loan_project.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 136,
   "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>clientid</th>\n",
       "      <th>income</th>\n",
       "      <th>age</th>\n",
       "      <th>loan</th>\n",
       "      <th>loan_to_income</th>\n",
       "      <th>known_value</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>66155.92510</td>\n",
       "      <td>59.017015</td>\n",
       "      <td>8106.532131</td>\n",
       "      <td>0.122537</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>34415.15397</td>\n",
       "      <td>48.117153</td>\n",
       "      <td>6564.745018</td>\n",
       "      <td>0.190752</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3</td>\n",
       "      <td>57317.17006</td>\n",
       "      <td>63.108049</td>\n",
       "      <td>8020.953296</td>\n",
       "      <td>0.139940</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>4</td>\n",
       "      <td>42709.53420</td>\n",
       "      <td>45.751972</td>\n",
       "      <td>6103.642260</td>\n",
       "      <td>0.142911</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5</td>\n",
       "      <td>66952.68885</td>\n",
       "      <td>18.584336</td>\n",
       "      <td>8770.099235</td>\n",
       "      <td>0.130990</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   clientid       income        age         loan  loan_to_income  known_value\n",
       "0         1  66155.92510  59.017015  8106.532131        0.122537            0\n",
       "1         2  34415.15397  48.117153  6564.745018        0.190752            0\n",
       "2         3  57317.17006  63.108049  8020.953296        0.139940            0\n",
       "3         4  42709.53420  45.751972  6103.642260        0.142911            0\n",
       "4         5  66952.68885  18.584336  8770.099235        0.130990            1"
      ]
     },
     "execution_count": 136,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import pandas as pd\n",
    "credit_data=pd.read_csv('credit_data.csv')\n",
    "credit_data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 137,
   "metadata": {},
   "outputs": [],
   "source": [
    "x=credit_card.iloc[:, 1:4].values\n",
    "y = credit_card.iloc[:,5].values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 138,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.preprocessing import MinMaxScaler\n",
    "scaler = MinMaxScaler()\n",
    "scaler.fit(x)\n",
    "x_scaled=scaler.transform(x)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 139,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[0.9231759 , 0.89209175, 0.58883739],\n",
       "       [0.28812165, 0.65470788, 0.47682695],\n",
       "       [0.74633429, 0.9811888 , 0.58262011],\n",
       "       ...,\n",
       "       [0.48612202, 0.21695807, 0.40112895],\n",
       "       [0.47500998, 1.        , 0.1177903 ],\n",
       "       [0.98881367, 0.82970913, 0.53597028]])"
      ]
     },
     "execution_count": 139,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x_scaled"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 140,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "train_ratio = 0.70\n",
    "validation_ratio = 0.10\n",
    "test_ratio = 0.20\n",
    "\n",
    "# train is now 75% of the entire data set\n",
    "# the _junk suffix means that we drop that variable completely\n",
    "x_train, x_test, y_train, y_test = train_test_split(x_scaled, y, test_size=1 - train_ratio)\n",
    "\n",
    "# test is now 10% of the initial data set\n",
    "# validation is now 15% of the initial data set\n",
    "x_val, x_test, y_val, y_test = train_test_split(x_test, y_test, test_size=test_ratio/(test_ratio + validation_ratio)) \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 141,
   "metadata": {},
   "outputs": [],
   "source": [
    "# import keras library and packages\n",
    "# Initializing the ANN\n",
    "# Adding the input layer and the first hidden layer\n",
    "# Adding second hidden layer\n",
    "# Adding output layer\n",
    "# Compiling the ANN\n",
    "import keras\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Dense\n",
    "model = Sequential()\n",
    "model.add(Dense(units = 5,activation = 'sigmoid', input_dim = 3))\n",
    "model.add(Dense(units = 1, activation = 'sigmoid'))\n",
    "model.compile(optimizer = 'adam', loss = 'binary_crossentropy', metrics = ['accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 142,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 1399 samples, validate on 200 samples\n",
      "Epoch 1/100\n",
      "1399/1399 [==============================] - 3s 2ms/step - loss: 0.6149 - accuracy: 0.6769 - val_loss: 0.4750 - val_accuracy: 0.8550\n",
      "Epoch 2/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.4328 - accuracy: 0.8585 - val_loss: 0.4155 - val_accuracy: 0.8550\n",
      "Epoch 3/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.4013 - accuracy: 0.8585 - val_loss: 0.4018 - val_accuracy: 0.8550\n",
      "Epoch 4/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.3920 - accuracy: 0.8585 - val_loss: 0.3951 - val_accuracy: 0.8550\n",
      "Epoch 5/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.3853 - accuracy: 0.8585 - val_loss: 0.3886 - val_accuracy: 0.8550\n",
      "Epoch 6/100\n",
      "1399/1399 [==============================] - 2s 2ms/step - loss: 0.3784 - accuracy: 0.8585 - val_loss: 0.3816 - val_accuracy: 0.8550\n",
      "Epoch 7/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.3710 - accuracy: 0.8585 - val_loss: 0.3740 - val_accuracy: 0.8550\n",
      "Epoch 8/100\n",
      "1399/1399 [==============================] - 2s 2ms/step - loss: 0.3630 - accuracy: 0.8585 - val_loss: 0.3657 - val_accuracy: 0.8550\n",
      "Epoch 9/100\n",
      "1399/1399 [==============================] - 2s 2ms/step - loss: 0.3540 - accuracy: 0.8585 - val_loss: 0.3563 - val_accuracy: 0.8550\n",
      "Epoch 10/100\n",
      "1399/1399 [==============================] - 2s 2ms/step - loss: 0.3442 - accuracy: 0.8585 - val_loss: 0.3459 - val_accuracy: 0.8550\n",
      "Epoch 11/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.3333 - accuracy: 0.8585 - val_loss: 0.3348 - val_accuracy: 0.8550\n",
      "Epoch 12/100\n",
      "1399/1399 [==============================] - 2s 2ms/step - loss: 0.3215 - accuracy: 0.8585 - val_loss: 0.3227 - val_accuracy: 0.8550\n",
      "Epoch 13/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.3094 - accuracy: 0.8585 - val_loss: 0.3102 - val_accuracy: 0.8550\n",
      "Epoch 14/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.2965 - accuracy: 0.8585 - val_loss: 0.2976 - val_accuracy: 0.8550\n",
      "Epoch 15/100\n",
      "1399/1399 [==============================] - 2s 2ms/step - loss: 0.2840 - accuracy: 0.8585 - val_loss: 0.2855 - val_accuracy: 0.8550\n",
      "Epoch 16/100\n",
      "1399/1399 [==============================] - 2s 2ms/step - loss: 0.2721 - accuracy: 0.8656 - val_loss: 0.2731 - val_accuracy: 0.8550\n",
      "Epoch 17/100\n",
      "1399/1399 [==============================] - 2s 2ms/step - loss: 0.2610 - accuracy: 0.8721 - val_loss: 0.2624 - val_accuracy: 0.8700\n",
      "Epoch 18/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.2506 - accuracy: 0.8799 - val_loss: 0.2524 - val_accuracy: 0.8800\n",
      "Epoch 19/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.2414 - accuracy: 0.8906 - val_loss: 0.2424 - val_accuracy: 0.8800\n",
      "Epoch 20/100\n",
      "1399/1399 [==============================] - 3s 2ms/step - loss: 0.2327 - accuracy: 0.8921 - val_loss: 0.2340 - val_accuracy: 0.9050\n",
      "Epoch 21/100\n",
      "1399/1399 [==============================] - 2s 2ms/step - loss: 0.2249 - accuracy: 0.8985 - val_loss: 0.2262 - val_accuracy: 0.9250\n",
      "Epoch 22/100\n",
      "1399/1399 [==============================] - 3s 2ms/step - loss: 0.2180 - accuracy: 0.9035 - val_loss: 0.2186 - val_accuracy: 0.9250\n",
      "Epoch 23/100\n",
      "1399/1399 [==============================] - 2s 2ms/step - loss: 0.2115 - accuracy: 0.9114 - val_loss: 0.2113 - val_accuracy: 0.9300\n",
      "Epoch 24/100\n",
      "1399/1399 [==============================] - 3s 2ms/step - loss: 0.2056 - accuracy: 0.9142 - val_loss: 0.2058 - val_accuracy: 0.9400\n",
      "Epoch 25/100\n",
      "1399/1399 [==============================] - 3s 2ms/step - loss: 0.2004 - accuracy: 0.9214 - val_loss: 0.1991 - val_accuracy: 0.9350\n",
      "Epoch 26/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.1959 - accuracy: 0.9221 - val_loss: 0.1942 - val_accuracy: 0.9400\n",
      "Epoch 27/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.1913 - accuracy: 0.9249 - val_loss: 0.1899 - val_accuracy: 0.9350\n",
      "Epoch 28/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.1871 - accuracy: 0.9278 - val_loss: 0.1864 - val_accuracy: 0.9450\n",
      "Epoch 29/100\n",
      "1399/1399 [==============================] - 1s 889us/step - loss: 0.1837 - accuracy: 0.9278 - val_loss: 0.1820 - val_accuracy: 0.9450\n",
      "Epoch 30/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.1800 - accuracy: 0.9307 - val_loss: 0.1764 - val_accuracy: 0.9400\n",
      "Epoch 31/100\n",
      "1399/1399 [==============================] - 3s 2ms/step - loss: 0.1771 - accuracy: 0.9299 - val_loss: 0.1730 - val_accuracy: 0.9400\n",
      "Epoch 32/100\n",
      "1399/1399 [==============================] - 3s 2ms/step - loss: 0.1740 - accuracy: 0.9328 - val_loss: 0.1703 - val_accuracy: 0.9500\n",
      "Epoch 33/100\n",
      "1399/1399 [==============================] - 2s 2ms/step - loss: 0.1713 - accuracy: 0.9271 - val_loss: 0.1665 - val_accuracy: 0.9450\n",
      "Epoch 34/100\n",
      "1399/1399 [==============================] - 3s 2ms/step - loss: 0.1690 - accuracy: 0.9285 - val_loss: 0.1638 - val_accuracy: 0.9500\n",
      "Epoch 35/100\n",
      "1399/1399 [==============================] - 3s 2ms/step - loss: 0.1666 - accuracy: 0.9299 - val_loss: 0.1618 - val_accuracy: 0.9550\n",
      "Epoch 36/100\n",
      "1399/1399 [==============================] - 3s 2ms/step - loss: 0.1644 - accuracy: 0.9278 - val_loss: 0.1587 - val_accuracy: 0.9550\n",
      "Epoch 37/100\n",
      "1399/1399 [==============================] - 3s 2ms/step - loss: 0.1625 - accuracy: 0.9271 - val_loss: 0.1564 - val_accuracy: 0.9550\n",
      "Epoch 38/100\n",
      "1399/1399 [==============================] - 3s 2ms/step - loss: 0.1603 - accuracy: 0.9278 - val_loss: 0.1540 - val_accuracy: 0.9550\n",
      "Epoch 39/100\n",
      "1399/1399 [==============================] - 3s 2ms/step - loss: 0.1585 - accuracy: 0.9307 - val_loss: 0.1526 - val_accuracy: 0.9500\n",
      "Epoch 40/100\n",
      "1399/1399 [==============================] - 3s 2ms/step - loss: 0.1567 - accuracy: 0.9321 - val_loss: 0.1499 - val_accuracy: 0.9550\n",
      "Epoch 41/100\n",
      "1399/1399 [==============================] - 4s 3ms/step - loss: 0.1555 - accuracy: 0.9328 - val_loss: 0.1484 - val_accuracy: 0.9500\n",
      "Epoch 42/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.1538 - accuracy: 0.9342 - val_loss: 0.1463 - val_accuracy: 0.9500\n",
      "Epoch 43/100\n",
      "1399/1399 [==============================] - 3s 2ms/step - loss: 0.1523 - accuracy: 0.9342 - val_loss: 0.1450 - val_accuracy: 0.9550\n",
      "Epoch 44/100\n",
      "1399/1399 [==============================] - 4s 3ms/step - loss: 0.1508 - accuracy: 0.9335 - val_loss: 0.1436 - val_accuracy: 0.9500\n",
      "Epoch 45/100\n",
      "1399/1399 [==============================] - 3s 2ms/step - loss: 0.1499 - accuracy: 0.9342 - val_loss: 0.1421 - val_accuracy: 0.9500\n",
      "Epoch 46/100\n",
      "1399/1399 [==============================] - 2s 2ms/step - loss: 0.1484 - accuracy: 0.9364 - val_loss: 0.1410 - val_accuracy: 0.9500\n",
      "Epoch 47/100\n",
      "1399/1399 [==============================] - 3s 2ms/step - loss: 0.1473 - accuracy: 0.9378 - val_loss: 0.1399 - val_accuracy: 0.9500\n",
      "Epoch 48/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.1460 - accuracy: 0.9342 - val_loss: 0.1381 - val_accuracy: 0.9500\n",
      "Epoch 49/100\n",
      "1399/1399 [==============================] - 3s 2ms/step - loss: 0.1451 - accuracy: 0.9357 - val_loss: 0.1372 - val_accuracy: 0.9500\n",
      "Epoch 50/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.1440 - accuracy: 0.9371 - val_loss: 0.1360 - val_accuracy: 0.9500\n",
      "Epoch 51/100\n",
      "1399/1399 [==============================] - 3s 2ms/step - loss: 0.1433 - accuracy: 0.9357 - val_loss: 0.1351 - val_accuracy: 0.9500\n",
      "Epoch 52/100\n",
      "1399/1399 [==============================] - 3s 2ms/step - loss: 0.1423 - accuracy: 0.9371 - val_loss: 0.1346 - val_accuracy: 0.9450\n",
      "Epoch 53/100\n",
      "1399/1399 [==============================] - 2s 2ms/step - loss: 0.1416 - accuracy: 0.9371 - val_loss: 0.1344 - val_accuracy: 0.9450\n",
      "Epoch 54/100\n",
      "1399/1399 [==============================] - 4s 3ms/step - loss: 0.1404 - accuracy: 0.9371 - val_loss: 0.1324 - val_accuracy: 0.9450\n",
      "Epoch 55/100\n",
      "1399/1399 [==============================] - 2s 2ms/step - loss: 0.1398 - accuracy: 0.9392 - val_loss: 0.1321 - val_accuracy: 0.9450\n",
      "Epoch 56/100\n",
      "1399/1399 [==============================] - 2s 2ms/step - loss: 0.1393 - accuracy: 0.9392 - val_loss: 0.1309 - val_accuracy: 0.9450\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 57/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.1385 - accuracy: 0.9378 - val_loss: 0.1303 - val_accuracy: 0.9450\n",
      "Epoch 58/100\n",
      "1399/1399 [==============================] - 3s 2ms/step - loss: 0.1375 - accuracy: 0.9400 - val_loss: 0.1293 - val_accuracy: 0.9450\n",
      "Epoch 59/100\n",
      "1399/1399 [==============================] - 3s 2ms/step - loss: 0.1372 - accuracy: 0.9407 - val_loss: 0.1292 - val_accuracy: 0.9450\n",
      "Epoch 60/100\n",
      "1399/1399 [==============================] - 3s 2ms/step - loss: 0.1365 - accuracy: 0.9378 - val_loss: 0.1289 - val_accuracy: 0.9450\n",
      "Epoch 61/100\n",
      "1399/1399 [==============================] - 4s 3ms/step - loss: 0.1360 - accuracy: 0.9392 - val_loss: 0.1286 - val_accuracy: 0.9500\n",
      "Epoch 62/100\n",
      "1399/1399 [==============================] - 3s 2ms/step - loss: 0.1353 - accuracy: 0.9414 - val_loss: 0.1277 - val_accuracy: 0.9450\n",
      "Epoch 63/100\n",
      "1399/1399 [==============================] - 7s 5ms/step - loss: 0.1350 - accuracy: 0.9414 - val_loss: 0.1267 - val_accuracy: 0.94501s - l\n",
      "Epoch 64/100\n",
      "1399/1399 [==============================] - 5s 3ms/step - loss: 0.1343 - accuracy: 0.9400 - val_loss: 0.1259 - val_accuracy: 0.9450\n",
      "Epoch 65/100\n",
      "1399/1399 [==============================] - 4s 3ms/step - loss: 0.1340 - accuracy: 0.9421 - val_loss: 0.1258 - val_accuracy: 0.9450\n",
      "Epoch 66/100\n",
      "1399/1399 [==============================] - 3s 2ms/step - loss: 0.1336 - accuracy: 0.9407 - val_loss: 0.1260 - val_accuracy: 0.9500\n",
      "Epoch 67/100\n",
      "1399/1399 [==============================] - 4s 3ms/step - loss: 0.1330 - accuracy: 0.9414 - val_loss: 0.1250 - val_accuracy: 0.9500\n",
      "Epoch 68/100\n",
      "1399/1399 [==============================] - 3s 2ms/step - loss: 0.1326 - accuracy: 0.9428 - val_loss: 0.1252 - val_accuracy: 0.9500\n",
      "Epoch 69/100\n",
      "1399/1399 [==============================] - 4s 3ms/step - loss: 0.1322 - accuracy: 0.9407 - val_loss: 0.1243 - val_accuracy: 0.9500\n",
      "Epoch 70/100\n",
      "1399/1399 [==============================] - 4s 3ms/step - loss: 0.1314 - accuracy: 0.9414 - val_loss: 0.1260 - val_accuracy: 0.9450\n",
      "Epoch 71/100\n",
      "1399/1399 [==============================] - 4s 3ms/step - loss: 0.1317 - accuracy: 0.9428 - val_loss: 0.1242 - val_accuracy: 0.9500\n",
      "Epoch 72/100\n",
      "1399/1399 [==============================] - 2s 2ms/step - loss: 0.1313 - accuracy: 0.9435 - val_loss: 0.1240 - val_accuracy: 0.9500\n",
      "Epoch 73/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.1308 - accuracy: 0.9442 - val_loss: 0.1235 - val_accuracy: 0.9500\n",
      "Epoch 74/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.1301 - accuracy: 0.9414 - val_loss: 0.1248 - val_accuracy: 0.9450\n",
      "Epoch 75/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.1303 - accuracy: 0.9442 - val_loss: 0.1234 - val_accuracy: 0.9450\n",
      "Epoch 76/100\n",
      "1399/1399 [==============================] - 2s 2ms/step - loss: 0.1299 - accuracy: 0.9442 - val_loss: 0.1236 - val_accuracy: 0.9450\n",
      "Epoch 77/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.1295 - accuracy: 0.9457 - val_loss: 0.1219 - val_accuracy: 0.9500\n",
      "Epoch 78/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.1292 - accuracy: 0.9421 - val_loss: 0.1232 - val_accuracy: 0.9450\n",
      "Epoch 79/100\n",
      "1399/1399 [==============================] - 2s 2ms/step - loss: 0.1291 - accuracy: 0.9442 - val_loss: 0.1215 - val_accuracy: 0.9500\n",
      "Epoch 80/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.1287 - accuracy: 0.9450 - val_loss: 0.1213 - val_accuracy: 0.9500\n",
      "Epoch 81/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.1283 - accuracy: 0.9421 - val_loss: 0.1224 - val_accuracy: 0.9450\n",
      "Epoch 82/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.1284 - accuracy: 0.9457 - val_loss: 0.1221 - val_accuracy: 0.9450\n",
      "Epoch 83/100\n",
      "1399/1399 [==============================] - 2s 2ms/step - loss: 0.1280 - accuracy: 0.9421 - val_loss: 0.1207 - val_accuracy: 0.9450\n",
      "Epoch 84/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.1281 - accuracy: 0.9442 - val_loss: 0.1206 - val_accuracy: 0.9450\n",
      "Epoch 85/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.1281 - accuracy: 0.9435 - val_loss: 0.1207 - val_accuracy: 0.9500\n",
      "Epoch 86/100\n",
      "1399/1399 [==============================] - 2s 2ms/step - loss: 0.1274 - accuracy: 0.9414 - val_loss: 0.1216 - val_accuracy: 0.9450\n",
      "Epoch 87/100\n",
      "1399/1399 [==============================] - 2s 2ms/step - loss: 0.1272 - accuracy: 0.9450 - val_loss: 0.1211 - val_accuracy: 0.9450\n",
      "Epoch 88/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.1271 - accuracy: 0.9457 - val_loss: 0.1215 - val_accuracy: 0.9500\n",
      "Epoch 89/100\n",
      "1399/1399 [==============================] - 2s 2ms/step - loss: 0.1272 - accuracy: 0.9442 - val_loss: 0.1206 - val_accuracy: 0.9450\n",
      "Epoch 90/100\n",
      "1399/1399 [==============================] - 2s 2ms/step - loss: 0.1268 - accuracy: 0.9450 - val_loss: 0.1206 - val_accuracy: 0.9450\n",
      "Epoch 91/100\n",
      "1399/1399 [==============================] - 2s 2ms/step - loss: 0.1267 - accuracy: 0.9421 - val_loss: 0.1207 - val_accuracy: 0.9500\n",
      "Epoch 92/100\n",
      "1399/1399 [==============================] - 2s 2ms/step - loss: 0.1266 - accuracy: 0.9450 - val_loss: 0.1207 - val_accuracy: 0.9500\n",
      "Epoch 93/100\n",
      "1399/1399 [==============================] - 2s 2ms/step - loss: 0.1262 - accuracy: 0.9435 - val_loss: 0.1196 - val_accuracy: 0.9450\n",
      "Epoch 94/100\n",
      "1399/1399 [==============================] - 2s 2ms/step - loss: 0.1262 - accuracy: 0.9435 - val_loss: 0.1198 - val_accuracy: 0.9450\n",
      "Epoch 95/100\n",
      "1399/1399 [==============================] - 3s 2ms/step - loss: 0.1260 - accuracy: 0.9428 - val_loss: 0.1198 - val_accuracy: 0.9500\n",
      "Epoch 96/100\n",
      "1399/1399 [==============================] - 2s 2ms/step - loss: 0.1261 - accuracy: 0.9428 - val_loss: 0.1206 - val_accuracy: 0.9500\n",
      "Epoch 97/100\n",
      "1399/1399 [==============================] - 2s 2ms/step - loss: 0.1260 - accuracy: 0.9428 - val_loss: 0.1197 - val_accuracy: 0.9500\n",
      "Epoch 98/100\n",
      "1399/1399 [==============================] - 2s 2ms/step - loss: 0.1258 - accuracy: 0.9435 - val_loss: 0.1194 - val_accuracy: 0.9450\n",
      "Epoch 99/100\n",
      "1399/1399 [==============================] - 3s 2ms/step - loss: 0.1254 - accuracy: 0.9435 - val_loss: 0.1188 - val_accuracy: 0.9450\n",
      "Epoch 100/100\n",
      "1399/1399 [==============================] - 3s 2ms/step - loss: 0.1256 - accuracy: 0.9435 - val_loss: 0.1194 - val_accuracy: 0.9450\n"
     ]
    }
   ],
   "source": [
    "# Fitting the ANN to the training set\n",
    "history=model.fit(x_train, y_train, validation_data=(x_val, y_val), batch_size = 4, epochs = 100)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 145,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "plt.plot(history.history['accuracy'])\n",
    "plt.plot(history.history['val_accuracy'])\n",
    "plt.title('model accuracy')\n",
    "plt.ylabel('accuracy')\n",
    "plt.xlabel('epoch')\n",
    "plt.legend(['train', 'validation'], loc='upper left')\n",
    "plt.show()\n",
    "# \"Loss\"\n",
    "plt.plot(history.history['loss'])\n",
    "plt.plot(history.history['val_loss'])\n",
    "plt.title('model loss')\n",
    "plt.ylabel('loss')\n",
    "plt.xlabel('epoch')\n",
    "plt.legend(['train', 'validation'], loc='upper left')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 146,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9476309226932669"
      ]
     },
     "execution_count": 146,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_predict=model.predict(x_test)\n",
    "#print(test_predict)\n",
    "test_predict = (test_predict > 0.5)\n",
    "from sklearn.metrics import confusion_matrix\n",
    "cm = confusion_matrix(y_test, test_predict)\n",
    "accuracy = (cm[0][0]+cm[1][1])/(cm[0][0]+cm[0][1]+cm[1][0]+cm[1][1])\n",
    "accuracy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 147,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.945"
      ]
     },
     "execution_count": 147,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "val_predict=model.predict(x_val)\n",
    "#print(test_predict)\n",
    "val_predict = (val_predict > 0.5)\n",
    "from sklearn.metrics import confusion_matrix\n",
    "cm = confusion_matrix(y_val, val_predict)\n",
    "accuracy = (cm[0][0]+cm[1][1])/(cm[0][0]+cm[0][1]+cm[1][0]+cm[1][1])\n",
    "accuracy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 155,
   "metadata": {},
   "outputs": [],
   "source": [
    "credit_data.to_csv('F:/Faizan/MyLoans.csv', encoding='utf-8', index=False)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 156,
   "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>clientid</th>\n",
       "      <th>income</th>\n",
       "      <th>age</th>\n",
       "      <th>loan</th>\n",
       "      <th>loan_to_income</th>\n",
       "      <th>known_value</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>66155.92510</td>\n",
       "      <td>59.017015</td>\n",
       "      <td>8106.532131</td>\n",
       "      <td>0.122537</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>34415.15397</td>\n",
       "      <td>48.117153</td>\n",
       "      <td>6564.745018</td>\n",
       "      <td>0.190752</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3</td>\n",
       "      <td>57317.17006</td>\n",
       "      <td>63.108049</td>\n",
       "      <td>8020.953296</td>\n",
       "      <td>0.139940</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>4</td>\n",
       "      <td>42709.53420</td>\n",
       "      <td>45.751972</td>\n",
       "      <td>6103.642260</td>\n",
       "      <td>0.142911</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5</td>\n",
       "      <td>66952.68885</td>\n",
       "      <td>18.584336</td>\n",
       "      <td>8770.099235</td>\n",
       "      <td>0.130990</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   clientid       income        age         loan  loan_to_income  known_value\n",
       "0         1  66155.92510  59.017015  8106.532131        0.122537            0\n",
       "1         2  34415.15397  48.117153  6564.745018        0.190752            0\n",
       "2         3  57317.17006  63.108049  8020.953296        0.139940            0\n",
       "3         4  42709.53420  45.751972  6103.642260        0.142911            0\n",
       "4         5  66952.68885  18.584336  8770.099235        0.130990            1"
      ]
     },
     "execution_count": 156,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "loan_data=pd.read_csv('MyLoans.csv')\n",
    "loan_data.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Change the last five value of the know_value"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 173,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2000\n"
     ]
    }
   ],
   "source": [
    "update_col=loan_data.iloc[:,5].values\n",
    "print(len(update_col))\n",
    "for i in range(len(update_col)):\n",
    "    if i > 1995:\n",
    "        update_col[i]=1;\n",
    "      "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Change column name "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 174,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Change column name \n",
    "loan_data = loan_data.rename(columns = {'income':'Inc','age':'atttained age','loan':'borrowed','loan_to_income':'redundant_feature','known_value':'answer'})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 175,
   "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>clientid</th>\n",
       "      <th>Inc</th>\n",
       "      <th>atttained age</th>\n",
       "      <th>borrowed</th>\n",
       "      <th>redundant_feature</th>\n",
       "      <th>answer</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>66155.92510</td>\n",
       "      <td>59.017015</td>\n",
       "      <td>8106.532131</td>\n",
       "      <td>0.122537</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>34415.15397</td>\n",
       "      <td>48.117153</td>\n",
       "      <td>6564.745018</td>\n",
       "      <td>0.190752</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3</td>\n",
       "      <td>57317.17006</td>\n",
       "      <td>63.108049</td>\n",
       "      <td>8020.953296</td>\n",
       "      <td>0.139940</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>4</td>\n",
       "      <td>42709.53420</td>\n",
       "      <td>45.751972</td>\n",
       "      <td>6103.642260</td>\n",
       "      <td>0.142911</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5</td>\n",
       "      <td>66952.68885</td>\n",
       "      <td>18.584336</td>\n",
       "      <td>8770.099235</td>\n",
       "      <td>0.130990</td>\n",
       "      <td>1</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",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1995</th>\n",
       "      <td>1996</td>\n",
       "      <td>59221.04487</td>\n",
       "      <td>48.518179</td>\n",
       "      <td>1926.729397</td>\n",
       "      <td>0.032535</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1996</th>\n",
       "      <td>1997</td>\n",
       "      <td>69516.12757</td>\n",
       "      <td>23.162104</td>\n",
       "      <td>3503.176156</td>\n",
       "      <td>0.050394</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1997</th>\n",
       "      <td>1998</td>\n",
       "      <td>44311.44926</td>\n",
       "      <td>28.017167</td>\n",
       "      <td>5522.786693</td>\n",
       "      <td>0.124636</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1998</th>\n",
       "      <td>1999</td>\n",
       "      <td>43756.05660</td>\n",
       "      <td>63.971796</td>\n",
       "      <td>1622.722598</td>\n",
       "      <td>0.037086</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1999</th>\n",
       "      <td>2000</td>\n",
       "      <td>69436.57955</td>\n",
       "      <td>56.152617</td>\n",
       "      <td>7378.833599</td>\n",
       "      <td>0.106267</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>2000 rows × 6 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "      clientid          Inc  atttained age     borrowed  redundant_feature  \\\n",
       "0            1  66155.92510      59.017015  8106.532131           0.122537   \n",
       "1            2  34415.15397      48.117153  6564.745018           0.190752   \n",
       "2            3  57317.17006      63.108049  8020.953296           0.139940   \n",
       "3            4  42709.53420      45.751972  6103.642260           0.142911   \n",
       "4            5  66952.68885      18.584336  8770.099235           0.130990   \n",
       "...        ...          ...            ...          ...                ...   \n",
       "1995      1996  59221.04487      48.518179  1926.729397           0.032535   \n",
       "1996      1997  69516.12757      23.162104  3503.176156           0.050394   \n",
       "1997      1998  44311.44926      28.017167  5522.786693           0.124636   \n",
       "1998      1999  43756.05660      63.971796  1622.722598           0.037086   \n",
       "1999      2000  69436.57955      56.152617  7378.833599           0.106267   \n",
       "\n",
       "      answer  \n",
       "0          0  \n",
       "1          0  \n",
       "2          0  \n",
       "3          0  \n",
       "4          1  \n",
       "...      ...  \n",
       "1995       1  \n",
       "1996       1  \n",
       "1997       1  \n",
       "1998       1  \n",
       "1999       1  \n",
       "\n",
       "[2000 rows x 6 columns]"
      ]
     },
     "execution_count": 175,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "loan_data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 176,
   "metadata": {},
   "outputs": [],
   "source": [
    "x=loan_data.iloc[:, 1:4].values\n",
    "y = update_col"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 177,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.preprocessing import MinMaxScaler\n",
    "scaler = MinMaxScaler()\n",
    "scaler.fit(x)\n",
    "x_scaled=scaler.transform(x)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 178,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "train_ratio = 0.70\n",
    "validation_ratio = 0.10\n",
    "test_ratio = 0.20\n",
    "\n",
    "# train is now 75% of the entire data set\n",
    "# the _junk suffix means that we drop that variable completely\n",
    "x_train, x_test, y_train, y_test = train_test_split(x_scaled, y, test_size=1 - train_ratio)\n",
    "\n",
    "# test is now 10% of the initial data set\n",
    "# validation is now 15% of the initial data set\n",
    "x_val, x_test, y_val, y_test = train_test_split(x_test, y_test, test_size=test_ratio/(test_ratio + validation_ratio)) \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 182,
   "metadata": {},
   "outputs": [],
   "source": [
    "# import keras library and packages\n",
    "# Initializing the ANN\n",
    "# Adding the input layer and the first hidden layer\n",
    "# Adding second hidden layer\n",
    "# Adding output layer\n",
    "# Compiling the ANN\n",
    "import keras\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Dense\n",
    "model2 = Sequential()\n",
    "model2.add(Dense(units = 5,activation = 'sigmoid', input_dim = 3))\n",
    "model2.add(Dense(units = 1, activation = 'sigmoid'))\n",
    "model2.compile(optimizer = 'adam', loss = 'binary_crossentropy', metrics = ['accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 183,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 1399 samples, validate on 200 samples\n",
      "Epoch 1/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.5157 - accuracy: 0.8435 - val_loss: 0.4396 - val_accuracy: 0.8500\n",
      "Epoch 2/100\n",
      "1399/1399 [==============================] - 1s 1ms/step - loss: 0.4254 - accuracy: 0.8449 - val_loss: 0.4095 - val_accuracy: 0.8500\n",
      "Epoch 3/100\n",
      "1399/1399 [==============================] - 1s 937us/step - loss: 0.4091 - accuracy: 0.8449 - val_loss: 0.3970 - val_accuracy: 0.8500\n",
      "Epoch 4/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.3983 - accuracy: 0.8449 - val_loss: 0.3858 - val_accuracy: 0.8500\n",
      "Epoch 5/100\n",
      "1399/1399 [==============================] - 1s 1ms/step - loss: 0.3864 - accuracy: 0.8449 - val_loss: 0.3739 - val_accuracy: 0.8500\n",
      "Epoch 6/100\n",
      "1399/1399 [==============================] - 1s 1ms/step - loss: 0.3740 - accuracy: 0.8449 - val_loss: 0.3616 - val_accuracy: 0.8500\n",
      "Epoch 7/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.3615 - accuracy: 0.8449 - val_loss: 0.3484 - val_accuracy: 0.8500\n",
      "Epoch 8/100\n",
      "1399/1399 [==============================] - 1s 1ms/step - loss: 0.3485 - accuracy: 0.8449 - val_loss: 0.3352 - val_accuracy: 0.8500\n",
      "Epoch 9/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.3358 - accuracy: 0.8449 - val_loss: 0.3224 - val_accuracy: 0.8500\n",
      "Epoch 10/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.3232 - accuracy: 0.8449 - val_loss: 0.3100 - val_accuracy: 0.8500\n",
      "Epoch 11/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.3113 - accuracy: 0.8456 - val_loss: 0.2982 - val_accuracy: 0.8500\n",
      "Epoch 12/100\n",
      "1399/1399 [==============================] - 1s 1ms/step - loss: 0.3004 - accuracy: 0.8513 - val_loss: 0.2872 - val_accuracy: 0.8650\n",
      "Epoch 13/100\n",
      "1399/1399 [==============================] - 1s 1ms/step - loss: 0.2895 - accuracy: 0.8549 - val_loss: 0.2772 - val_accuracy: 0.8750\n",
      "Epoch 14/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.2799 - accuracy: 0.8699 - val_loss: 0.2676 - val_accuracy: 0.8800\n",
      "Epoch 15/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.2709 - accuracy: 0.8756 - val_loss: 0.2589 - val_accuracy: 0.8800\n",
      "Epoch 16/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.2625 - accuracy: 0.8821 - val_loss: 0.2511 - val_accuracy: 0.8850\n",
      "Epoch 17/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.2556 - accuracy: 0.8813 - val_loss: 0.2438 - val_accuracy: 0.8950\n",
      "Epoch 18/100\n",
      "1399/1399 [==============================] - 3s 2ms/step - loss: 0.2483 - accuracy: 0.8885 - val_loss: 0.2374 - val_accuracy: 0.9050\n",
      "Epoch 19/100\n",
      "1399/1399 [==============================] - 3s 2ms/step - loss: 0.2422 - accuracy: 0.8978 - val_loss: 0.2312 - val_accuracy: 0.9100\n",
      "Epoch 20/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.2367 - accuracy: 0.8999 - val_loss: 0.2258 - val_accuracy: 0.9150\n",
      "Epoch 21/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.2314 - accuracy: 0.9042 - val_loss: 0.2207 - val_accuracy: 0.9250\n",
      "Epoch 22/100\n",
      "1399/1399 [==============================] - 3s 2ms/step - loss: 0.2267 - accuracy: 0.9078 - val_loss: 0.2161 - val_accuracy: 0.9250\n",
      "Epoch 23/100\n",
      "1399/1399 [==============================] - 3s 2ms/step - loss: 0.2227 - accuracy: 0.9071 - val_loss: 0.2119 - val_accuracy: 0.9350\n",
      "Epoch 24/100\n",
      "1399/1399 [==============================] - 3s 2ms/step - loss: 0.2182 - accuracy: 0.9107 - val_loss: 0.2079 - val_accuracy: 0.9350\n",
      "Epoch 25/100\n",
      "1399/1399 [==============================] - 2s 2ms/step - loss: 0.2143 - accuracy: 0.9085 - val_loss: 0.2045 - val_accuracy: 0.9350\n",
      "Epoch 26/100\n",
      "1399/1399 [==============================] - 2s 2ms/step - loss: 0.2113 - accuracy: 0.9164 - val_loss: 0.2007 - val_accuracy: 0.9400\n",
      "Epoch 27/100\n",
      "1399/1399 [==============================] - 3s 2ms/step - loss: 0.2078 - accuracy: 0.9157 - val_loss: 0.1973 - val_accuracy: 0.9400\n",
      "Epoch 28/100\n",
      "1399/1399 [==============================] - 3s 2ms/step - loss: 0.2046 - accuracy: 0.9185 - val_loss: 0.1942 - val_accuracy: 0.9400\n",
      "Epoch 29/100\n",
      "1399/1399 [==============================] - 3s 2ms/step - loss: 0.2020 - accuracy: 0.9185 - val_loss: 0.1913 - val_accuracy: 0.9350\n",
      "Epoch 30/100\n",
      "1399/1399 [==============================] - 1s 964us/step - loss: 0.1990 - accuracy: 0.9178 - val_loss: 0.1886 - val_accuracy: 0.9350\n",
      "Epoch 31/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.1968 - accuracy: 0.9171 - val_loss: 0.1859 - val_accuracy: 0.9350\n",
      "Epoch 32/100\n",
      "1399/1399 [==============================] - 1s 862us/step - loss: 0.1945 - accuracy: 0.9171 - val_loss: 0.1834 - val_accuracy: 0.9400\n",
      "Epoch 33/100\n",
      "1399/1399 [==============================] - 1s 867us/step - loss: 0.1923 - accuracy: 0.9178 - val_loss: 0.1811 - val_accuracy: 0.9400\n",
      "Epoch 34/100\n",
      "1399/1399 [==============================] - 1s 868us/step - loss: 0.1901 - accuracy: 0.9171 - val_loss: 0.1789 - val_accuracy: 0.9400\n",
      "Epoch 35/100\n",
      "1399/1399 [==============================] - 1s 1ms/step - loss: 0.1883 - accuracy: 0.9164 - val_loss: 0.1769 - val_accuracy: 0.9400\n",
      "Epoch 36/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.1866 - accuracy: 0.9157 - val_loss: 0.1749 - val_accuracy: 0.9400\n",
      "Epoch 37/100\n",
      "1399/1399 [==============================] - 2s 2ms/step - loss: 0.1848 - accuracy: 0.9199 - val_loss: 0.1732 - val_accuracy: 0.9450\n",
      "Epoch 38/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.1833 - accuracy: 0.9178 - val_loss: 0.1713 - val_accuracy: 0.9450\n",
      "Epoch 39/100\n",
      "1399/1399 [==============================] - 2s 2ms/step - loss: 0.1816 - accuracy: 0.9157 - val_loss: 0.1696 - val_accuracy: 0.9450\n",
      "Epoch 40/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.1802 - accuracy: 0.9171 - val_loss: 0.1680 - val_accuracy: 0.9500\n",
      "Epoch 41/100\n",
      "1399/1399 [==============================] - 3s 2ms/step - loss: 0.1788 - accuracy: 0.9192 - val_loss: 0.1665 - val_accuracy: 0.9450\n",
      "Epoch 42/100\n",
      "1399/1399 [==============================] - 3s 2ms/step - loss: 0.1775 - accuracy: 0.9221 - val_loss: 0.1652 - val_accuracy: 0.9450\n",
      "Epoch 43/100\n",
      "1399/1399 [==============================] - 3s 2ms/step - loss: 0.1763 - accuracy: 0.9221 - val_loss: 0.1637 - val_accuracy: 0.9550\n",
      "Epoch 44/100\n",
      "1399/1399 [==============================] - 3s 2ms/step - loss: 0.1753 - accuracy: 0.9221 - val_loss: 0.1624 - val_accuracy: 0.9550\n",
      "Epoch 45/100\n",
      "1399/1399 [==============================] - 2s 2ms/step - loss: 0.1743 - accuracy: 0.9242 - val_loss: 0.1612 - val_accuracy: 0.9500\n",
      "Epoch 46/100\n",
      "1399/1399 [==============================] - 2s 2ms/step - loss: 0.1731 - accuracy: 0.9214 - val_loss: 0.1599 - val_accuracy: 0.9550\n",
      "Epoch 47/100\n",
      "1399/1399 [==============================] - 2s 2ms/step - loss: 0.1722 - accuracy: 0.9264 - val_loss: 0.1587 - val_accuracy: 0.9550\n",
      "Epoch 48/100\n",
      "1399/1399 [==============================] - 2s 2ms/step - loss: 0.1713 - accuracy: 0.9235 - val_loss: 0.1576 - val_accuracy: 0.9550\n",
      "Epoch 49/100\n",
      "1399/1399 [==============================] - 3s 2ms/step - loss: 0.1703 - accuracy: 0.9264 - val_loss: 0.1565 - val_accuracy: 0.9550\n",
      "Epoch 50/100\n",
      "1399/1399 [==============================] - 3s 2ms/step - loss: 0.1695 - accuracy: 0.9257 - val_loss: 0.1555 - val_accuracy: 0.9550\n",
      "Epoch 51/100\n",
      "1399/1399 [==============================] - 2s 2ms/step - loss: 0.1687 - accuracy: 0.9278 - val_loss: 0.1547 - val_accuracy: 0.9600\n",
      "Epoch 52/100\n",
      "1399/1399 [==============================] - 3s 2ms/step - loss: 0.1681 - accuracy: 0.9235 - val_loss: 0.1536 - val_accuracy: 0.9550\n",
      "Epoch 53/100\n",
      "1399/1399 [==============================] - 2s 2ms/step - loss: 0.1675 - accuracy: 0.9307 - val_loss: 0.1528 - val_accuracy: 0.9600\n",
      "Epoch 54/100\n",
      "1399/1399 [==============================] - 3s 2ms/step - loss: 0.1664 - accuracy: 0.9321 - val_loss: 0.1520 - val_accuracy: 0.9550\n",
      "Epoch 55/100\n",
      "1399/1399 [==============================] - 3s 2ms/step - loss: 0.1663 - accuracy: 0.9271 - val_loss: 0.1511 - val_accuracy: 0.9600\n",
      "Epoch 56/100\n",
      "1399/1399 [==============================] - 3s 2ms/step - loss: 0.1655 - accuracy: 0.9328 - val_loss: 0.1509 - val_accuracy: 0.9600\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 57/100\n",
      "1399/1399 [==============================] - 1s 1ms/step - loss: 0.1652 - accuracy: 0.9292 - val_loss: 0.1497 - val_accuracy: 0.9600\n",
      "Epoch 58/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.1646 - accuracy: 0.9299 - val_loss: 0.1490 - val_accuracy: 0.9600\n",
      "Epoch 59/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.1640 - accuracy: 0.9299 - val_loss: 0.1483 - val_accuracy: 0.9600\n",
      "Epoch 60/100\n",
      "1399/1399 [==============================] - 1s 929us/step - loss: 0.1635 - accuracy: 0.9307 - val_loss: 0.1478 - val_accuracy: 0.9600\n",
      "Epoch 61/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.1633 - accuracy: 0.9299 - val_loss: 0.1471 - val_accuracy: 0.9600\n",
      "Epoch 62/100\n",
      "1399/1399 [==============================] - 2s 2ms/step - loss: 0.1626 - accuracy: 0.9321 - val_loss: 0.1464 - val_accuracy: 0.9650\n",
      "Epoch 63/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.1623 - accuracy: 0.9299 - val_loss: 0.1457 - val_accuracy: 0.9650\n",
      "Epoch 64/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.1620 - accuracy: 0.9299 - val_loss: 0.1452 - val_accuracy: 0.9650\n",
      "Epoch 65/100\n",
      "1399/1399 [==============================] - 2s 2ms/step - loss: 0.1614 - accuracy: 0.9307 - val_loss: 0.1453 - val_accuracy: 0.9600\n",
      "Epoch 66/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.1611 - accuracy: 0.9314 - val_loss: 0.1441 - val_accuracy: 0.9650\n",
      "Epoch 67/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.1608 - accuracy: 0.9328 - val_loss: 0.1436 - val_accuracy: 0.9650\n",
      "Epoch 68/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.1607 - accuracy: 0.9307 - val_loss: 0.1432 - val_accuracy: 0.9650\n",
      "Epoch 69/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.1605 - accuracy: 0.9321 - val_loss: 0.1427 - val_accuracy: 0.9650\n",
      "Epoch 70/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.1602 - accuracy: 0.9321 - val_loss: 0.1426 - val_accuracy: 0.9650\n",
      "Epoch 71/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.1594 - accuracy: 0.9328 - val_loss: 0.1417 - val_accuracy: 0.9650\n",
      "Epoch 72/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.1593 - accuracy: 0.9328 - val_loss: 0.1417 - val_accuracy: 0.9650\n",
      "Epoch 73/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.1591 - accuracy: 0.9342 - val_loss: 0.1412 - val_accuracy: 0.9650\n",
      "Epoch 74/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.1588 - accuracy: 0.9321 - val_loss: 0.1407 - val_accuracy: 0.9650\n",
      "Epoch 75/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.1587 - accuracy: 0.9335 - val_loss: 0.1402 - val_accuracy: 0.9650\n",
      "Epoch 76/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.1585 - accuracy: 0.9335 - val_loss: 0.1396 - val_accuracy: 0.9650\n",
      "Epoch 77/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.1581 - accuracy: 0.9335 - val_loss: 0.1392 - val_accuracy: 0.9650\n",
      "Epoch 78/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.1579 - accuracy: 0.9328 - val_loss: 0.1394 - val_accuracy: 0.9650\n",
      "Epoch 79/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.1577 - accuracy: 0.9328 - val_loss: 0.1390 - val_accuracy: 0.9650\n",
      "Epoch 80/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.1579 - accuracy: 0.9342 - val_loss: 0.1387 - val_accuracy: 0.9650\n",
      "Epoch 81/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.1574 - accuracy: 0.9357 - val_loss: 0.1378 - val_accuracy: 0.9650\n",
      "Epoch 82/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.1573 - accuracy: 0.9335 - val_loss: 0.1377 - val_accuracy: 0.9650\n",
      "Epoch 83/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.1572 - accuracy: 0.9321 - val_loss: 0.1374 - val_accuracy: 0.9650\n",
      "Epoch 84/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.1569 - accuracy: 0.9321 - val_loss: 0.1370 - val_accuracy: 0.9650\n",
      "Epoch 85/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.1567 - accuracy: 0.9321 - val_loss: 0.1369 - val_accuracy: 0.9650\n",
      "Epoch 86/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.1567 - accuracy: 0.9335 - val_loss: 0.1364 - val_accuracy: 0.9650\n",
      "Epoch 87/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.1563 - accuracy: 0.9314 - val_loss: 0.1364 - val_accuracy: 0.9650\n",
      "Epoch 88/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.1563 - accuracy: 0.9328 - val_loss: 0.1359 - val_accuracy: 0.9650\n",
      "Epoch 89/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.1562 - accuracy: 0.9335 - val_loss: 0.1358 - val_accuracy: 0.9650\n",
      "Epoch 90/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.1561 - accuracy: 0.9342 - val_loss: 0.1357 - val_accuracy: 0.9650\n",
      "Epoch 91/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.1557 - accuracy: 0.9335 - val_loss: 0.1350 - val_accuracy: 0.9700\n",
      "Epoch 92/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.1558 - accuracy: 0.9342 - val_loss: 0.1347 - val_accuracy: 0.9700\n",
      "Epoch 93/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.1555 - accuracy: 0.9350 - val_loss: 0.1351 - val_accuracy: 0.9650\n",
      "Epoch 94/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.1556 - accuracy: 0.9335 - val_loss: 0.1346 - val_accuracy: 0.9700\n",
      "Epoch 95/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.1554 - accuracy: 0.9342 - val_loss: 0.1339 - val_accuracy: 0.9700\n",
      "Epoch 96/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.1554 - accuracy: 0.9335 - val_loss: 0.1340 - val_accuracy: 0.9700\n",
      "Epoch 97/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.1551 - accuracy: 0.9328 - val_loss: 0.1344 - val_accuracy: 0.9600\n",
      "Epoch 98/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.1551 - accuracy: 0.9335 - val_loss: 0.1336 - val_accuracy: 0.9700\n",
      "Epoch 99/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.1550 - accuracy: 0.9342 - val_loss: 0.1334 - val_accuracy: 0.9700\n",
      "Epoch 100/100\n",
      "1399/1399 [==============================] - 2s 1ms/step - loss: 0.1548 - accuracy: 0.9328 - val_loss: 0.1329 - val_accuracy: 0.9700\n"
     ]
    }
   ],
   "source": [
    "# Fitting the ANN to the training set\n",
    "history1=model2.fit(x_train, y_train, validation_data=(x_val, y_val), batch_size = 4, epochs = 100)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 184,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "plt.plot(history1.history['accuracy'])\n",
    "plt.plot(history1.history['val_accuracy'])\n",
    "plt.title('model accuracy')\n",
    "plt.ylabel('accuracy')\n",
    "plt.xlabel('epoch')\n",
    "plt.legend(['train', 'validation'], loc='upper left')\n",
    "plt.show()\n",
    "# \"Loss\"\n",
    "plt.plot(history1.history['loss'])\n",
    "plt.plot(history1.history['val_loss'])\n",
    "plt.title('model loss')\n",
    "plt.ylabel('loss')\n",
    "plt.xlabel('epoch')\n",
    "plt.legend(['train', 'validation'], loc='upper left')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 185,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9526184538653366"
      ]
     },
     "execution_count": 185,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Test Accuracy for laon dataset\n",
    "test_predict=model2.predict(x_test)\n",
    "#print(test_predict)\n",
    "test_predict = (test_predict > 0.5)\n",
    "from sklearn.metrics import confusion_matrix\n",
    "cm = confusion_matrix(y_test, test_predict)\n",
    "accuracy = (cm[0][0]+cm[1][1])/(cm[0][0]+cm[0][1]+cm[1][0]+cm[1][1])\n",
    "accuracy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 186,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.97"
      ]
     },
     "execution_count": 186,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Valdation accuracy for laon dataset\n",
    "val_predict=model2.predict(x_val)\n",
    "#print(test_predict)\n",
    "val_predict = (val_predict > 0.5)\n",
    "from sklearn.metrics import confusion_matrix\n",
    "cm = confusion_matrix(y_val, val_predict)\n",
    "accuracy = (cm[0][0]+cm[1][1])/(cm[0][0]+cm[0][1]+cm[1][0]+cm[1][1])\n",
    "accuracy"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Regression Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 187,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LogisticRegression(C=1.0, class_weight=None, dual=False, fit_intercept=True,\n",
       "                   intercept_scaling=1, l1_ratio=None, max_iter=100,\n",
       "                   multi_class='auto', n_jobs=None, penalty='l2',\n",
       "                   random_state=None, solver='lbfgs', tol=0.0001, verbose=0,\n",
       "                   warm_start=False)"
      ]
     },
     "execution_count": 187,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn import metrics\n",
    " \n",
    "model3 = LogisticRegression()\n",
    "model3.fit(x_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 188,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9476309226932669\n"
     ]
    }
   ],
   "source": [
    "# Test data accuracy against loan dataset\n",
    "pred=model3.predict(x_test)\n",
    "# Print the prediction accuracy\n",
    "print (metrics.accuracy_score(y_test, pred))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 189,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.945\n"
     ]
    }
   ],
   "source": [
    "# Validation data accuracy against loan dataset\n",
    "pred=model3.predict(x_val)\n",
    "# Print the prediction accuracy\n",
    "print (metrics.accuracy_score(y_val, pred))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# What was the purpose of this assignment? What can you do now as a result of accomplishing it?"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### The purpose of this assignment is to analyze the effect of different features in the dataset. Changing the last five value of the known value column increase the accuracy of the model because now the probiblity of both 0 and 1 are equally same. Regression model result are less then classification result."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}