071 - accuracy: 0.9977\n",
      "Epoch 477/500\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.0072 - accuracy: 0.9977\n",
      "Epoch 478/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0062 - accuracy: 0.9977\n",
      "Epoch 479/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0059 - accuracy: 0.9977\n",
      "Epoch 480/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0079 - accuracy: 0.9977\n",
      "Epoch 481/500\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.0058 - accuracy: 1.0000\n",
      "Epoch 482/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0066 - accuracy: 0.9977\n",
      "Epoch 483/500\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.0074 - accuracy: 0.9977\n",
      "Epoch 484/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0075 - accuracy: 0.9977\n",
      "Epoch 485/500\n",
      "437/437 [==============================] - 0s 37us/step - loss: 0.0083 - accuracy: 0.9954\n",
      "Epoch 486/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0060 - accuracy: 0.9977\n",
      "Epoch 487/500\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.0059 - accuracy: 0.9977\n",
      "Epoch 488/500\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.0081 - accuracy: 0.9954\n",
      "Epoch 489/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0052 - accuracy: 0.9977\n",
      "Epoch 490/500\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0084 - accuracy: 0.9954\n",
      "Epoch 491/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0053 - accuracy: 0.9977\n",
      "Epoch 492/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0072 - accuracy: 0.9954\n",
      "Epoch 493/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0050 - accuracy: 1.0000\n",
      "Epoch 494/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0070 - accuracy: 0.9977\n",
      "Epoch 495/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0068 - accuracy: 0.9954\n",
      "Epoch 496/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0061 - accuracy: 0.9977\n",
      "Epoch 497/500\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.0058 - accuracy: 0.9977\n",
      "Epoch 498/500\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0075 - accuracy: 0.9977\n",
      "Epoch 499/500\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.0039 - accuracy: 1.0000\n",
      "Epoch 500/500\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0078 - accuracy: 0.9954\n"
     ]
    }
   ],
   "source": [
    "nb_epoch = 500 \n",
    "batch_size = 32\n",
    "\n",
    "#Get the ANN parameters                   \n",
    "ANNModel.fit(X_train, y_train,\n",
    "              batch_size = batch_size,\n",
    "              epochs = nb_epoch)\n",
    "\n",
    "y_pred = ANNModel.predict_classes(X_test)\n",
    "\n",
    "# print(y_pred)\n",
    "# print(y_test)\n",
    "\n",
    "predy = []\n",
    "testy = []\n",
    "for i in range(len(y_pred)):\n",
    "    predy.append(y_pred[i][0])\n",
    "    testy.append(y_test.loc[i])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 155,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.856481\n",
      "Precision: 0.875000\n",
      "Recall: 0.867769\n",
      "F1 score: 0.871369\n"
     ]
    }
   ],
   "source": [
    "accuracy = accuracy_score(testy, predy)\n",
    "print('Accuracy: %f' % accuracy)\n",
    "# precision tp / (tp + fp)\n",
    "precision = precision_score(testy, predy)\n",
    "print('Precision: %f' % precision)\n",
    "# recall: tp / (tp + fn)\n",
    "recall = recall_score(testy, predy)\n",
    "print('Recall: %f' % recall)\n",
    "# f1: 2 tp / (2 tp + fp + fn)\n",
    "f1 = f1_score(testy, predy)\n",
    "print('F1 score: %f' % f1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Increase the Learning rate"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 156,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_10\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "dense_28 (Dense)             (None, 30)                480       \n",
      "_________________________________________________________________\n",
      "dense_29 (Dense)             (None, 3)                 93        \n",
      "_________________________________________________________________\n",
      "dense_30 (Dense)             (None, 1)                 4         \n",
      "=================================================================\n",
      "Total params: 577\n",
      "Trainable params: 577\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "#Define the ANN\n",
    "ANNModel = Sequential() #Sequential ANN\n",
    "\n",
    "#Layers of the ANN (Dense in this case)\n",
    "ANNModel.add(Dense(30, input_dim=15, activation='relu'))\n",
    "ANNModel.add(Dense(3, input_dim=30, activation='relu'))\n",
    "ANNModel.add(Dense(1, input_dim=3, activation='sigmoid'))\n",
    "adam = optimizers.Adam(lr=0.1, beta_1=0.9, beta_2=0.999, amsgrad=False)\n",
    "ANNModel.compile(loss='binary_crossentropy',\n",
    "              optimizer= adam,\n",
    "              metrics=['accuracy'])\n",
    "ANNModel.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 157,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/200\n",
      "437/437 [==============================] - 0s 199us/step - loss: 0.5888 - accuracy: 0.6064\n",
      "Epoch 2/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.4013 - accuracy: 0.8741\n",
      "Epoch 3/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.3708 - accuracy: 0.8764\n",
      "Epoch 4/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.3472 - accuracy: 0.8719\n",
      "Epoch 5/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.3371 - accuracy: 0.8764\n",
      "Epoch 6/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.3798 - accuracy: 0.8535\n",
      "Epoch 7/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.3244 - accuracy: 0.8879\n",
      "Epoch 8/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.3277 - accuracy: 0.8856\n",
      "Epoch 9/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.3248 - accuracy: 0.8924\n",
      "Epoch 10/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.3195 - accuracy: 0.8741\n",
      "Epoch 11/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.3049 - accuracy: 0.8787\n",
      "Epoch 12/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.3005 - accuracy: 0.8810\n",
      "Epoch 13/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.3198 - accuracy: 0.8741\n",
      "Epoch 14/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.3273 - accuracy: 0.8719\n",
      "Epoch 15/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.3106 - accuracy: 0.9016\n",
      "Epoch 16/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.3184 - accuracy: 0.8650\n",
      "Epoch 17/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.3084 - accuracy: 0.8856\n",
      "Epoch 18/200\n",
      "437/437 [==============================] - 0s 52us/step - loss: 0.3201 - accuracy: 0.8764\n",
      "Epoch 19/200\n",
      "437/437 [==============================] - 0s 37us/step - loss: 0.2871 - accuracy: 0.8924\n",
      "Epoch 20/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2944 - accuracy: 0.8902\n",
      "Epoch 21/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.3334 - accuracy: 0.8787\n",
      "Epoch 22/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.3069 - accuracy: 0.8902\n",
      "Epoch 23/200\n",
      "437/437 [==============================] - 0s 39us/step - loss: 0.2801 - accuracy: 0.8970\n",
      "Epoch 24/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2838 - accuracy: 0.8970\n",
      "Epoch 25/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.3133 - accuracy: 0.8764\n",
      "Epoch 26/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2839 - accuracy: 0.8993\n",
      "Epoch 27/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2686 - accuracy: 0.8993\n",
      "Epoch 28/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2910 - accuracy: 0.8924\n",
      "Epoch 29/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2652 - accuracy: 0.9016\n",
      "Epoch 30/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2754 - accuracy: 0.8902\n",
      "Epoch 31/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2838 - accuracy: 0.8879\n",
      "Epoch 32/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2584 - accuracy: 0.9085\n",
      "Epoch 33/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2605 - accuracy: 0.8947\n",
      "Epoch 34/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2613 - accuracy: 0.9039\n",
      "Epoch 35/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2448 - accuracy: 0.9222\n",
      "Epoch 36/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2374 - accuracy: 0.9153\n",
      "Epoch 37/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2370 - accuracy: 0.9199\n",
      "Epoch 38/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2547 - accuracy: 0.9039\n",
      "Epoch 39/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2471 - accuracy: 0.9153\n",
      "Epoch 40/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2364 - accuracy: 0.9176\n",
      "Epoch 41/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2484 - accuracy: 0.9130\n",
      "Epoch 42/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2350 - accuracy: 0.9085\n",
      "Epoch 43/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2307 - accuracy: 0.9245\n",
      "Epoch 44/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2246 - accuracy: 0.9199\n",
      "Epoch 45/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2408 - accuracy: 0.9108\n",
      "Epoch 46/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2414 - accuracy: 0.9108\n",
      "Epoch 47/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.2291 - accuracy: 0.9245\n",
      "Epoch 48/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2311 - accuracy: 0.9085\n",
      "Epoch 49/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2325 - accuracy: 0.9039\n",
      "Epoch 50/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2352 - accuracy: 0.9130\n",
      "Epoch 51/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2381 - accuracy: 0.9062\n",
      "Epoch 52/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2220 - accuracy: 0.9176\n",
      "Epoch 53/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2504 - accuracy: 0.9085\n",
      "Epoch 54/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2172 - accuracy: 0.9222\n",
      "Epoch 55/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.2155 - accuracy: 0.9153\n",
      "Epoch 56/200\n",
      "437/437 [==============================] - ETA: 0s - loss: 0.0610 - accuracy: 1.00 - 0s 37us/step - loss: 0.2306 - accuracy: 0.9176\n",
      "Epoch 57/200\n",
      "437/437 [==============================] - 0s 37us/step - loss: 0.2231 - accuracy: 0.9245\n",
      "Epoch 58/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2802 - accuracy: 0.8856\n",
      "Epoch 59/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.2771 - accuracy: 0.8833\n",
      "Epoch 60/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2649 - accuracy: 0.9039\n",
      "Epoch 61/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.2481 - accuracy: 0.9039\n",
      "Epoch 62/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2265 - accuracy: 0.9153\n",
      "Epoch 63/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2385 - accuracy: 0.9085\n",
      "Epoch 64/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2212 - accuracy: 0.9108\n",
      "Epoch 65/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2163 - accuracy: 0.9199\n",
      "Epoch 66/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2140 - accuracy: 0.9108\n",
      "Epoch 67/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.3109 - accuracy: 0.8764\n",
      "Epoch 68/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.3328 - accuracy: 0.8924\n",
      "Epoch 69/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2261 - accuracy: 0.9062\n",
      "Epoch 70/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2251 - accuracy: 0.9199\n",
      "Epoch 71/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.2372 - accuracy: 0.9245\n",
      "Epoch 72/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2273 - accuracy: 0.9176\n",
      "Epoch 73/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2151 - accuracy: 0.9291\n",
      "Epoch 74/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2229 - accuracy: 0.9130\n",
      "Epoch 75/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2433 - accuracy: 0.9039\n",
      "Epoch 76/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2410 - accuracy: 0.9153\n",
      "Epoch 77/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2324 - accuracy: 0.9153\n",
      "Epoch 78/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2147 - accuracy: 0.9245\n",
      "Epoch 79/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2077 - accuracy: 0.9268\n",
      "Epoch 80/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2080 - accuracy: 0.9245\n",
      "Epoch 81/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2095 - accuracy: 0.9268\n",
      "Epoch 82/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2060 - accuracy: 0.9268\n",
      "Epoch 83/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2116 - accuracy: 0.9336\n",
      "Epoch 84/200\n",
      "437/437 [==============================] - 0s 37us/step - loss: 0.2028 - accuracy: 0.9245\n",
      "Epoch 85/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.1999 - accuracy: 0.9291\n",
      "Epoch 86/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.2099 - accuracy: 0.9199\n",
      "Epoch 87/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2273 - accuracy: 0.9245\n",
      "Epoch 88/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2695 - accuracy: 0.8993\n",
      "Epoch 89/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2251 - accuracy: 0.9199\n",
      "Epoch 90/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.2217 - accuracy: 0.9130\n",
      "Epoch 91/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2051 - accuracy: 0.9336\n",
      "Epoch 92/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2007 - accuracy: 0.9291\n",
      "Epoch 93/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.1982 - accuracy: 0.9268\n",
      "Epoch 94/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.1948 - accuracy: 0.9336\n",
      "Epoch 95/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.2043 - accuracy: 0.9291\n",
      "Epoch 96/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.1949 - accuracy: 0.9336\n",
      "Epoch 97/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2348 - accuracy: 0.9245\n",
      "Epoch 98/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2490 - accuracy: 0.9062\n",
      "Epoch 99/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2815 - accuracy: 0.9039\n",
      "Epoch 100/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2391 - accuracy: 0.9108\n",
      "Epoch 101/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.2319 - accuracy: 0.9245\n",
      "Epoch 102/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2351 - accuracy: 0.9268\n",
      "Epoch 103/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.2223 - accuracy: 0.9245\n",
      "Epoch 104/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2332 - accuracy: 0.9176\n",
      "Epoch 105/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2655 - accuracy: 0.9085\n",
      "Epoch 106/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2357 - accuracy: 0.9176\n",
      "Epoch 107/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2165 - accuracy: 0.9291\n",
      "Epoch 108/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2038 - accuracy: 0.9314\n",
      "Epoch 109/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2011 - accuracy: 0.9291\n",
      "Epoch 110/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2036 - accuracy: 0.9268\n",
      "Epoch 111/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2022 - accuracy: 0.9336\n",
      "Epoch 112/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2051 - accuracy: 0.9336\n",
      "Epoch 113/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2056 - accuracy: 0.9199\n",
      "Epoch 114/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2124 - accuracy: 0.9222\n",
      "Epoch 115/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.1967 - accuracy: 0.9291\n",
      "Epoch 116/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.1945 - accuracy: 0.9268\n",
      "Epoch 117/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.1893 - accuracy: 0.9359\n",
      "Epoch 118/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.1959 - accuracy: 0.9382\n",
      "Epoch 119/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.1946 - accuracy: 0.9314\n",
      "Epoch 120/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.1970 - accuracy: 0.9245\n",
      "Epoch 121/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.1885 - accuracy: 0.9405\n",
      "Epoch 122/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.2050 - accuracy: 0.9176\n",
      "Epoch 123/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.1990 - accuracy: 0.9291\n",
      "Epoch 124/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.2372 - accuracy: 0.8993\n",
      "Epoch 125/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2167 - accuracy: 0.9199\n",
      "Epoch 126/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2179 - accuracy: 0.9039\n",
      "Epoch 127/200\n",
      "437/437 [==============================] - 0s 23us/step - loss: 0.2130 - accuracy: 0.9268\n",
      "Epoch 128/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.2486 - accuracy: 0.8947\n",
      "Epoch 129/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2702 - accuracy: 0.8810\n",
      "Epoch 130/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2252 - accuracy: 0.9153\n",
      "Epoch 131/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.2295 - accuracy: 0.9153\n",
      "Epoch 132/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.2139 - accuracy: 0.9268\n",
      "Epoch 133/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.2094 - accuracy: 0.9176\n",
      "Epoch 134/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.2221 - accuracy: 0.9291\n",
      "Epoch 135/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.2064 - accuracy: 0.9314\n",
      "Epoch 136/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.2239 - accuracy: 0.9291\n",
      "Epoch 137/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2087 - accuracy: 0.9405\n",
      "Epoch 138/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2174 - accuracy: 0.9268\n",
      "Epoch 139/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.2206 - accuracy: 0.9268\n",
      "Epoch 140/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.2464 - accuracy: 0.9130\n",
      "Epoch 141/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.2617 - accuracy: 0.9016\n",
      "Epoch 142/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2529 - accuracy: 0.9153\n",
      "Epoch 143/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.3172 - accuracy: 0.8879\n",
      "Epoch 144/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.3153 - accuracy: 0.8879\n",
      "Epoch 145/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.2621 - accuracy: 0.8993\n",
      "Epoch 146/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2724 - accuracy: 0.9062\n",
      "Epoch 147/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2563 - accuracy: 0.9085\n",
      "Epoch 148/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.2561 - accuracy: 0.9108\n",
      "Epoch 149/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.2540 - accuracy: 0.9062\n",
      "Epoch 150/200\n",
      "437/437 [==============================] - ETA: 0s - loss: 0.0932 - accuracy: 1.00 - 0s 25us/step - loss: 0.2227 - accuracy: 0.9199\n",
      "Epoch 151/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.2172 - accuracy: 0.9199\n",
      "Epoch 152/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.2168 - accuracy: 0.9268\n",
      "Epoch 153/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.2197 - accuracy: 0.9176\n",
      "Epoch 154/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.2257 - accuracy: 0.9199\n",
      "Epoch 155/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.2140 - accuracy: 0.9268\n",
      "Epoch 156/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2135 - accuracy: 0.9268\n",
      "Epoch 157/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2144 - accuracy: 0.9176\n",
      "Epoch 158/200\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "437/437 [==============================] - 0s 34us/step - loss: 0.2213 - accuracy: 0.9199\n",
      "Epoch 159/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2280 - accuracy: 0.9245\n",
      "Epoch 160/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.2198 - accuracy: 0.9222\n",
      "Epoch 161/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.1971 - accuracy: 0.9268\n",
      "Epoch 162/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2057 - accuracy: 0.9359\n",
      "Epoch 163/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1946 - accuracy: 0.9291\n",
      "Epoch 164/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2087 - accuracy: 0.9291\n",
      "Epoch 165/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.2042 - accuracy: 0.9222\n",
      "Epoch 166/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2046 - accuracy: 0.9245\n",
      "Epoch 167/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.2038 - accuracy: 0.9268\n",
      "Epoch 168/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.2112 - accuracy: 0.9268\n",
      "Epoch 169/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.1971 - accuracy: 0.9359\n",
      "Epoch 170/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.2057 - accuracy: 0.9245\n",
      "Epoch 171/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.2046 - accuracy: 0.9291\n",
      "Epoch 172/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.1953 - accuracy: 0.9291\n",
      "Epoch 173/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.1958 - accuracy: 0.9336\n",
      "Epoch 174/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1908 - accuracy: 0.9314\n",
      "Epoch 175/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1872 - accuracy: 0.9359\n",
      "Epoch 176/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.1903 - accuracy: 0.9291\n",
      "Epoch 177/200\n",
      "437/437 [==============================] - 0s 37us/step - loss: 0.1863 - accuracy: 0.9336\n",
      "Epoch 178/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.1953 - accuracy: 0.9336\n",
      "Epoch 179/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.1877 - accuracy: 0.9314\n",
      "Epoch 180/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.1995 - accuracy: 0.9314\n",
      "Epoch 181/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2006 - accuracy: 0.9336\n",
      "Epoch 182/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.1997 - accuracy: 0.9336\n",
      "Epoch 183/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2170 - accuracy: 0.9245\n",
      "Epoch 184/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2061 - accuracy: 0.9268\n",
      "Epoch 185/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.1961 - accuracy: 0.9291\n",
      "Epoch 186/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.2135 - accuracy: 0.9268\n",
      "Epoch 187/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.2059 - accuracy: 0.9268\n",
      "Epoch 188/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.1964 - accuracy: 0.9336\n",
      "Epoch 189/200\n",
      "437/437 [==============================] - 0s 37us/step - loss: 0.1965 - accuracy: 0.9268\n",
      "Epoch 190/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.1847 - accuracy: 0.9336\n",
      "Epoch 191/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.1984 - accuracy: 0.9222\n",
      "Epoch 192/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2144 - accuracy: 0.9199\n",
      "Epoch 193/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.1840 - accuracy: 0.9268\n",
      "Epoch 194/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.1862 - accuracy: 0.9336\n",
      "Epoch 195/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.1906 - accuracy: 0.9405\n",
      "Epoch 196/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.1855 - accuracy: 0.9336\n",
      "Epoch 197/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.1830 - accuracy: 0.9336\n",
      "Epoch 198/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.1834 - accuracy: 0.9359\n",
      "Epoch 199/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1879 - accuracy: 0.9359\n",
      "Epoch 200/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.1887 - accuracy: 0.9336\n"
     ]
    }
   ],
   "source": [
    "nb_epoch = 200 \n",
    "batch_size = 32\n",
    "\n",
    "#Get the ANN parameters                   \n",
    "ANNModel.fit(X_train, y_train,\n",
    "              batch_size = batch_size,\n",
    "              epochs = nb_epoch)\n",
    "\n",
    "y_pred = ANNModel.predict_classes(X_test)\n",
    "\n",
    "# print(y_pred)\n",
    "# print(y_test)\n",
    "\n",
    "predy = []\n",
    "testy = []\n",
    "for i in range(len(y_pred)):\n",
    "    predy.append(y_pred[i][0])\n",
    "    testy.append(y_test.loc[i])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 158,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.856481\n",
      "Precision: 0.881356\n",
      "Recall: 0.859504\n",
      "F1 score: 0.870293\n"
     ]
    }
   ],
   "source": [
    "accuracy = accuracy_score(testy, predy)\n",
    "print('Accuracy: %f' % accuracy)\n",
    "# precision tp / (tp + fp)\n",
    "precision = precision_score(testy, predy)\n",
    "print('Precision: %f' % precision)\n",
    "# recall: tp / (tp + fn)\n",
    "recall = recall_score(testy, predy)\n",
    "print('Recall: %f' % recall)\n",
    "# f1: 2 tp / (2 tp + fp + fn)\n",
    "f1 = f1_score(testy, predy)\n",
    "print('F1 score: %f' % f1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Change the numer of layer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 159,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_11\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "dense_31 (Dense)             (None, 30)                480       \n",
      "_________________________________________________________________\n",
      "dense_32 (Dense)             (None, 20)                620       \n",
      "_________________________________________________________________\n",
      "dense_33 (Dense)             (None, 10)                210       \n",
      "_________________________________________________________________\n",
      "dense_34 (Dense)             (None, 3)                 33        \n",
      "_________________________________________________________________\n",
      "dense_35 (Dense)             (None, 1)                 4         \n",
      "=================================================================\n",
      "Total params: 1,347\n",
      "Trainable params: 1,347\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "#Define the ANN\n",
    "ANNModel = Sequential() #Sequential ANN\n",
    "\n",
    "#Layers of the ANN (Dense in this case)\n",
    "ANNModel.add(Dense(30, input_dim=15, activation='relu'))\n",
    "ANNModel.add(Dense(20, input_dim=30, activation='relu'))\n",
    "ANNModel.add(Dense(10, input_dim=20, activation='relu'))\n",
    "ANNModel.add(Dense(3, input_dim=10, activation='relu'))\n",
    "ANNModel.add(Dense(1, input_dim=3, activation='sigmoid'))\n",
    "adam = optimizers.Adam(lr=0.01, beta_1=0.9, beta_2=0.999, amsgrad=False)\n",
    "ANNModel.compile(loss='binary_crossentropy',\n",
    "              optimizer= adam,\n",
    "              metrics=['accuracy'])\n",
    "ANNModel.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 160,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/200\n",
      "437/437 [==============================] - 0s 274us/step - loss: 0.5302 - accuracy: 0.6636\n",
      "Epoch 2/200\n",
      "437/437 [==============================] - 0s 39us/step - loss: 0.4509 - accuracy: 0.8787\n",
      "Epoch 3/200\n",
      "437/437 [==============================] - 0s 41us/step - loss: 0.4337 - accuracy: 0.8741\n",
      "Epoch 4/200\n",
      "437/437 [==============================] - 0s 39us/step - loss: 0.4187 - accuracy: 0.8741\n",
      "Epoch 5/200\n",
      "437/437 [==============================] - 0s 36us/step - loss: 0.3913 - accuracy: 0.8764\n",
      "Epoch 6/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.3794 - accuracy: 0.8787\n",
      "Epoch 7/200\n",
      "437/437 [==============================] - 0s 37us/step - loss: 0.3633 - accuracy: 0.8787\n",
      "Epoch 8/200\n",
      "437/437 [==============================] - 0s 37us/step - loss: 0.3643 - accuracy: 0.8856\n",
      "Epoch 9/200\n",
      "437/437 [==============================] - 0s 37us/step - loss: 0.3530 - accuracy: 0.8764\n",
      "Epoch 10/200\n",
      "437/437 [==============================] - 0s 36us/step - loss: 0.3506 - accuracy: 0.8810\n",
      "Epoch 11/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.3300 - accuracy: 0.8902\n",
      "Epoch 12/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.3238 - accuracy: 0.8902\n",
      "Epoch 13/200\n",
      "437/437 [==============================] - 0s 39us/step - loss: 0.3198 - accuracy: 0.8833\n",
      "Epoch 14/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.3209 - accuracy: 0.8947\n",
      "Epoch 15/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.3136 - accuracy: 0.8856\n",
      "Epoch 16/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2972 - accuracy: 0.8947\n",
      "Epoch 17/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.2980 - accuracy: 0.8924\n",
      "Epoch 18/200\n",
      "437/437 [==============================] - 0s 37us/step - loss: 0.2858 - accuracy: 0.9039\n",
      "Epoch 19/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.2986 - accuracy: 0.8833\n",
      "Epoch 20/200\n",
      "437/437 [==============================] - 0s 37us/step - loss: 0.2923 - accuracy: 0.8924\n",
      "Epoch 21/200\n",
      "437/437 [==============================] - 0s 37us/step - loss: 0.2747 - accuracy: 0.8924\n",
      "Epoch 22/200\n",
      "437/437 [==============================] - 0s 37us/step - loss: 0.2852 - accuracy: 0.8787\n",
      "Epoch 23/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.2873 - accuracy: 0.8947\n",
      "Epoch 24/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.3024 - accuracy: 0.8947\n",
      "Epoch 25/200\n",
      "437/437 [==============================] - 0s 41us/step - loss: 0.2955 - accuracy: 0.8787\n",
      "Epoch 26/200\n",
      "437/437 [==============================] - 0s 37us/step - loss: 0.2877 - accuracy: 0.8879\n",
      "Epoch 27/200\n",
      "437/437 [==============================] - 0s 37us/step - loss: 0.3144 - accuracy: 0.8856\n",
      "Epoch 28/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.2617 - accuracy: 0.8970\n",
      "Epoch 29/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.2523 - accuracy: 0.9016\n",
      "Epoch 30/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.2562 - accuracy: 0.8947\n",
      "Epoch 31/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2503 - accuracy: 0.8947\n",
      "Epoch 32/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2451 - accuracy: 0.9062\n",
      "Epoch 33/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2604 - accuracy: 0.8993\n",
      "Epoch 34/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.2528 - accuracy: 0.9062\n",
      "Epoch 35/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.2255 - accuracy: 0.9108\n",
      "Epoch 36/200\n",
      "437/437 [==============================] - 0s 37us/step - loss: 0.2519 - accuracy: 0.8970\n",
      "Epoch 37/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2621 - accuracy: 0.8970\n",
      "Epoch 38/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.2410 - accuracy: 0.9153\n",
      "Epoch 39/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2174 - accuracy: 0.9130\n",
      "Epoch 40/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2081 - accuracy: 0.9222\n",
      "Epoch 41/200\n",
      "437/437 [==============================] - 0s 37us/step - loss: 0.2118 - accuracy: 0.9199\n",
      "Epoch 42/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.2284 - accuracy: 0.9108\n",
      "Epoch 43/200\n",
      "437/437 [==============================] - 0s 39us/step - loss: 0.2304 - accuracy: 0.9130\n",
      "Epoch 44/200\n",
      "437/437 [==============================] - 0s 37us/step - loss: 0.2264 - accuracy: 0.9199\n",
      "Epoch 45/200\n",
      "437/437 [==============================] - 0s 37us/step - loss: 0.2183 - accuracy: 0.9176\n",
      "Epoch 46/200\n",
      "437/437 [==============================] - 0s 46us/step - loss: 0.2050 - accuracy: 0.9359\n",
      "Epoch 47/200\n",
      "437/437 [==============================] - 0s 41us/step - loss: 0.1988 - accuracy: 0.9291\n",
      "Epoch 48/200\n",
      "437/437 [==============================] - 0s 37us/step - loss: 0.2057 - accuracy: 0.9291\n",
      "Epoch 49/200\n",
      "437/437 [==============================] - 0s 37us/step - loss: 0.2171 - accuracy: 0.9291\n",
      "Epoch 50/200\n",
      "437/437 [==============================] - 0s 37us/step - loss: 0.2146 - accuracy: 0.9222\n",
      "Epoch 51/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.2101 - accuracy: 0.9199\n",
      "Epoch 52/200\n",
      "437/437 [==============================] - 0s 37us/step - loss: 0.2089 - accuracy: 0.9268\n",
      "Epoch 53/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.1878 - accuracy: 0.9268\n",
      "Epoch 54/200\n",
      "437/437 [==============================] - 0s 37us/step - loss: 0.1837 - accuracy: 0.9405\n",
      "Epoch 55/200\n",
      "437/437 [==============================] - 0s 37us/step - loss: 0.1782 - accuracy: 0.9428\n",
      "Epoch 56/200\n",
      "437/437 [==============================] - 0s 37us/step - loss: 0.1825 - accuracy: 0.9382\n",
      "Epoch 57/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.1649 - accuracy: 0.9451\n",
      "Epoch 58/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.1837 - accuracy: 0.9405\n",
      "Epoch 59/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.1962 - accuracy: 0.9314\n",
      "Epoch 60/200\n",
      "437/437 [==============================] - 0s 37us/step - loss: 0.1947 - accuracy: 0.9382\n",
      "Epoch 61/200\n",
      "437/437 [==============================] - 0s 41us/step - loss: 0.1912 - accuracy: 0.9359\n",
      "Epoch 62/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2048 - accuracy: 0.9336\n",
      "Epoch 63/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.1876 - accuracy: 0.9474\n",
      "Epoch 64/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.1940 - accuracy: 0.9314\n",
      "Epoch 65/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.1947 - accuracy: 0.9405\n",
      "Epoch 66/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.1669 - accuracy: 0.9497\n",
      "Epoch 67/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.1734 - accuracy: 0.9336\n",
      "Epoch 68/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1522 - accuracy: 0.9497\n",
      "Epoch 69/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1622 - accuracy: 0.9474\n",
      "Epoch 70/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.1437 - accuracy: 0.9474\n",
      "Epoch 71/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.1404 - accuracy: 0.9565\n",
      "Epoch 72/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.1506 - accuracy: 0.9542\n",
      "Epoch 73/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1571 - accuracy: 0.9451\n",
      "Epoch 74/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.1533 - accuracy: 0.9565\n",
      "Epoch 75/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.1519 - accuracy: 0.9474\n",
      "Epoch 76/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.1550 - accuracy: 0.9474\n",
      "Epoch 77/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.1530 - accuracy: 0.9565\n",
      "Epoch 78/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.1429 - accuracy: 0.9542\n",
      "Epoch 79/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.1347 - accuracy: 0.9611\n",
      "Epoch 80/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.1277 - accuracy: 0.9680\n",
      "Epoch 81/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.1293 - accuracy: 0.9611\n",
      "Epoch 82/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.1324 - accuracy: 0.9611\n",
      "Epoch 83/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.1416 - accuracy: 0.9565\n",
      "Epoch 84/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2232 - accuracy: 0.9176\n",
      "Epoch 85/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2377 - accuracy: 0.9222\n",
      "Epoch 86/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1806 - accuracy: 0.9451\n",
      "Epoch 87/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1613 - accuracy: 0.9474\n",
      "Epoch 88/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1539 - accuracy: 0.9588\n",
      "Epoch 89/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.1348 - accuracy: 0.9634\n",
      "Epoch 90/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1268 - accuracy: 0.9657\n",
      "Epoch 91/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1264 - accuracy: 0.9657\n",
      "Epoch 92/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1179 - accuracy: 0.9680\n",
      "Epoch 93/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.1230 - accuracy: 0.9634\n",
      "Epoch 94/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1274 - accuracy: 0.9565\n",
      "Epoch 95/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1214 - accuracy: 0.9657\n",
      "Epoch 96/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1182 - accuracy: 0.9680\n",
      "Epoch 97/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.1227 - accuracy: 0.9588\n",
      "Epoch 98/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1261 - accuracy: 0.9680\n",
      "Epoch 99/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1223 - accuracy: 0.9680\n",
      "Epoch 100/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1165 - accuracy: 0.9703\n",
      "Epoch 101/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1169 - accuracy: 0.9680\n",
      "Epoch 102/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1147 - accuracy: 0.9703\n",
      "Epoch 103/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1119 - accuracy: 0.9703\n",
      "Epoch 104/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.1124 - accuracy: 0.9680\n",
      "Epoch 105/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.1129 - accuracy: 0.9680\n",
      "Epoch 106/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1102 - accuracy: 0.9703\n",
      "Epoch 107/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.1136 - accuracy: 0.9680\n",
      "Epoch 108/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.1101 - accuracy: 0.9703\n",
      "Epoch 109/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1113 - accuracy: 0.9680\n",
      "Epoch 110/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.1189 - accuracy: 0.9634\n",
      "Epoch 111/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1143 - accuracy: 0.9703\n",
      "Epoch 112/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.1147 - accuracy: 0.9657\n",
      "Epoch 113/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1547 - accuracy: 0.9588\n",
      "Epoch 114/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1255 - accuracy: 0.9611\n",
      "Epoch 115/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.1535 - accuracy: 0.9519\n",
      "Epoch 116/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2038 - accuracy: 0.9291\n",
      "Epoch 117/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.1865 - accuracy: 0.9428\n",
      "Epoch 118/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.1495 - accuracy: 0.9497\n",
      "Epoch 119/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.1839 - accuracy: 0.9359\n",
      "Epoch 120/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1468 - accuracy: 0.9519\n",
      "Epoch 121/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1338 - accuracy: 0.9634\n",
      "Epoch 122/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1323 - accuracy: 0.9634\n",
      "Epoch 123/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.1377 - accuracy: 0.9588\n",
      "Epoch 124/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1285 - accuracy: 0.9634\n",
      "Epoch 125/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1277 - accuracy: 0.9657\n",
      "Epoch 126/200\n",
      "437/437 [==============================] - 0s 41us/step - loss: 0.1181 - accuracy: 0.9680\n",
      "Epoch 127/200\n",
      "437/437 [==============================] - 0s 37us/step - loss: 0.1160 - accuracy: 0.9703\n",
      "Epoch 128/200\n",
      "437/437 [==============================] - 0s 37us/step - loss: 0.1165 - accuracy: 0.9634\n",
      "Epoch 129/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.1211 - accuracy: 0.9634\n",
      "Epoch 130/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1345 - accuracy: 0.9519\n",
      "Epoch 131/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1695 - accuracy: 0.9451\n",
      "Epoch 132/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.1886 - accuracy: 0.9428\n",
      "Epoch 133/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.2020 - accuracy: 0.9519\n",
      "Epoch 134/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1654 - accuracy: 0.9519\n",
      "Epoch 135/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1673 - accuracy: 0.9451\n",
      "Epoch 136/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1953 - accuracy: 0.9382\n",
      "Epoch 137/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2055 - accuracy: 0.9336\n",
      "Epoch 138/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1350 - accuracy: 0.9634\n",
      "Epoch 139/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.1220 - accuracy: 0.9703\n",
      "Epoch 140/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1162 - accuracy: 0.9703\n",
      "Epoch 141/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1127 - accuracy: 0.9703\n",
      "Epoch 142/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.1117 - accuracy: 0.9703\n",
      "Epoch 143/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1120 - accuracy: 0.9703\n",
      "Epoch 144/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.1104 - accuracy: 0.9703\n",
      "Epoch 145/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.1086 - accuracy: 0.9703\n",
      "Epoch 146/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.1083 - accuracy: 0.9703\n",
      "Epoch 147/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.1109 - accuracy: 0.9703\n",
      "Epoch 148/200\n",
      "437/437 [==============================] - 0s 43us/step - loss: 0.1109 - accuracy: 0.9680\n",
      "Epoch 149/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.1082 - accuracy: 0.9680\n",
      "Epoch 150/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.1076 - accuracy: 0.9725\n",
      "Epoch 151/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.1066 - accuracy: 0.9703\n",
      "Epoch 152/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.1074 - accuracy: 0.9703\n",
      "Epoch 153/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.1083 - accuracy: 0.9703\n",
      "Epoch 154/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1065 - accuracy: 0.9703\n",
      "Epoch 155/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.1056 - accuracy: 0.9725\n",
      "Epoch 156/200\n",
      "437/437 [==============================] - 0s 37us/step - loss: 0.1056 - accuracy: 0.9703\n",
      "Epoch 157/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.1066 - accuracy: 0.9725\n",
      "Epoch 158/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.1096 - accuracy: 0.9703\n",
      "Epoch 159/200\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "437/437 [==============================] - 0s 34us/step - loss: 0.1099 - accuracy: 0.9703\n",
      "Epoch 160/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1076 - accuracy: 0.9680\n",
      "Epoch 161/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.1067 - accuracy: 0.9703\n",
      "Epoch 162/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.1055 - accuracy: 0.9703\n",
      "Epoch 163/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.1051 - accuracy: 0.9725\n",
      "Epoch 164/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.1046 - accuracy: 0.9725\n",
      "Epoch 165/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1041 - accuracy: 0.9725\n",
      "Epoch 166/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1043 - accuracy: 0.9748\n",
      "Epoch 167/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1045 - accuracy: 0.9725\n",
      "Epoch 168/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1069 - accuracy: 0.9725\n",
      "Epoch 169/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.1222 - accuracy: 0.9680\n",
      "Epoch 170/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1149 - accuracy: 0.9680\n",
      "Epoch 171/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.1098 - accuracy: 0.9680\n",
      "Epoch 172/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.1186 - accuracy: 0.9634\n",
      "Epoch 173/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1537 - accuracy: 0.9634\n",
      "Epoch 174/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1261 - accuracy: 0.9634\n",
      "Epoch 175/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2000 - accuracy: 0.9451\n",
      "Epoch 176/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.1877 - accuracy: 0.9405\n",
      "Epoch 177/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.1516 - accuracy: 0.9519\n",
      "Epoch 178/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1850 - accuracy: 0.9405\n",
      "Epoch 179/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1786 - accuracy: 0.9382\n",
      "Epoch 180/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.1478 - accuracy: 0.9565\n",
      "Epoch 181/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1394 - accuracy: 0.9634\n",
      "Epoch 182/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.1189 - accuracy: 0.9680\n",
      "Epoch 183/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.1142 - accuracy: 0.9680\n",
      "Epoch 184/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.1107 - accuracy: 0.9703\n",
      "Epoch 185/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1103 - accuracy: 0.9703\n",
      "Epoch 186/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.1103 - accuracy: 0.9725\n",
      "Epoch 187/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1097 - accuracy: 0.9703\n",
      "Epoch 188/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1096 - accuracy: 0.9680\n",
      "Epoch 189/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.1061 - accuracy: 0.9725\n",
      "Epoch 190/200\n",
      "437/437 [==============================] - 0s 39us/step - loss: 0.1087 - accuracy: 0.9680\n",
      "Epoch 191/200\n",
      "437/437 [==============================] - 0s 37us/step - loss: 0.1072 - accuracy: 0.9703\n",
      "Epoch 192/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1042 - accuracy: 0.9748\n",
      "Epoch 193/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.1053 - accuracy: 0.9725\n",
      "Epoch 194/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.1048 - accuracy: 0.9725\n",
      "Epoch 195/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.1039 - accuracy: 0.9703\n",
      "Epoch 196/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.1057 - accuracy: 0.9725\n",
      "Epoch 197/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1049 - accuracy: 0.9703\n",
      "Epoch 198/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.1032 - accuracy: 0.9725\n",
      "Epoch 199/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1040 - accuracy: 0.9725\n",
      "Epoch 200/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1046 - accuracy: 0.9725\n"
     ]
    }
   ],
   "source": [
    "nb_epoch = 200 \n",
    "batch_size = 32\n",
    "\n",
    "#Get the ANN parameters                   \n",
    "ANNModel.fit(X_train, y_train,\n",
    "              batch_size = batch_size,\n",
    "              epochs = nb_epoch)\n",
    "\n",
    "y_pred = ANNModel.predict_classes(X_test)\n",
    "\n",
    "# print(y_pred)\n",
    "# print(y_test)\n",
    "\n",
    "predy = []\n",
    "testy = []\n",
    "for i in range(len(y_pred)):\n",
    "    predy.append(y_pred[i][0])\n",
    "    testy.append(y_test.loc[i])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 161,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.870370\n",
      "Precision: 0.897436\n",
      "Recall: 0.867769\n",
      "F1 score: 0.882353\n"
     ]
    }
   ],
   "source": [
    "accuracy = accuracy_score(testy, predy)\n",
    "print('Accuracy: %f' % accuracy)\n",
    "# precision tp / (tp + fp)\n",
    "precision = precision_score(testy, predy)\n",
    "print('Precision: %f' % precision)\n",
    "# recall: tp / (tp + fn)\n",
    "recall = recall_score(testy, predy)\n",
    "print('Recall: %f' % recall)\n",
    "# f1: 2 tp / (2 tp + fp + fn)\n",
    "f1 = f1_score(testy, predy)\n",
    "print('F1 score: %f' % f1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 162,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_12\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "dense_36 (Dense)             (None, 50)                800       \n",
      "_________________________________________________________________\n",
      "dense_37 (Dense)             (None, 10)                510       \n",
      "_________________________________________________________________\n",
      "dense_38 (Dense)             (None, 1)                 11        \n",
      "=================================================================\n",
      "Total params: 1,321\n",
      "Trainable params: 1,321\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "#Define the ANN\n",
    "ANNModel = Sequential() #Sequential ANN\n",
    "\n",
    "#Layers of the ANN (Dense in this case)\n",
    "ANNModel.add(Dense(50, input_dim=15, activation='relu'))\n",
    "ANNModel.add(Dense(10, input_dim=50, activation='relu'))\n",
    "ANNModel.add(Dense(1, input_dim=10, activation='sigmoid'))\n",
    "adam = optimizers.Adam(lr=0.01, beta_1=0.9, beta_2=0.999, amsgrad=False)\n",
    "ANNModel.compile(loss='binary_crossentropy',\n",
    "              optimizer= adam,\n",
    "              metrics=['accuracy'])\n",
    "ANNModel.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 163,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/200\n",
      "437/437 [==============================] - 0s 201us/step - loss: 0.5535 - accuracy: 0.7185\n",
      "Epoch 2/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.3560 - accuracy: 0.8696\n",
      "Epoch 3/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.3332 - accuracy: 0.8741\n",
      "Epoch 4/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.3200 - accuracy: 0.8764\n",
      "Epoch 5/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.3050 - accuracy: 0.8810\n",
      "Epoch 6/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.3076 - accuracy: 0.8856\n",
      "Epoch 7/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.2998 - accuracy: 0.8947\n",
      "Epoch 8/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.2903 - accuracy: 0.8902\n",
      "Epoch 9/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.2776 - accuracy: 0.8902\n",
      "Epoch 10/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.2712 - accuracy: 0.8856\n",
      "Epoch 11/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2619 - accuracy: 0.8856\n",
      "Epoch 12/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.2622 - accuracy: 0.8902\n",
      "Epoch 13/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.2615 - accuracy: 0.8947\n",
      "Epoch 14/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2524 - accuracy: 0.8947\n",
      "Epoch 15/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2546 - accuracy: 0.8993\n",
      "Epoch 16/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2424 - accuracy: 0.8947\n",
      "Epoch 17/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2339 - accuracy: 0.8970\n",
      "Epoch 18/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2272 - accuracy: 0.8993\n",
      "Epoch 19/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.2386 - accuracy: 0.8947\n",
      "Epoch 20/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2342 - accuracy: 0.9108\n",
      "Epoch 21/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2298 - accuracy: 0.9085\n",
      "Epoch 22/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2175 - accuracy: 0.9085\n",
      "Epoch 23/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2118 - accuracy: 0.9016\n",
      "Epoch 24/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.2217 - accuracy: 0.9176\n",
      "Epoch 25/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.2072 - accuracy: 0.9062\n",
      "Epoch 26/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.1947 - accuracy: 0.9130\n",
      "Epoch 27/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.1971 - accuracy: 0.9222\n",
      "Epoch 28/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.1875 - accuracy: 0.9199\n",
      "Epoch 29/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1793 - accuracy: 0.9245\n",
      "Epoch 30/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.1941 - accuracy: 0.9085\n",
      "Epoch 31/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1973 - accuracy: 0.9176\n",
      "Epoch 32/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.1850 - accuracy: 0.9222\n",
      "Epoch 33/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.1830 - accuracy: 0.9291\n",
      "Epoch 34/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.1585 - accuracy: 0.9314\n",
      "Epoch 35/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1615 - accuracy: 0.9245\n",
      "Epoch 36/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1535 - accuracy: 0.9291\n",
      "Epoch 37/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.1599 - accuracy: 0.9268\n",
      "Epoch 38/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.1574 - accuracy: 0.9291\n",
      "Epoch 39/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.1418 - accuracy: 0.9405\n",
      "Epoch 40/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.1351 - accuracy: 0.9405\n",
      "Epoch 41/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.1277 - accuracy: 0.9497\n",
      "Epoch 42/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.1464 - accuracy: 0.9382\n",
      "Epoch 43/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.1439 - accuracy: 0.9382\n",
      "Epoch 44/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.1258 - accuracy: 0.9565\n",
      "Epoch 45/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.1134 - accuracy: 0.9451\n",
      "Epoch 46/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.1088 - accuracy: 0.9565\n",
      "Epoch 47/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.1025 - accuracy: 0.9611\n",
      "Epoch 48/200\n",
      "437/437 [==============================] - 0s 37us/step - loss: 0.0916 - accuracy: 0.9634\n",
      "Epoch 49/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.1049 - accuracy: 0.9542\n",
      "Epoch 50/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1066 - accuracy: 0.9428\n",
      "Epoch 51/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.1121 - accuracy: 0.9497\n",
      "Epoch 52/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.1169 - accuracy: 0.9588\n",
      "Epoch 53/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.1055 - accuracy: 0.9588\n",
      "Epoch 54/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0927 - accuracy: 0.9703\n",
      "Epoch 55/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.0863 - accuracy: 0.9771\n",
      "Epoch 56/200\n",
      "437/437 [==============================] - 0s 39us/step - loss: 0.0892 - accuracy: 0.9588\n",
      "Epoch 57/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.0879 - accuracy: 0.9634\n",
      "Epoch 58/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0919 - accuracy: 0.9611\n",
      "Epoch 59/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0880 - accuracy: 0.9565\n",
      "Epoch 60/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0742 - accuracy: 0.9703\n",
      "Epoch 61/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.0861 - accuracy: 0.9703\n",
      "Epoch 62/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.1353 - accuracy: 0.9451\n",
      "Epoch 63/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1383 - accuracy: 0.9497\n",
      "Epoch 64/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.1162 - accuracy: 0.9565\n",
      "Epoch 65/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.1058 - accuracy: 0.9542\n",
      "Epoch 66/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0726 - accuracy: 0.9794\n",
      "Epoch 67/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0701 - accuracy: 0.9771\n",
      "Epoch 68/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0659 - accuracy: 0.9794\n",
      "Epoch 69/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0593 - accuracy: 0.9840\n",
      "Epoch 70/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0604 - accuracy: 0.9817\n",
      "Epoch 71/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0723 - accuracy: 0.9748\n",
      "Epoch 72/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.1071 - accuracy: 0.9542\n",
      "Epoch 73/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.1079 - accuracy: 0.9588\n",
      "Epoch 74/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0828 - accuracy: 0.9771\n",
      "Epoch 75/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0631 - accuracy: 0.9771\n",
      "Epoch 76/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0530 - accuracy: 0.9863\n",
      "Epoch 77/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0558 - accuracy: 0.9840\n",
      "Epoch 78/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0594 - accuracy: 0.9725\n",
      "Epoch 79/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0578 - accuracy: 0.9771\n",
      "Epoch 80/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0511 - accuracy: 0.9817\n",
      "Epoch 81/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0413 - accuracy: 0.9863\n",
      "Epoch 82/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0453 - accuracy: 0.9863\n",
      "Epoch 83/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0411 - accuracy: 0.9886\n",
      "Epoch 84/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0389 - accuracy: 0.9863\n",
      "Epoch 85/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0363 - accuracy: 0.9908\n",
      "Epoch 86/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.0346 - accuracy: 0.9886\n",
      "Epoch 87/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.0361 - accuracy: 0.9908\n",
      "Epoch 88/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0381 - accuracy: 0.9863\n",
      "Epoch 89/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0366 - accuracy: 0.9886\n",
      "Epoch 90/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0336 - accuracy: 0.9886\n",
      "Epoch 91/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0324 - accuracy: 0.9886\n",
      "Epoch 92/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.0366 - accuracy: 0.9931\n",
      "Epoch 93/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0381 - accuracy: 0.9863\n",
      "Epoch 94/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0355 - accuracy: 0.9886\n",
      "Epoch 95/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0389 - accuracy: 0.9840\n",
      "Epoch 96/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0489 - accuracy: 0.9840\n",
      "Epoch 97/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0374 - accuracy: 0.9863\n",
      "Epoch 98/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0350 - accuracy: 0.9863\n",
      "Epoch 99/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0283 - accuracy: 0.9886\n",
      "Epoch 100/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0297 - accuracy: 0.9908\n",
      "Epoch 101/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0513 - accuracy: 0.9863\n",
      "Epoch 102/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0569 - accuracy: 0.9794\n",
      "Epoch 103/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0471 - accuracy: 0.9863\n",
      "Epoch 104/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0510 - accuracy: 0.9817\n",
      "Epoch 105/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0607 - accuracy: 0.9794\n",
      "Epoch 106/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0550 - accuracy: 0.9817\n",
      "Epoch 107/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0865 - accuracy: 0.9611\n",
      "Epoch 108/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0734 - accuracy: 0.9703\n",
      "Epoch 109/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0460 - accuracy: 0.9794\n",
      "Epoch 110/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.0355 - accuracy: 0.9908\n",
      "Epoch 111/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0402 - accuracy: 0.9840\n",
      "Epoch 112/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0313 - accuracy: 0.9931\n",
      "Epoch 113/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.0260 - accuracy: 0.9908\n",
      "Epoch 114/200\n",
      "437/437 [==============================] - 0s 23us/step - loss: 0.0261 - accuracy: 0.9908\n",
      "Epoch 115/200\n",
      "437/437 [==============================] - 0s 23us/step - loss: 0.0244 - accuracy: 0.9931\n",
      "Epoch 116/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.0243 - accuracy: 0.9886\n",
      "Epoch 117/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.0227 - accuracy: 0.9908\n",
      "Epoch 118/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.0281 - accuracy: 0.9931\n",
      "Epoch 119/200\n",
      "437/437 [==============================] - 0s 23us/step - loss: 0.0227 - accuracy: 0.9954\n",
      "Epoch 120/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.0235 - accuracy: 0.9931\n",
      "Epoch 121/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0239 - accuracy: 0.9954\n",
      "Epoch 122/200\n",
      "437/437 [==============================] - 0s 23us/step - loss: 0.0208 - accuracy: 0.9931\n",
      "Epoch 123/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.0193 - accuracy: 0.9931\n",
      "Epoch 124/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0177 - accuracy: 0.9954\n",
      "Epoch 125/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0172 - accuracy: 0.9954\n",
      "Epoch 126/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.0192 - accuracy: 0.9954\n",
      "Epoch 127/200\n",
      "437/437 [==============================] - 0s 55us/step - loss: 0.0190 - accuracy: 0.9931\n",
      "Epoch 128/200\n",
      "437/437 [==============================] - 0s 37us/step - loss: 0.0195 - accuracy: 0.9908\n",
      "Epoch 129/200\n",
      "437/437 [==============================] - 0s 37us/step - loss: 0.0186 - accuracy: 0.9954\n",
      "Epoch 130/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0149 - accuracy: 0.9954\n",
      "Epoch 131/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0196 - accuracy: 0.9931\n",
      "Epoch 132/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0238 - accuracy: 0.9908\n",
      "Epoch 133/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0154 - accuracy: 0.9931\n",
      "Epoch 134/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0196 - accuracy: 0.9931\n",
      "Epoch 135/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0229 - accuracy: 0.9931\n",
      "Epoch 136/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0155 - accuracy: 0.9931\n",
      "Epoch 137/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0166 - accuracy: 0.9954\n",
      "Epoch 138/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0143 - accuracy: 0.9954\n",
      "Epoch 139/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0141 - accuracy: 0.9954\n",
      "Epoch 140/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.0140 - accuracy: 0.9954\n",
      "Epoch 141/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0150 - accuracy: 0.9954\n",
      "Epoch 142/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0137 - accuracy: 0.9954\n",
      "Epoch 143/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0148 - accuracy: 0.9954\n",
      "Epoch 144/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0150 - accuracy: 0.9954\n",
      "Epoch 145/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.0192 - accuracy: 0.9931\n",
      "Epoch 146/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.0153 - accuracy: 0.9908\n",
      "Epoch 147/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0198 - accuracy: 0.9886\n",
      "Epoch 148/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0208 - accuracy: 0.9954\n",
      "Epoch 149/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0239 - accuracy: 0.9908\n",
      "Epoch 150/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.0169 - accuracy: 0.9908\n",
      "Epoch 151/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0223 - accuracy: 0.9931\n",
      "Epoch 152/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0306 - accuracy: 0.9886\n",
      "Epoch 153/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0245 - accuracy: 0.9886\n",
      "Epoch 154/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0272 - accuracy: 0.9908\n",
      "Epoch 155/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0521 - accuracy: 0.9817\n",
      "Epoch 156/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0404 - accuracy: 0.9863\n",
      "Epoch 157/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0783 - accuracy: 0.9817\n",
      "Epoch 158/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2643 - accuracy: 0.9336\n",
      "Epoch 159/200\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "437/437 [==============================] - 0s 30us/step - loss: 0.3892 - accuracy: 0.8947\n",
      "Epoch 160/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.1550 - accuracy: 0.9359\n",
      "Epoch 161/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0970 - accuracy: 0.9611\n",
      "Epoch 162/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.0697 - accuracy: 0.9748\n",
      "Epoch 163/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0568 - accuracy: 0.9840\n",
      "Epoch 164/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0442 - accuracy: 0.9886\n",
      "Epoch 165/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0387 - accuracy: 0.9931\n",
      "Epoch 166/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0324 - accuracy: 0.9908\n",
      "Epoch 167/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0284 - accuracy: 0.9908\n",
      "Epoch 168/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.0224 - accuracy: 0.9954\n",
      "Epoch 169/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0214 - accuracy: 0.9954\n",
      "Epoch 170/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0181 - accuracy: 0.9954\n",
      "Epoch 171/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0190 - accuracy: 0.9954\n",
      "Epoch 172/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0183 - accuracy: 0.9931\n",
      "Epoch 173/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.0162 - accuracy: 0.9954\n",
      "Epoch 174/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.0165 - accuracy: 0.9954\n",
      "Epoch 175/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0175 - accuracy: 0.9931\n",
      "Epoch 176/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0180 - accuracy: 0.9977\n",
      "Epoch 177/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0171 - accuracy: 0.9954\n",
      "Epoch 178/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0158 - accuracy: 0.9931\n",
      "Epoch 179/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0145 - accuracy: 0.9954\n",
      "Epoch 180/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.0139 - accuracy: 0.9954\n",
      "Epoch 181/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0167 - accuracy: 0.9886\n",
      "Epoch 182/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0141 - accuracy: 0.9954\n",
      "Epoch 183/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0126 - accuracy: 0.9977\n",
      "Epoch 184/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0121 - accuracy: 0.9954\n",
      "Epoch 185/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0107 - accuracy: 0.9977\n",
      "Epoch 186/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0153 - accuracy: 0.9954\n",
      "Epoch 187/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0129 - accuracy: 0.9977\n",
      "Epoch 188/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0145 - accuracy: 0.9954\n",
      "Epoch 189/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0125 - accuracy: 0.9977\n",
      "Epoch 190/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0103 - accuracy: 0.9954\n",
      "Epoch 191/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0099 - accuracy: 0.9954\n",
      "Epoch 192/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0100 - accuracy: 0.9954\n",
      "Epoch 193/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0119 - accuracy: 0.9977\n",
      "Epoch 194/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0135 - accuracy: 0.9954\n",
      "Epoch 195/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0131 - accuracy: 0.9954\n",
      "Epoch 196/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0139 - accuracy: 0.9954\n",
      "Epoch 197/200\n",
      "437/437 [==============================] - 0s 37us/step - loss: 0.0119 - accuracy: 0.9954\n",
      "Epoch 198/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.0139 - accuracy: 0.9977\n",
      "Epoch 199/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.0138 - accuracy: 0.9977\n",
      "Epoch 200/200\n",
      "437/437 [==============================] - ETA: 0s - loss: 0.0096 - accuracy: 1.00 - 0s 27us/step - loss: 0.0140 - accuracy: 0.9931\n"
     ]
    }
   ],
   "source": [
    "nb_epoch = 200 \n",
    "batch_size = 32\n",
    "\n",
    "#Get the ANN parameters                   \n",
    "ANNModel.fit(X_train, y_train,\n",
    "              batch_size = batch_size,\n",
    "              epochs = nb_epoch)\n",
    "\n",
    "y_pred = ANNModel.predict_classes(X_test)\n",
    "\n",
    "# print(y_pred)\n",
    "# print(y_test)\n",
    "\n",
    "predy = []\n",
    "testy = []\n",
    "for i in range(len(y_pred)):\n",
    "    predy.append(y_pred[i][0])\n",
    "    testy.append(y_test.loc[i])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 164,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.856481\n",
      "Precision: 0.875000\n",
      "Recall: 0.867769\n",
      "F1 score: 0.871369\n"
     ]
    }
   ],
   "source": [
    "accuracy = accuracy_score(testy, predy)\n",
    "print('Accuracy: %f' % accuracy)\n",
    "# precision tp / (tp + fp)\n",
    "precision = precision_score(testy, predy)\n",
    "print('Precision: %f' % precision)\n",
    "# recall: tp / (tp + fn)\n",
    "recall = recall_score(testy, predy)\n",
    "print('Recall: %f' % recall)\n",
    "# f1: 2 tp / (2 tp + fp + fn)\n",
    "f1 = f1_score(testy, predy)\n",
    "print('F1 score: %f' % f1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 165,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_13\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "dense_39 (Dense)             (None, 30)                480       \n",
      "_________________________________________________________________\n",
      "dense_40 (Dense)             (None, 3)                 93        \n",
      "_________________________________________________________________\n",
      "dense_41 (Dense)             (None, 1)                 4         \n",
      "=================================================================\n",
      "Total params: 577\n",
      "Trainable params: 577\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "#Define the ANN\n",
    "ANNModel = Sequential() #Sequential ANN\n",
    "\n",
    "#Layers of the ANN (Dense in this case)\n",
    "ANNModel.add(Dense(30, input_dim=15, activation='sigmoid'))\n",
    "ANNModel.add(Dense(3, input_dim=30, activation='sigmoid'))\n",
    "ANNModel.add(Dense(1, input_dim=3, activation='sigmoid'))\n",
    "adam = optimizers.Adam(lr=0.01, beta_1=0.9, beta_2=0.999, amsgrad=False)\n",
    "ANNModel.compile(loss='binary_crossentropy',\n",
    "              optimizer= adam,\n",
    "              metrics=['accuracy'])\n",
    "ANNModel.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 166,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/200\n",
      "437/437 [==============================] - 0s 203us/step - loss: 0.8642 - accuracy: 0.4600\n",
      "Epoch 2/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.6872 - accuracy: 0.5767\n",
      "Epoch 3/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.6805 - accuracy: 0.5469\n",
      "Epoch 4/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.6778 - accuracy: 0.5423\n",
      "Epoch 5/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.6702 - accuracy: 0.6247\n",
      "Epoch 6/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.6540 - accuracy: 0.8398\n",
      "Epoch 7/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.6306 - accuracy: 0.7849\n",
      "Epoch 8/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.5965 - accuracy: 0.8055\n",
      "Epoch 9/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.5545 - accuracy: 0.8513\n",
      "Epoch 10/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.5142 - accuracy: 0.8604\n",
      "Epoch 11/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.4804 - accuracy: 0.8719\n",
      "Epoch 12/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.4509 - accuracy: 0.8764\n",
      "Epoch 13/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.4287 - accuracy: 0.8810\n",
      "Epoch 14/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.4089 - accuracy: 0.8764\n",
      "Epoch 15/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.3922 - accuracy: 0.8856\n",
      "Epoch 16/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.3800 - accuracy: 0.8810\n",
      "Epoch 17/200\n",
      "437/437 [==============================] - 0s 46us/step - loss: 0.3707 - accuracy: 0.8833\n",
      "Epoch 18/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.3655 - accuracy: 0.8787\n",
      "Epoch 19/200\n",
      "437/437 [==============================] - 0s 37us/step - loss: 0.3554 - accuracy: 0.8902\n",
      "Epoch 20/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.3470 - accuracy: 0.8856\n",
      "Epoch 21/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.3453 - accuracy: 0.8787\n",
      "Epoch 22/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.3396 - accuracy: 0.8879\n",
      "Epoch 23/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.3331 - accuracy: 0.8856\n",
      "Epoch 24/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.3323 - accuracy: 0.8947\n",
      "Epoch 25/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.3283 - accuracy: 0.8787\n",
      "Epoch 26/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.3260 - accuracy: 0.8879\n",
      "Epoch 27/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.3234 - accuracy: 0.8787\n",
      "Epoch 28/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.3211 - accuracy: 0.8902\n",
      "Epoch 29/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.3175 - accuracy: 0.8856\n",
      "Epoch 30/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.3177 - accuracy: 0.8787\n",
      "Epoch 31/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.3155 - accuracy: 0.8947\n",
      "Epoch 32/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.3172 - accuracy: 0.8764\n",
      "Epoch 33/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.3103 - accuracy: 0.8924\n",
      "Epoch 34/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.3114 - accuracy: 0.8924\n",
      "Epoch 35/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.3089 - accuracy: 0.8947\n",
      "Epoch 36/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.3086 - accuracy: 0.8924\n",
      "Epoch 37/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.3069 - accuracy: 0.8947\n",
      "Epoch 38/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.3129 - accuracy: 0.8856\n",
      "Epoch 39/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.3127 - accuracy: 0.8947\n",
      "Epoch 40/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.3069 - accuracy: 0.8879\n",
      "Epoch 41/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.3050 - accuracy: 0.8947\n",
      "Epoch 42/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.3031 - accuracy: 0.8947\n",
      "Epoch 43/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.3050 - accuracy: 0.8856\n",
      "Epoch 44/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.3037 - accuracy: 0.8947\n",
      "Epoch 45/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.3024 - accuracy: 0.8902\n",
      "Epoch 46/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.3022 - accuracy: 0.8970\n",
      "Epoch 47/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.3024 - accuracy: 0.8947\n",
      "Epoch 48/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.3024 - accuracy: 0.8879\n",
      "Epoch 49/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.3030 - accuracy: 0.8924\n",
      "Epoch 50/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2997 - accuracy: 0.8902\n",
      "Epoch 51/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.3000 - accuracy: 0.9016\n",
      "Epoch 52/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.3025 - accuracy: 0.8879\n",
      "Epoch 53/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.2974 - accuracy: 0.8947\n",
      "Epoch 54/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2972 - accuracy: 0.8902\n",
      "Epoch 55/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2967 - accuracy: 0.8970\n",
      "Epoch 56/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.2968 - accuracy: 0.8947\n",
      "Epoch 57/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.2968 - accuracy: 0.8902\n",
      "Epoch 58/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.2952 - accuracy: 0.8993\n",
      "Epoch 59/200\n",
      "437/437 [==============================] - 0s 37us/step - loss: 0.2974 - accuracy: 0.8947\n",
      "Epoch 60/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2983 - accuracy: 0.8970\n",
      "Epoch 61/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2954 - accuracy: 0.8902\n",
      "Epoch 62/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.2942 - accuracy: 0.8993\n",
      "Epoch 63/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2971 - accuracy: 0.8924\n",
      "Epoch 64/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2959 - accuracy: 0.8879\n",
      "Epoch 65/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.2960 - accuracy: 0.8902\n",
      "Epoch 66/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2966 - accuracy: 0.8924\n",
      "Epoch 67/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.3087 - accuracy: 0.8856\n",
      "Epoch 68/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2994 - accuracy: 0.8856\n",
      "Epoch 69/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2938 - accuracy: 0.8947\n",
      "Epoch 70/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2965 - accuracy: 0.9039\n",
      "Epoch 71/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2952 - accuracy: 0.8924\n",
      "Epoch 72/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2938 - accuracy: 0.8924\n",
      "Epoch 73/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2928 - accuracy: 0.8970\n",
      "Epoch 74/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2947 - accuracy: 0.8833\n",
      "Epoch 75/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2933 - accuracy: 0.8993\n",
      "Epoch 76/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2925 - accuracy: 0.8993\n",
      "Epoch 77/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2922 - accuracy: 0.8924\n",
      "Epoch 78/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2940 - accuracy: 0.8856\n",
      "Epoch 79/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2911 - accuracy: 0.8924\n",
      "Epoch 80/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2950 - accuracy: 0.8947\n",
      "Epoch 81/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2955 - accuracy: 0.8993\n",
      "Epoch 82/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2992 - accuracy: 0.8833\n",
      "Epoch 83/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2930 - accuracy: 0.8879\n",
      "Epoch 84/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.2925 - accuracy: 0.8902\n",
      "Epoch 85/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.2909 - accuracy: 0.8970\n",
      "Epoch 86/200\n",
      "437/437 [==============================] - 0s 23us/step - loss: 0.2898 - accuracy: 0.8970\n",
      "Epoch 87/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2946 - accuracy: 0.8924\n",
      "Epoch 88/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2953 - accuracy: 0.8856\n",
      "Epoch 89/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2990 - accuracy: 0.8902\n",
      "Epoch 90/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2940 - accuracy: 0.8924\n",
      "Epoch 91/200\n",
      "437/437 [==============================] - 0s 23us/step - loss: 0.2903 - accuracy: 0.8902\n",
      "Epoch 92/200\n",
      "437/437 [==============================] - 0s 34us/step - loss: 0.2900 - accuracy: 0.9016\n",
      "Epoch 93/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2910 - accuracy: 0.8924\n",
      "Epoch 94/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2959 - accuracy: 0.8902\n",
      "Epoch 95/200\n",
      "437/437 [==============================] - 0s 23us/step - loss: 0.2944 - accuracy: 0.8924\n",
      "Epoch 96/200\n",
      "437/437 [==============================] - 0s 23us/step - loss: 0.2937 - accuracy: 0.8947\n",
      "Epoch 97/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.2909 - accuracy: 0.8947\n",
      "Epoch 98/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2903 - accuracy: 0.8993\n",
      "Epoch 99/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.2894 - accuracy: 0.8947\n",
      "Epoch 100/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.2898 - accuracy: 0.8993\n",
      "Epoch 101/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2911 - accuracy: 0.8993\n",
      "Epoch 102/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2889 - accuracy: 0.8947\n",
      "Epoch 103/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.2895 - accuracy: 0.8970\n",
      "Epoch 104/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.2896 - accuracy: 0.8902\n",
      "Epoch 105/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2952 - accuracy: 0.8947\n",
      "Epoch 106/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2922 - accuracy: 0.8947\n",
      "Epoch 107/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2908 - accuracy: 0.8902\n",
      "Epoch 108/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.2902 - accuracy: 0.8970\n",
      "Epoch 109/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.2880 - accuracy: 0.8970\n",
      "Epoch 110/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2968 - accuracy: 0.8924\n",
      "Epoch 111/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2885 - accuracy: 0.8970\n",
      "Epoch 112/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.2928 - accuracy: 0.8879\n",
      "Epoch 113/200\n",
      "437/437 [==============================] - 0s 23us/step - loss: 0.2901 - accuracy: 0.8924\n",
      "Epoch 114/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2903 - accuracy: 0.8970\n",
      "Epoch 115/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2920 - accuracy: 0.8902\n",
      "Epoch 116/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2928 - accuracy: 0.8924\n",
      "Epoch 117/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.2921 - accuracy: 0.9039\n",
      "Epoch 118/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2895 - accuracy: 0.8902\n",
      "Epoch 119/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2883 - accuracy: 0.8947\n",
      "Epoch 120/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2880 - accuracy: 0.8993\n",
      "Epoch 121/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2872 - accuracy: 0.9016\n",
      "Epoch 122/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2947 - accuracy: 0.8947\n",
      "Epoch 123/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2896 - accuracy: 0.8902\n",
      "Epoch 124/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.2885 - accuracy: 0.8947\n",
      "Epoch 125/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2872 - accuracy: 0.8970\n",
      "Epoch 126/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2874 - accuracy: 0.8970\n",
      "Epoch 127/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2891 - accuracy: 0.8993\n",
      "Epoch 128/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2867 - accuracy: 0.8993\n",
      "Epoch 129/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2902 - accuracy: 0.8924\n",
      "Epoch 130/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2858 - accuracy: 0.8970\n",
      "Epoch 131/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2878 - accuracy: 0.8947\n",
      "Epoch 132/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.2850 - accuracy: 0.9016\n",
      "Epoch 133/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2861 - accuracy: 0.8993\n",
      "Epoch 134/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.2843 - accuracy: 0.9016\n",
      "Epoch 135/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2850 - accuracy: 0.8970\n",
      "Epoch 136/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2832 - accuracy: 0.8993\n",
      "Epoch 137/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2844 - accuracy: 0.8947\n",
      "Epoch 138/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2853 - accuracy: 0.8970\n",
      "Epoch 139/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2874 - accuracy: 0.8947\n",
      "Epoch 140/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2832 - accuracy: 0.9016\n",
      "Epoch 141/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2844 - accuracy: 0.8970\n",
      "Epoch 142/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2856 - accuracy: 0.8924\n",
      "Epoch 143/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2894 - accuracy: 0.8924\n",
      "Epoch 144/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2859 - accuracy: 0.8993\n",
      "Epoch 145/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2815 - accuracy: 0.9062\n",
      "Epoch 146/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2817 - accuracy: 0.9039\n",
      "Epoch 147/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2808 - accuracy: 0.9039\n",
      "Epoch 148/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2814 - accuracy: 0.9016\n",
      "Epoch 149/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2819 - accuracy: 0.9016\n",
      "Epoch 150/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2802 - accuracy: 0.8993\n",
      "Epoch 151/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2803 - accuracy: 0.9016\n",
      "Epoch 152/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.2821 - accuracy: 0.9039\n",
      "Epoch 153/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2789 - accuracy: 0.8970\n",
      "Epoch 154/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2796 - accuracy: 0.9039\n",
      "Epoch 155/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2798 - accuracy: 0.9016\n",
      "Epoch 156/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2824 - accuracy: 0.8970\n",
      "Epoch 157/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2811 - accuracy: 0.8947\n",
      "Epoch 158/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2823 - accuracy: 0.9016\n",
      "Epoch 159/200\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "437/437 [==============================] - 0s 37us/step - loss: 0.2808 - accuracy: 0.8970\n",
      "Epoch 160/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2795 - accuracy: 0.8993\n",
      "Epoch 161/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2767 - accuracy: 0.9039\n",
      "Epoch 162/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2774 - accuracy: 0.9085\n",
      "Epoch 163/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2782 - accuracy: 0.9016\n",
      "Epoch 164/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2937 - accuracy: 0.8924\n",
      "Epoch 165/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2867 - accuracy: 0.9039\n",
      "Epoch 166/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.2798 - accuracy: 0.8993\n",
      "Epoch 167/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2754 - accuracy: 0.8993\n",
      "Epoch 168/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2764 - accuracy: 0.9108\n",
      "Epoch 169/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2760 - accuracy: 0.9062\n",
      "Epoch 170/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2746 - accuracy: 0.9062\n",
      "Epoch 171/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2766 - accuracy: 0.9085\n",
      "Epoch 172/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2762 - accuracy: 0.9039\n",
      "Epoch 173/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2769 - accuracy: 0.9108\n",
      "Epoch 174/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2744 - accuracy: 0.9062\n",
      "Epoch 175/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2730 - accuracy: 0.9108\n",
      "Epoch 176/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2725 - accuracy: 0.9039\n",
      "Epoch 177/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2749 - accuracy: 0.9039\n",
      "Epoch 178/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2754 - accuracy: 0.9062\n",
      "Epoch 179/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2726 - accuracy: 0.9062\n",
      "Epoch 180/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2748 - accuracy: 0.9085\n",
      "Epoch 181/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.2716 - accuracy: 0.9085\n",
      "Epoch 182/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2705 - accuracy: 0.9062\n",
      "Epoch 183/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2707 - accuracy: 0.9130\n",
      "Epoch 184/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2718 - accuracy: 0.9130\n",
      "Epoch 185/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2735 - accuracy: 0.9085\n",
      "Epoch 186/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2729 - accuracy: 0.9108\n",
      "Epoch 187/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2703 - accuracy: 0.9108\n",
      "Epoch 188/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2730 - accuracy: 0.9062\n",
      "Epoch 189/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2689 - accuracy: 0.9085\n",
      "Epoch 190/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2701 - accuracy: 0.9085\n",
      "Epoch 191/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2700 - accuracy: 0.9085\n",
      "Epoch 192/200\n",
      "437/437 [==============================] - 0s 25us/step - loss: 0.2699 - accuracy: 0.9108\n",
      "Epoch 193/200\n",
      "437/437 [==============================] - 0s 32us/step - loss: 0.2686 - accuracy: 0.9130\n",
      "Epoch 194/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2677 - accuracy: 0.9153\n",
      "Epoch 195/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2683 - accuracy: 0.9108\n",
      "Epoch 196/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2684 - accuracy: 0.9108\n",
      "Epoch 197/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2680 - accuracy: 0.9153\n",
      "Epoch 198/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2683 - accuracy: 0.9108\n",
      "Epoch 199/200\n",
      "437/437 [==============================] - 0s 30us/step - loss: 0.2681 - accuracy: 0.9062\n",
      "Epoch 200/200\n",
      "437/437 [==============================] - 0s 27us/step - loss: 0.2645 - accuracy: 0.9176\n"
     ]
    }
   ],
   "source": [
    "nb_epoch = 200 \n",
    "batch_size = 32\n",
    "\n",
    "#Get the ANN parameters                   \n",
    "ANNModel.fit(X_train, y_train,\n",
    "              batch_size = batch_size,\n",
    "              epochs = nb_epoch)\n",
    "\n",
    "y_pred = ANNModel.predict_classes(X_test)\n",
    "\n",
    "# print(y_pred)\n",
    "# print(y_test)\n",
    "\n",
    "predy = []\n",
    "testy = []\n",
    "for i in range(len(y_pred)):\n",
    "    predy.append(y_pred[i][0])\n",
    "    testy.append(y_test.loc[i])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 167,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.870370\n",
      "Precision: 0.897436\n",
      "Recall: 0.867769\n",
      "F1 score: 0.882353\n"
     ]
    }
   ],
   "source": [
    "accuracy = accuracy_score(testy, predy)\n",
    "print('Accuracy: %f' % accuracy)\n",
    "# precision tp / (tp + fp)\n",
    "precision = precision_score(testy, predy)\n",
    "print('Precision: %f' % precision)\n",
    "# recall: tp / (tp + fn)\n",
    "recall = recall_score(testy, predy)\n",
    "print('Recall: %f' % recall)\n",
    "# f1: 2 tp / (2 tp + fp + fn)\n",
    "f1 = f1_score(testy, predy)\n",
    "print('F1 score: %f' % f1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Q_2: Please list potential ethical bias inherent to solving the problem"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- Bias can also be introduced into the data through how they are collected or selected for use. Oversampling certain neighborhoods because they are overpoliced can result in recording more crime, which results in less credit-worthiness.\n",
    "\n",
    "- Older individuals have a higher likelihood of defaulting and reduces lending based on age, society and legal institutions may consider this to be illegal age discrimination.\n",
    "\n",
    "Conclusion:  Setting different decision thresholds for different groups may achieve the best balance, particularly if we believe some of the underlying variables in the model may be biased. Different metrics and standards will likely be required, depending on the use case and circumstances."
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.9"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}