01-sonar3.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# sonar3 data, classification"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Library import"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import tensorflow as tf\n",
    "from tensorflow.keras.models import Sequential\n",
    "from tensorflow.keras.layers import Dense, Input\n",
    "from sklearn.model_selection import KFold"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Random seed"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* 동일한 난수 결과를 얻기 위함\n",
    "* TensorFlow 모델 가중치 초기화, 랜덤 텐서 생성 등에서 동일한 결과 보장 "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "np.random.seed(3)\n",
    "tf.random.set_seed(3)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Data import"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* sonar3.csv "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "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>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "      <th>9</th>\n",
       "      <th>10</th>\n",
       "      <th>11</th>\n",
       "      <th>12</th>\n",
       "      <th>13</th>\n",
       "      <th>14</th>\n",
       "      <th>15</th>\n",
       "      <th>16</th>\n",
       "      <th>17</th>\n",
       "      <th>18</th>\n",
       "      <th>19</th>\n",
       "      <th>20</th>\n",
       "      <th>21</th>\n",
       "      <th>22</th>\n",
       "      <th>23</th>\n",
       "      <th>24</th>\n",
       "      <th>25</th>\n",
       "      <th>26</th>\n",
       "      <th>27</th>\n",
       "      <th>28</th>\n",
       "      <th>29</th>\n",
       "      <th>30</th>\n",
       "      <th>31</th>\n",
       "      <th>32</th>\n",
       "      <th>33</th>\n",
       "      <th>34</th>\n",
       "      <th>35</th>\n",
       "      <th>36</th>\n",
       "      <th>37</th>\n",
       "      <th>38</th>\n",
       "      <th>39</th>\n",
       "      <th>40</th>\n",
       "      <th>41</th>\n",
       "      <th>42</th>\n",
       "      <th>43</th>\n",
       "      <th>44</th>\n",
       "      <th>45</th>\n",
       "      <th>46</th>\n",
       "      <th>47</th>\n",
       "      <th>48</th>\n",
       "      <th>49</th>\n",
       "      <th>50</th>\n",
       "      <th>51</th>\n",
       "      <th>52</th>\n",
       "      <th>53</th>\n",
       "      <th>54</th>\n",
       "      <th>55</th>\n",
       "      <th>56</th>\n",
       "      <th>57</th>\n",
       "      <th>58</th>\n",
       "      <th>59</th>\n",
       "      <th>60</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.0200</td>\n",
       "      <td>0.0371</td>\n",
       "      <td>0.0428</td>\n",
       "      <td>0.0207</td>\n",
       "      <td>0.0954</td>\n",
       "      <td>0.0986</td>\n",
       "      <td>0.1539</td>\n",
       "      <td>0.1601</td>\n",
       "      <td>0.3109</td>\n",
       "      <td>0.2111</td>\n",
       "      <td>0.1609</td>\n",
       "      <td>0.1582</td>\n",
       "      <td>0.2238</td>\n",
       "      <td>0.0645</td>\n",
       "      <td>0.0660</td>\n",
       "      <td>0.2273</td>\n",
       "      <td>0.3100</td>\n",
       "      <td>0.2999</td>\n",
       "      <td>0.5078</td>\n",
       "      <td>0.4797</td>\n",
       "      <td>0.5783</td>\n",
       "      <td>0.5071</td>\n",
       "      <td>0.4328</td>\n",
       "      <td>0.5550</td>\n",
       "      <td>0.6711</td>\n",
       "      <td>0.6415</td>\n",
       "      <td>0.7104</td>\n",
       "      <td>0.8080</td>\n",
       "      <td>0.6791</td>\n",
       "      <td>0.3857</td>\n",
       "      <td>0.1307</td>\n",
       "      <td>0.2604</td>\n",
       "      <td>0.5121</td>\n",
       "      <td>0.7547</td>\n",
       "      <td>0.8537</td>\n",
       "      <td>0.8507</td>\n",
       "      <td>0.6692</td>\n",
       "      <td>0.6097</td>\n",
       "      <td>0.4943</td>\n",
       "      <td>0.2744</td>\n",
       "      <td>0.0510</td>\n",
       "      <td>0.2834</td>\n",
       "      <td>0.2825</td>\n",
       "      <td>0.4256</td>\n",
       "      <td>0.2641</td>\n",
       "      <td>0.1386</td>\n",
       "      <td>0.1051</td>\n",
       "      <td>0.1343</td>\n",
       "      <td>0.0383</td>\n",
       "      <td>0.0324</td>\n",
       "      <td>0.0232</td>\n",
       "      <td>0.0027</td>\n",
       "      <td>0.0065</td>\n",
       "      <td>0.0159</td>\n",
       "      <td>0.0072</td>\n",
       "      <td>0.0167</td>\n",
       "      <td>0.0180</td>\n",
       "      <td>0.0084</td>\n",
       "      <td>0.0090</td>\n",
       "      <td>0.0032</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.0453</td>\n",
       "      <td>0.0523</td>\n",
       "      <td>0.0843</td>\n",
       "      <td>0.0689</td>\n",
       "      <td>0.1183</td>\n",
       "      <td>0.2583</td>\n",
       "      <td>0.2156</td>\n",
       "      <td>0.3481</td>\n",
       "      <td>0.3337</td>\n",
       "      <td>0.2872</td>\n",
       "      <td>0.4918</td>\n",
       "      <td>0.6552</td>\n",
       "      <td>0.6919</td>\n",
       "      <td>0.7797</td>\n",
       "      <td>0.7464</td>\n",
       "      <td>0.9444</td>\n",
       "      <td>1.0000</td>\n",
       "      <td>0.8874</td>\n",
       "      <td>0.8024</td>\n",
       "      <td>0.7818</td>\n",
       "      <td>0.5212</td>\n",
       "      <td>0.4052</td>\n",
       "      <td>0.3957</td>\n",
       "      <td>0.3914</td>\n",
       "      <td>0.3250</td>\n",
       "      <td>0.3200</td>\n",
       "      <td>0.3271</td>\n",
       "      <td>0.2767</td>\n",
       "      <td>0.4423</td>\n",
       "      <td>0.2028</td>\n",
       "      <td>0.3788</td>\n",
       "      <td>0.2947</td>\n",
       "      <td>0.1984</td>\n",
       "      <td>0.2341</td>\n",
       "      <td>0.1306</td>\n",
       "      <td>0.4182</td>\n",
       "      <td>0.3835</td>\n",
       "      <td>0.1057</td>\n",
       "      <td>0.1840</td>\n",
       "      <td>0.1970</td>\n",
       "      <td>0.1674</td>\n",
       "      <td>0.0583</td>\n",
       "      <td>0.1401</td>\n",
       "      <td>0.1628</td>\n",
       "      <td>0.0621</td>\n",
       "      <td>0.0203</td>\n",
       "      <td>0.0530</td>\n",
       "      <td>0.0742</td>\n",
       "      <td>0.0409</td>\n",
       "      <td>0.0061</td>\n",
       "      <td>0.0125</td>\n",
       "      <td>0.0084</td>\n",
       "      <td>0.0089</td>\n",
       "      <td>0.0048</td>\n",
       "      <td>0.0094</td>\n",
       "      <td>0.0191</td>\n",
       "      <td>0.0140</td>\n",
       "      <td>0.0049</td>\n",
       "      <td>0.0052</td>\n",
       "      <td>0.0044</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.0262</td>\n",
       "      <td>0.0582</td>\n",
       "      <td>0.1099</td>\n",
       "      <td>0.1083</td>\n",
       "      <td>0.0974</td>\n",
       "      <td>0.2280</td>\n",
       "      <td>0.2431</td>\n",
       "      <td>0.3771</td>\n",
       "      <td>0.5598</td>\n",
       "      <td>0.6194</td>\n",
       "      <td>0.6333</td>\n",
       "      <td>0.7060</td>\n",
       "      <td>0.5544</td>\n",
       "      <td>0.5320</td>\n",
       "      <td>0.6479</td>\n",
       "      <td>0.6931</td>\n",
       "      <td>0.6759</td>\n",
       "      <td>0.7551</td>\n",
       "      <td>0.8929</td>\n",
       "      <td>0.8619</td>\n",
       "      <td>0.7974</td>\n",
       "      <td>0.6737</td>\n",
       "      <td>0.4293</td>\n",
       "      <td>0.3648</td>\n",
       "      <td>0.5331</td>\n",
       "      <td>0.2413</td>\n",
       "      <td>0.5070</td>\n",
       "      <td>0.8533</td>\n",
       "      <td>0.6036</td>\n",
       "      <td>0.8514</td>\n",
       "      <td>0.8512</td>\n",
       "      <td>0.5045</td>\n",
       "      <td>0.1862</td>\n",
       "      <td>0.2709</td>\n",
       "      <td>0.4232</td>\n",
       "      <td>0.3043</td>\n",
       "      <td>0.6116</td>\n",
       "      <td>0.6756</td>\n",
       "      <td>0.5375</td>\n",
       "      <td>0.4719</td>\n",
       "      <td>0.4647</td>\n",
       "      <td>0.2587</td>\n",
       "      <td>0.2129</td>\n",
       "      <td>0.2222</td>\n",
       "      <td>0.2111</td>\n",
       "      <td>0.0176</td>\n",
       "      <td>0.1348</td>\n",
       "      <td>0.0744</td>\n",
       "      <td>0.0130</td>\n",
       "      <td>0.0106</td>\n",
       "      <td>0.0033</td>\n",
       "      <td>0.0232</td>\n",
       "      <td>0.0166</td>\n",
       "      <td>0.0095</td>\n",
       "      <td>0.0180</td>\n",
       "      <td>0.0244</td>\n",
       "      <td>0.0316</td>\n",
       "      <td>0.0164</td>\n",
       "      <td>0.0095</td>\n",
       "      <td>0.0078</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.0100</td>\n",
       "      <td>0.0171</td>\n",
       "      <td>0.0623</td>\n",
       "      <td>0.0205</td>\n",
       "      <td>0.0205</td>\n",
       "      <td>0.0368</td>\n",
       "      <td>0.1098</td>\n",
       "      <td>0.1276</td>\n",
       "      <td>0.0598</td>\n",
       "      <td>0.1264</td>\n",
       "      <td>0.0881</td>\n",
       "      <td>0.1992</td>\n",
       "      <td>0.0184</td>\n",
       "      <td>0.2261</td>\n",
       "      <td>0.1729</td>\n",
       "      <td>0.2131</td>\n",
       "      <td>0.0693</td>\n",
       "      <td>0.2281</td>\n",
       "      <td>0.4060</td>\n",
       "      <td>0.3973</td>\n",
       "      <td>0.2741</td>\n",
       "      <td>0.3690</td>\n",
       "      <td>0.5556</td>\n",
       "      <td>0.4846</td>\n",
       "      <td>0.3140</td>\n",
       "      <td>0.5334</td>\n",
       "      <td>0.5256</td>\n",
       "      <td>0.2520</td>\n",
       "      <td>0.2090</td>\n",
       "      <td>0.3559</td>\n",
       "      <td>0.6260</td>\n",
       "      <td>0.7340</td>\n",
       "      <td>0.6120</td>\n",
       "      <td>0.3497</td>\n",
       "      <td>0.3953</td>\n",
       "      <td>0.3012</td>\n",
       "      <td>0.5408</td>\n",
       "      <td>0.8814</td>\n",
       "      <td>0.9857</td>\n",
       "      <td>0.9167</td>\n",
       "      <td>0.6121</td>\n",
       "      <td>0.5006</td>\n",
       "      <td>0.3210</td>\n",
       "      <td>0.3202</td>\n",
       "      <td>0.4295</td>\n",
       "      <td>0.3654</td>\n",
       "      <td>0.2655</td>\n",
       "      <td>0.1576</td>\n",
       "      <td>0.0681</td>\n",
       "      <td>0.0294</td>\n",
       "      <td>0.0241</td>\n",
       "      <td>0.0121</td>\n",
       "      <td>0.0036</td>\n",
       "      <td>0.0150</td>\n",
       "      <td>0.0085</td>\n",
       "      <td>0.0073</td>\n",
       "      <td>0.0050</td>\n",
       "      <td>0.0044</td>\n",
       "      <td>0.0040</td>\n",
       "      <td>0.0117</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.0762</td>\n",
       "      <td>0.0666</td>\n",
       "      <td>0.0481</td>\n",
       "      <td>0.0394</td>\n",
       "      <td>0.0590</td>\n",
       "      <td>0.0649</td>\n",
       "      <td>0.1209</td>\n",
       "      <td>0.2467</td>\n",
       "      <td>0.3564</td>\n",
       "      <td>0.4459</td>\n",
       "      <td>0.4152</td>\n",
       "      <td>0.3952</td>\n",
       "      <td>0.4256</td>\n",
       "      <td>0.4135</td>\n",
       "      <td>0.4528</td>\n",
       "      <td>0.5326</td>\n",
       "      <td>0.7306</td>\n",
       "      <td>0.6193</td>\n",
       "      <td>0.2032</td>\n",
       "      <td>0.4636</td>\n",
       "      <td>0.4148</td>\n",
       "      <td>0.4292</td>\n",
       "      <td>0.5730</td>\n",
       "      <td>0.5399</td>\n",
       "      <td>0.3161</td>\n",
       "      <td>0.2285</td>\n",
       "      <td>0.6995</td>\n",
       "      <td>1.0000</td>\n",
       "      <td>0.7262</td>\n",
       "      <td>0.4724</td>\n",
       "      <td>0.5103</td>\n",
       "      <td>0.5459</td>\n",
       "      <td>0.2881</td>\n",
       "      <td>0.0981</td>\n",
       "      <td>0.1951</td>\n",
       "      <td>0.4181</td>\n",
       "      <td>0.4604</td>\n",
       "      <td>0.3217</td>\n",
       "      <td>0.2828</td>\n",
       "      <td>0.2430</td>\n",
       "      <td>0.1979</td>\n",
       "      <td>0.2444</td>\n",
       "      <td>0.1847</td>\n",
       "      <td>0.0841</td>\n",
       "      <td>0.0692</td>\n",
       "      <td>0.0528</td>\n",
       "      <td>0.0357</td>\n",
       "      <td>0.0085</td>\n",
       "      <td>0.0230</td>\n",
       "      <td>0.0046</td>\n",
       "      <td>0.0156</td>\n",
       "      <td>0.0031</td>\n",
       "      <td>0.0054</td>\n",
       "      <td>0.0105</td>\n",
       "      <td>0.0110</td>\n",
       "      <td>0.0015</td>\n",
       "      <td>0.0072</td>\n",
       "      <td>0.0048</td>\n",
       "      <td>0.0107</td>\n",
       "      <td>0.0094</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       0       1       2       3       4   ...      56      57      58      59  60\n",
       "0  0.0200  0.0371  0.0428  0.0207  0.0954  ...  0.0180  0.0084  0.0090  0.0032   0\n",
       "1  0.0453  0.0523  0.0843  0.0689  0.1183  ...  0.0140  0.0049  0.0052  0.0044   0\n",
       "2  0.0262  0.0582  0.1099  0.1083  0.0974  ...  0.0316  0.0164  0.0095  0.0078   0\n",
       "3  0.0100  0.0171  0.0623  0.0205  0.0205  ...  0.0050  0.0044  0.0040  0.0117   0\n",
       "4  0.0762  0.0666  0.0481  0.0394  0.0590  ...  0.0072  0.0048  0.0107  0.0094   0\n",
       "\n",
       "[5 rows x 61 columns]"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = pd.read_csv('./data/sonar3.csv', header=None)\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* y data counting"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "60\n",
       "1    111\n",
       "0     97\n",
       "Name: count, dtype: int64"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df[60].value_counts()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Feature, Target seperate"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(       0       1       2       3   ...      56      57      58      59\n",
       " 0  0.0200  0.0371  0.0428  0.0207  ...  0.0180  0.0084  0.0090  0.0032\n",
       " 1  0.0453  0.0523  0.0843  0.0689  ...  0.0140  0.0049  0.0052  0.0044\n",
       " 2  0.0262  0.0582  0.1099  0.1083  ...  0.0316  0.0164  0.0095  0.0078\n",
       " 3  0.0100  0.0171  0.0623  0.0205  ...  0.0050  0.0044  0.0040  0.0117\n",
       " 4  0.0762  0.0666  0.0481  0.0394  ...  0.0072  0.0048  0.0107  0.0094\n",
       " \n",
       " [5 rows x 60 columns],\n",
       " 0    0\n",
       " 1    0\n",
       " 2    0\n",
       " 3    0\n",
       " 4    0\n",
       " Name: 60, dtype: int64)"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X = df.iloc[:, 0:60]\n",
    "y = df.iloc[:, 60]\n",
    "\n",
    "X.head(), y.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Train / Test split"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* Prevent overfitting\n",
    "* Model train -> train data / Model evaluation -> test data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3, random_state=13)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Define model"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* Binary classification => sigmoid (activation) \n",
    "* in general, the ReLU activation function is commonly used alongside the sigmoid function"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = Sequential()\n",
    "\n",
    "# 입력층\n",
    "model.add(Input(shape=(60,)))  # Input 객체를 첫 레이어로 사용\n",
    "\n",
    "# 은닉층\n",
    "model.add(Dense(24, activation=\"relu\"))\n",
    "model.add(Dense(10, activation=\"relu\"))\n",
    "\n",
    "# 출력층\n",
    "model.add(Dense(1, activation=\"sigmoid\"))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* Loss function => binary_crossentropy\n",
    "    * calculate the error between model expected values and real values\n",
    "\n",
    "* Optimizer => adam\n",
    "    * Based on SGD, automatically tunes learning rates\n",
    "\n",
    "* Metrics => accuracy\n",
    "    * Measure how accurately the model predictions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "model.compile(loss=\"binary_crossentropy\", optimizer=\"adam\", metrics=[\"accuracy\"])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* model summary"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\">Model: \"sequential_1\"</span>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[1mModel: \"sequential_1\"\u001b[0m\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\">┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓\n",
       "┃<span style=\"font-weight: bold\"> Layer (type)                    </span>┃<span style=\"font-weight: bold\"> Output Shape           </span>┃<span style=\"font-weight: bold\">       Param # </span>┃\n",
       "┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩\n",
       "│ dense_3 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dense</span>)                 │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">24</span>)             │         <span style=\"color: #00af00; text-decoration-color: #00af00\">1,464</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dense_4 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dense</span>)                 │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">10</span>)             │           <span style=\"color: #00af00; text-decoration-color: #00af00\">250</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dense_5 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dense</span>)                 │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">1</span>)              │            <span style=\"color: #00af00; text-decoration-color: #00af00\">11</span> │\n",
       "└─────────────────────────────────┴────────────────────────┴───────────────┘\n",
       "</pre>\n"
      ],
      "text/plain": [
       "┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓\n",
       "┃\u001b[1m \u001b[0m\u001b[1mLayer (type)                   \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mOutput Shape          \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1m      Param #\u001b[0m\u001b[1m \u001b[0m┃\n",
       "┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩\n",
       "│ dense_3 (\u001b[38;5;33mDense\u001b[0m)                 │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m24\u001b[0m)             │         \u001b[38;5;34m1,464\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dense_4 (\u001b[38;5;33mDense\u001b[0m)                 │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m10\u001b[0m)             │           \u001b[38;5;34m250\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dense_5 (\u001b[38;5;33mDense\u001b[0m)                 │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m1\u001b[0m)              │            \u001b[38;5;34m11\u001b[0m │\n",
       "└─────────────────────────────────┴────────────────────────┴───────────────┘\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\"> Total params: </span><span style=\"color: #00af00; text-decoration-color: #00af00\">1,725</span> (6.74 KB)\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[1m Total params: \u001b[0m\u001b[38;5;34m1,725\u001b[0m (6.74 KB)\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\"> Trainable params: </span><span style=\"color: #00af00; text-decoration-color: #00af00\">1,725</span> (6.74 KB)\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[1m Trainable params: \u001b[0m\u001b[38;5;34m1,725\u001b[0m (6.74 KB)\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\"> Non-trainable params: </span><span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> (0.00 B)\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[1m Non-trainable params: \u001b[0m\u001b[38;5;34m0\u001b[0m (0.00 B)\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "model.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* epochs\n",
    "    * Count the number of times the entire dataset is trained\n",
    "\n",
    "* batch_size\n",
    "    * Input data size for each weight update"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* entire dataset is trained 80 times\n",
    "* batch size => 10 \n",
    "    * train data length is 145\n",
    "    * 145 / 10 => 14.5 (15)\n",
    "    * The number of time the weight update is 15 (each epochs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 510us/step - accuracy: 0.4682 - loss: 0.6959\n",
      "Epoch 2/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 2ms/step - accuracy: 0.5647 - loss: 0.6735\n",
      "Epoch 3/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 344us/step - accuracy: 0.6272 - loss: 0.6628\n",
      "Epoch 4/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 390us/step - accuracy: 0.6364 - loss: 0.6526\n",
      "Epoch 5/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 359us/step - accuracy: 0.6729 - loss: 0.6420\n",
      "Epoch 6/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 458us/step - accuracy: 0.6798 - loss: 0.6309\n",
      "Epoch 7/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 337us/step - accuracy: 0.7035 - loss: 0.6208\n",
      "Epoch 8/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 303us/step - accuracy: 0.7065 - loss: 0.6102\n",
      "Epoch 9/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 297us/step - accuracy: 0.7262 - loss: 0.5993\n",
      "Epoch 10/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 316us/step - accuracy: 0.7182 - loss: 0.5878\n",
      "Epoch 11/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 312us/step - accuracy: 0.7340 - loss: 0.5764\n",
      "Epoch 12/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 408us/step - accuracy: 0.7335 - loss: 0.5653\n",
      "Epoch 13/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 328us/step - accuracy: 0.7335 - loss: 0.5532\n",
      "Epoch 14/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 302us/step - accuracy: 0.7463 - loss: 0.5400\n",
      "Epoch 15/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 317us/step - accuracy: 0.7564 - loss: 0.5282\n",
      "Epoch 16/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 312us/step - accuracy: 0.7682 - loss: 0.5172\n",
      "Epoch 17/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 654us/step - accuracy: 0.7763 - loss: 0.5076\n",
      "Epoch 18/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 301us/step - accuracy: 0.7763 - loss: 0.4980\n",
      "Epoch 19/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 310us/step - accuracy: 0.7552 - loss: 0.4890\n",
      "Epoch 20/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 316us/step - accuracy: 0.7552 - loss: 0.4807\n",
      "Epoch 21/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 296us/step - accuracy: 0.7610 - loss: 0.4730\n",
      "Epoch 22/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 319us/step - accuracy: 0.7822 - loss: 0.4653\n",
      "Epoch 23/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 347us/step - accuracy: 0.7940 - loss: 0.4587\n",
      "Epoch 24/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 315us/step - accuracy: 0.8022 - loss: 0.4518\n",
      "Epoch 25/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 299us/step - accuracy: 0.8022 - loss: 0.4452\n",
      "Epoch 26/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 2ms/step - accuracy: 0.8022 - loss: 0.4393 \n",
      "Epoch 27/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 299us/step - accuracy: 0.8022 - loss: 0.4345\n",
      "Epoch 28/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 328us/step - accuracy: 0.8104 - loss: 0.4281\n",
      "Epoch 29/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 315us/step - accuracy: 0.8022 - loss: 0.4241\n",
      "Epoch 30/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 342us/step - accuracy: 0.8139 - loss: 0.4189\n",
      "Epoch 31/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 285us/step - accuracy: 0.8139 - loss: 0.4147\n",
      "Epoch 32/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 318us/step - accuracy: 0.8224 - loss: 0.4103\n",
      "Epoch 33/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 307us/step - accuracy: 0.8224 - loss: 0.4064\n",
      "Epoch 34/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 315us/step - accuracy: 0.8224 - loss: 0.4025\n",
      "Epoch 35/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 371us/step - accuracy: 0.8365 - loss: 0.3987\n",
      "Epoch 36/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 301us/step - accuracy: 0.8365 - loss: 0.3949\n",
      "Epoch 37/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 316us/step - accuracy: 0.8512 - loss: 0.3922\n",
      "Epoch 38/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 324us/step - accuracy: 0.8512 - loss: 0.3883\n",
      "Epoch 39/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 348us/step - accuracy: 0.8512 - loss: 0.3852\n",
      "Epoch 40/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 277us/step - accuracy: 0.8512 - loss: 0.3819\n",
      "Epoch 41/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 300us/step - accuracy: 0.8547 - loss: 0.3793\n",
      "Epoch 42/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 590us/step - accuracy: 0.8547 - loss: 0.3760\n",
      "Epoch 43/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 317us/step - accuracy: 0.8597 - loss: 0.3730\n",
      "Epoch 44/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 332us/step - accuracy: 0.8597 - loss: 0.3700\n",
      "Epoch 45/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 290us/step - accuracy: 0.8590 - loss: 0.3672\n",
      "Epoch 46/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 290us/step - accuracy: 0.8590 - loss: 0.3646\n",
      "Epoch 47/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 315us/step - accuracy: 0.8590 - loss: 0.3615\n",
      "Epoch 48/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 306us/step - accuracy: 0.8590 - loss: 0.3589\n",
      "Epoch 49/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 282us/step - accuracy: 0.8590 - loss: 0.3558\n",
      "Epoch 50/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 320us/step - accuracy: 0.8590 - loss: 0.3534\n",
      "Epoch 51/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 316us/step - accuracy: 0.8590 - loss: 0.3508\n",
      "Epoch 52/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 2ms/step - accuracy: 0.8619 - loss: 0.3482\n",
      "Epoch 53/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 316us/step - accuracy: 0.8619 - loss: 0.3454\n",
      "Epoch 54/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 362us/step - accuracy: 0.8619 - loss: 0.3430\n",
      "Epoch 55/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 310us/step - accuracy: 0.8619 - loss: 0.3400\n",
      "Epoch 56/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 300us/step - accuracy: 0.8619 - loss: 0.3376\n",
      "Epoch 57/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 347us/step - accuracy: 0.8619 - loss: 0.3349\n",
      "Epoch 58/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 339us/step - accuracy: 0.8677 - loss: 0.3324\n",
      "Epoch 59/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 362us/step - accuracy: 0.8677 - loss: 0.3303\n",
      "Epoch 60/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 299us/step - accuracy: 0.8677 - loss: 0.3272\n",
      "Epoch 61/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 293us/step - accuracy: 0.8677 - loss: 0.3251\n",
      "Epoch 62/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 295us/step - accuracy: 0.8677 - loss: 0.3222\n",
      "Epoch 63/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 367us/step - accuracy: 0.8746 - loss: 0.3202\n",
      "Epoch 64/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 323us/step - accuracy: 0.8746 - loss: 0.3176\n",
      "Epoch 65/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 571us/step - accuracy: 0.8746 - loss: 0.3151\n",
      "Epoch 66/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 298us/step - accuracy: 0.8746 - loss: 0.3124\n",
      "Epoch 67/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 301us/step - accuracy: 0.8746 - loss: 0.3101\n",
      "Epoch 68/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 293us/step - accuracy: 0.8746 - loss: 0.3072\n",
      "Epoch 69/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 336us/step - accuracy: 0.8746 - loss: 0.3051\n",
      "Epoch 70/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 299us/step - accuracy: 0.8805 - loss: 0.3021\n",
      "Epoch 71/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 289us/step - accuracy: 0.8805 - loss: 0.3004\n",
      "Epoch 72/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 315us/step - accuracy: 0.8805 - loss: 0.2972\n",
      "Epoch 73/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 345us/step - accuracy: 0.8805 - loss: 0.2954\n",
      "Epoch 74/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 329us/step - accuracy: 0.8805 - loss: 0.2927\n",
      "Epoch 75/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 305us/step - accuracy: 0.8805 - loss: 0.2908\n",
      "Epoch 76/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 313us/step - accuracy: 0.8805 - loss: 0.2877\n",
      "Epoch 77/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 305us/step - accuracy: 0.8847 - loss: 0.2853\n",
      "Epoch 78/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 2ms/step - accuracy: 0.8847 - loss: 0.2840\n",
      "Epoch 79/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 342us/step - accuracy: 0.8847 - loss: 0.2804\n",
      "Epoch 80/80\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 345us/step - accuracy: 0.8847 - loss: 0.2783\n"
     ]
    }
   ],
   "source": [
    "hist = model.fit(X_train, y_train, epochs=80, batch_size=10)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* evaluate\n",
    "    * [Loss, Accuracy]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 1ms/step - accuracy: 0.7999 - loss: 0.4473 \n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[0.45234987139701843, 0.7936508059501648]"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.evaluate(X_test, y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1200x500 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "loss = hist.history['loss'] ## loss\n",
    "accuracy = hist.history['accuracy'] ## accuracy\n",
    "epochs = range(1, len(loss) + 1) ## epoch\n",
    "plt.figure(figsize=(12, 5))\n",
    "plt.subplot(1, 2, 1)\n",
    "plt.plot(epochs, loss, label='Training loss')\n",
    "plt.title('Training Loss')\n",
    "plt.xlabel('Epochs')\n",
    "plt.ylabel('Loss')\n",
    "plt.legend()\n",
    "\n",
    "plt.subplot(1, 2, 2)\n",
    "plt.plot(epochs, accuracy, label='Training accuracy')\n",
    "plt.title('Training Accuracy')\n",
    "plt.xlabel('Epochs')\n",
    "plt.ylabel('Accuracy')\n",
    "plt.legend()\n",
    "plt.tight_layout()\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# KFold cross validation"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* KFold CV\n",
    "    * Entire dataset split by \"k\" (k is int)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "k=5\n",
    "kfold = KFold(n_splits=k, shuffle=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [],
   "source": [
    "def model_fn():\n",
    "    model = Sequential()\n",
    "    model.add(Input(shape=(60,)))\n",
    "    model.add(Dense(24, activation=\"relu\"))\n",
    "    model.add(Dense(10, activation=\"relu\"))\n",
    "    model.add(Dense(1, activation=\"sigmoid\"))\n",
    "\n",
    "    return model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 1ms/step - accuracy: 0.8001 - loss: 0.7900 \n",
      "[0.7226483225822449, 0.8095238208770752]\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 1ms/step - accuracy: 0.8948 - loss: 0.6216 \n",
      "[0.5652133226394653, 0.9047619104385376]\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 1ms/step - accuracy: 0.8264 - loss: 0.7995 \n",
      "[0.7338466644287109, 0.8333333134651184]\n",
      "WARNING:tensorflow:5 out of the last 9 calls to <function TensorFlowTrainer.make_test_function.<locals>.one_step_on_iterator at 0x306c69300> triggered tf.function retracing. Tracing is expensive and the excessive number of tracings could be due to (1) creating @tf.function repeatedly in a loop, (2) passing tensors with different shapes, (3) passing Python objects instead of tensors. For (1), please define your @tf.function outside of the loop. For (2), @tf.function has reduce_retracing=True option that can avoid unnecessary retracing. For (3), please refer to https://www.tensorflow.org/guide/function#controlling_retracing and https://www.tensorflow.org/api_docs/python/tf/function for  more details.\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 1ms/step - accuracy: 0.7703 - loss: 0.5658 \n",
      "[0.5507982969284058, 0.7804877758026123]\n",
      "WARNING:tensorflow:6 out of the last 11 calls to <function TensorFlowTrainer.make_test_function.<locals>.one_step_on_iterator at 0x30553fec0> triggered tf.function retracing. Tracing is expensive and the excessive number of tracings could be due to (1) creating @tf.function repeatedly in a loop, (2) passing tensors with different shapes, (3) passing Python objects instead of tensors. For (1), please define your @tf.function outside of the loop. For (2), @tf.function has reduce_retracing=True option that can avoid unnecessary retracing. For (3), please refer to https://www.tensorflow.org/guide/function#controlling_retracing and https://www.tensorflow.org/api_docs/python/tf/function for  more details.\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 1ms/step - accuracy: 0.7645 - loss: 0.9384 \n",
      "[0.9043940305709839, 0.7560975551605225]\n"
     ]
    }
   ],
   "source": [
    "acc_score = []\n",
    "\n",
    "# kfold split -> train_index, test_index\n",
    "for train_index, test_index in kfold.split(X):\n",
    "\n",
    "    # define train index, test index\n",
    "    X_train, X_test = X.iloc[train_index, :], X.iloc[test_index, :]\n",
    "    y_train, y_test = y.iloc[train_index], y.iloc[test_index]\n",
    "    \n",
    "    # model\n",
    "    model = model_fn()\n",
    "    \n",
    "    # compile\n",
    "    model.compile(loss=\"binary_crossentropy\", optimizer='adam', metrics=['accuracy'])\n",
    "\n",
    "    # train\n",
    "    hist = model.fit(X_train, y_train, epochs=200, batch_size=10, verbose=0)\n",
    "\n",
    "    # accuracy\n",
    "    accuracy = model.evaluate(X_test, y_test)\n",
    "    print(accuracy)\n",
    "    acc_score.append(accuracy[1])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[0.8095238208770752,\n",
       " 0.9047619104385376,\n",
       " 0.8333333134651184,\n",
       " 0.7804877758026123,\n",
       " 0.7560975551605225]"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "acc_score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.8168408751487732\n"
     ]
    }
   ],
   "source": [
    "avg_acc_score = sum(acc_score) / k\n",
    "print(avg_acc_score)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "dev",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.12.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}