CostPred_SimpleNN.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Simple Cost prediction with Neural Networks"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 157,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "\n",
    "import seaborn as sns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 158,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.read_csv('./DATA/fake_reg.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 159,
   "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>price</th>\n",
       "      <th>feature1</th>\n",
       "      <th>feature2</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>461.527929</td>\n",
       "      <td>999.787558</td>\n",
       "      <td>999.766096</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>548.130011</td>\n",
       "      <td>998.861615</td>\n",
       "      <td>1001.042403</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>410.297162</td>\n",
       "      <td>1000.070267</td>\n",
       "      <td>998.844015</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>540.382220</td>\n",
       "      <td>999.952251</td>\n",
       "      <td>1000.440940</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>546.024553</td>\n",
       "      <td>1000.446011</td>\n",
       "      <td>1000.338531</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        price     feature1     feature2\n",
       "0  461.527929   999.787558   999.766096\n",
       "1  548.130011   998.861615  1001.042403\n",
       "2  410.297162  1000.070267   998.844015\n",
       "3  540.382220   999.952251  1000.440940\n",
       "4  546.024553  1000.446011  1000.338531"
      ]
     },
     "execution_count": 159,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 160,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<seaborn.axisgrid.PairGrid at 0x7fb99899ed50>"
      ]
     },
     "execution_count": 160,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 540x540 with 12 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.pairplot(df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 161,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 162,
   "metadata": {},
   "outputs": [],
   "source": [
    "X = df[['feature1','feature2']].values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 163,
   "metadata": {},
   "outputs": [],
   "source": [
    "y = df['price'].values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 164,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[ 999.78755752,  999.7660962 ],\n",
       "       [ 998.86161491, 1001.04240315],\n",
       "       [1000.07026691,  998.84401463],\n",
       "       ...,\n",
       "       [1001.45164617,  998.84760554],\n",
       "       [1000.77102275,  998.56285086],\n",
       "       [ 999.2322436 , 1001.45140713]])"
      ]
     },
     "execution_count": 164,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 165,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3, random_state=42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 166,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(700, 2)"
      ]
     },
     "execution_count": 166,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 167,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(300, 2)"
      ]
     },
     "execution_count": 167,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_test.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 168,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Normalize the data\n",
    "from sklearn.preprocessing import MinMaxScaler\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 169,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Instance of the scaler\n",
    "scaler = MinMaxScaler()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 170,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "MinMaxScaler()"
      ]
     },
     "execution_count": 170,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Fit the data\n",
    "scaler.fit(X_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 171,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Transform the data\n",
    "X_train = scaler.transform(X_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 172,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_test = scaler.transform(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 173,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Create the neural network\n",
    "# Add the Sequential model and layers\n",
    "import tensorflow as tf\n",
    "from tensorflow.keras.models import Sequential\n",
    "from tensorflow.keras.layers import Dense, Activation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 174,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Keras model\n",
    "# The forst layer has 4 neurons, the second layer has 2 and the last one has 1\n",
    "#model = Sequential([Dense(4, activation = 'relu'), \n",
    "#                   Dense(2, activation = 'relu'),\n",
    " #                  Dense(1)])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 175,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Another way to create the model.\n",
    "# this makes it easier to edit the model\n",
    "model = Sequential()\n",
    "\n",
    "# Add every layer\n",
    "model.add(Dense(4, activation = 'relu'))\n",
    "model.add(Dense(4, activation = 'relu'))\n",
    "model.add(Dense(4, activation = 'relu'))\n",
    "\n",
    "# This final layer depends on the prediction. In this case one single number\n",
    "model.add(Dense(1))\n",
    "\n",
    "# This compiles the model\n",
    "# Loss changes depending what we want to accomplish\n",
    "# loss = mean squared error = mse\n",
    "model.compile(optimizer = 'rmsprop', loss = 'mse')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 176,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 700 samples\n",
      "Epoch 1/250\n",
      "700/700 [==============================] - 1s 1ms/sample - loss: 256301.3108\n",
      "Epoch 2/250\n",
      "700/700 [==============================] - 0s 56us/sample - loss: 256100.7289\n",
      "Epoch 3/250\n",
      "700/700 [==============================] - 0s 56us/sample - loss: 255858.1760\n",
      "Epoch 4/250\n",
      "700/700 [==============================] - 0s 57us/sample - loss: 255566.8693\n",
      "Epoch 5/250\n",
      "700/700 [==============================] - 0s 58us/sample - loss: 255246.3437\n",
      "Epoch 6/250\n",
      "700/700 [==============================] - 0s 58us/sample - loss: 254889.7630\n",
      "Epoch 7/250\n",
      "700/700 [==============================] - 0s 56us/sample - loss: 254492.8210\n",
      "Epoch 8/250\n",
      "700/700 [==============================] - 0s 60us/sample - loss: 254052.6620\n",
      "Epoch 9/250\n",
      "700/700 [==============================] - 0s 56us/sample - loss: 253565.2518\n",
      "Epoch 10/250\n",
      "700/700 [==============================] - 0s 57us/sample - loss: 253025.9091\n",
      "Epoch 11/250\n",
      "700/700 [==============================] - 0s 59us/sample - loss: 252432.3197\n",
      "Epoch 12/250\n",
      "700/700 [==============================] - 0s 64us/sample - loss: 251784.7539\n",
      "Epoch 13/250\n",
      "700/700 [==============================] - 0s 61us/sample - loss: 251069.9255\n",
      "Epoch 14/250\n",
      "700/700 [==============================] - 0s 58us/sample - loss: 250296.0462\n",
      "Epoch 15/250\n",
      "700/700 [==============================] - 0s 62us/sample - loss: 249451.2331\n",
      "Epoch 16/250\n",
      "700/700 [==============================] - 0s 60us/sample - loss: 248532.4795\n",
      "Epoch 17/250\n",
      "700/700 [==============================] - 0s 61us/sample - loss: 247518.1521\n",
      "Epoch 18/250\n",
      "700/700 [==============================] - 0s 57us/sample - loss: 246411.0802\n",
      "Epoch 19/250\n",
      "700/700 [==============================] - 0s 61us/sample - loss: 245195.2170\n",
      "Epoch 20/250\n",
      "700/700 [==============================] - 0s 58us/sample - loss: 243882.2034\n",
      "Epoch 21/250\n",
      "700/700 [==============================] - 0s 62us/sample - loss: 242462.4672\n",
      "Epoch 22/250\n",
      "700/700 [==============================] - 0s 58us/sample - loss: 240923.1343\n",
      "Epoch 23/250\n",
      "700/700 [==============================] - 0s 63us/sample - loss: 239270.1344\n",
      "Epoch 24/250\n",
      "700/700 [==============================] - 0s 58us/sample - loss: 237477.4034\n",
      "Epoch 25/250\n",
      "700/700 [==============================] - 0s 67us/sample - loss: 235565.6604\n",
      "Epoch 26/250\n",
      "700/700 [==============================] - 0s 70us/sample - loss: 233526.1088\n",
      "Epoch 27/250\n",
      "700/700 [==============================] - 0s 73us/sample - loss: 231339.0092\n",
      "Epoch 28/250\n",
      "700/700 [==============================] - 0s 76us/sample - loss: 228999.5151\n",
      "Epoch 29/250\n",
      "700/700 [==============================] - 0s 72us/sample - loss: 226497.2308\n",
      "Epoch 30/250\n",
      "700/700 [==============================] - 0s 63us/sample - loss: 223846.6647\n",
      "Epoch 31/250\n",
      "700/700 [==============================] - 0s 70us/sample - loss: 221033.9281\n",
      "Epoch 32/250\n",
      "700/700 [==============================] - 0s 69us/sample - loss: 218051.4188\n",
      "Epoch 33/250\n",
      "700/700 [==============================] - 0s 70us/sample - loss: 214910.7974\n",
      "Epoch 34/250\n",
      "700/700 [==============================] - 0s 63us/sample - loss: 211582.9837\n",
      "Epoch 35/250\n",
      "700/700 [==============================] - 0s 64us/sample - loss: 208104.9430\n",
      "Epoch 36/250\n",
      "700/700 [==============================] - 0s 60us/sample - loss: 204428.7979\n",
      "Epoch 37/250\n",
      "700/700 [==============================] - 0s 56us/sample - loss: 200560.0138\n",
      "Epoch 38/250\n",
      "700/700 [==============================] - 0s 58us/sample - loss: 196530.0447\n",
      "Epoch 39/250\n",
      "700/700 [==============================] - 0s 55us/sample - loss: 192289.2798\n",
      "Epoch 40/250\n",
      "700/700 [==============================] - 0s 53us/sample - loss: 187874.8922\n",
      "Epoch 41/250\n",
      "700/700 [==============================] - 0s 55us/sample - loss: 183270.6157\n",
      "Epoch 42/250\n",
      "700/700 [==============================] - 0s 49us/sample - loss: 178476.8934\n",
      "Epoch 43/250\n",
      "700/700 [==============================] - 0s 50us/sample - loss: 173510.3901\n",
      "Epoch 44/250\n",
      "700/700 [==============================] - 0s 48us/sample - loss: 168335.1458\n",
      "Epoch 45/250\n",
      "700/700 [==============================] - 0s 51us/sample - loss: 163007.9980\n",
      "Epoch 46/250\n",
      "700/700 [==============================] - 0s 65us/sample - loss: 157513.8913\n",
      "Epoch 47/250\n",
      "700/700 [==============================] - 0s 68us/sample - loss: 151844.8836\n",
      "Epoch 48/250\n",
      "700/700 [==============================] - 0s 67us/sample - loss: 145992.9478\n",
      "Epoch 49/250\n",
      "700/700 [==============================] - 0s 69us/sample - loss: 139986.6426\n",
      "Epoch 50/250\n",
      "700/700 [==============================] - 0s 71us/sample - loss: 133860.0401\n",
      "Epoch 51/250\n",
      "700/700 [==============================] - 0s 71us/sample - loss: 127571.9142\n",
      "Epoch 52/250\n",
      "700/700 [==============================] - 0s 69us/sample - loss: 121192.1574\n",
      "Epoch 53/250\n",
      "700/700 [==============================] - 0s 69us/sample - loss: 114704.2312\n",
      "Epoch 54/250\n",
      "700/700 [==============================] - 0s 68us/sample - loss: 108098.9265\n",
      "Epoch 55/250\n",
      "700/700 [==============================] - 0s 67us/sample - loss: 101465.4235\n",
      "Epoch 56/250\n",
      "700/700 [==============================] - 0s 69us/sample - loss: 94747.0492\n",
      "Epoch 57/250\n",
      "700/700 [==============================] - 0s 68us/sample - loss: 88037.8175\n",
      "Epoch 58/250\n",
      "700/700 [==============================] - 0s 70us/sample - loss: 81304.7759\n",
      "Epoch 59/250\n",
      "700/700 [==============================] - 0s 66us/sample - loss: 74595.8971\n",
      "Epoch 60/250\n",
      "700/700 [==============================] - 0s 68us/sample - loss: 67950.8492\n",
      "Epoch 61/250\n",
      "700/700 [==============================] - 0s 65us/sample - loss: 61359.7955\n",
      "Epoch 62/250\n",
      "700/700 [==============================] - 0s 69us/sample - loss: 54903.1852\n",
      "Epoch 63/250\n",
      "700/700 [==============================] - 0s 64us/sample - loss: 48593.1218\n",
      "Epoch 64/250\n",
      "700/700 [==============================] - 0s 64us/sample - loss: 42474.6518\n",
      "Epoch 65/250\n",
      "700/700 [==============================] - 0s 61us/sample - loss: 36602.2838\n",
      "Epoch 66/250\n",
      "700/700 [==============================] - 0s 55us/sample - loss: 31039.6335\n",
      "Epoch 67/250\n",
      "700/700 [==============================] - 0s 53us/sample - loss: 25770.1604\n",
      "Epoch 68/250\n",
      "700/700 [==============================] - 0s 50us/sample - loss: 20904.4042\n",
      "Epoch 69/250\n",
      "700/700 [==============================] - 0s 46us/sample - loss: 16468.4929\n",
      "Epoch 70/250\n",
      "700/700 [==============================] - 0s 47us/sample - loss: 12531.3102\n",
      "Epoch 71/250\n",
      "700/700 [==============================] - 0s 47us/sample - loss: 9156.6278\n",
      "Epoch 72/250\n",
      "700/700 [==============================] - 0s 50us/sample - loss: 6374.2735\n",
      "Epoch 73/250\n",
      "700/700 [==============================] - 0s 53us/sample - loss: 4235.3418\n",
      "Epoch 74/250\n",
      "700/700 [==============================] - 0s 55us/sample - loss: 2738.6112\n",
      "Epoch 75/250\n",
      "700/700 [==============================] - 0s 50us/sample - loss: 1930.7964\n",
      "Epoch 76/250\n",
      "700/700 [==============================] - 0s 45us/sample - loss: 1630.7870\n",
      "Epoch 77/250\n",
      "700/700 [==============================] - 0s 45us/sample - loss: 1570.3565\n",
      "Epoch 78/250\n",
      "700/700 [==============================] - 0s 45us/sample - loss: 1541.6147\n",
      "Epoch 79/250\n",
      "700/700 [==============================] - 0s 45us/sample - loss: 1506.0142\n",
      "Epoch 80/250\n",
      "700/700 [==============================] - 0s 46us/sample - loss: 1474.8395\n",
      "Epoch 81/250\n",
      "700/700 [==============================] - 0s 45us/sample - loss: 1443.6334\n",
      "Epoch 82/250\n",
      "700/700 [==============================] - 0s 45us/sample - loss: 1413.4439\n",
      "Epoch 83/250\n",
      "700/700 [==============================] - 0s 45us/sample - loss: 1381.8411\n",
      "Epoch 84/250\n",
      "700/700 [==============================] - 0s 46us/sample - loss: 1353.4282\n",
      "Epoch 85/250\n",
      "700/700 [==============================] - 0s 46us/sample - loss: 1323.7415\n",
      "Epoch 86/250\n",
      "700/700 [==============================] - 0s 46us/sample - loss: 1292.0913\n",
      "Epoch 87/250\n",
      "700/700 [==============================] - 0s 46us/sample - loss: 1261.5963\n",
      "Epoch 88/250\n",
      "700/700 [==============================] - 0s 52us/sample - loss: 1232.5017\n",
      "Epoch 89/250\n",
      "700/700 [==============================] - 0s 67us/sample - loss: 1202.8668\n",
      "Epoch 90/250\n",
      "700/700 [==============================] - 0s 67us/sample - loss: 1170.9230\n",
      "Epoch 91/250\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "700/700 [==============================] - 0s 69us/sample - loss: 1141.8230\n",
      "Epoch 92/250\n",
      "700/700 [==============================] - 0s 69us/sample - loss: 1113.2946\n",
      "Epoch 93/250\n",
      "700/700 [==============================] - 0s 77us/sample - loss: 1091.5366\n",
      "Epoch 94/250\n",
      "700/700 [==============================] - 0s 76us/sample - loss: 1060.9329\n",
      "Epoch 95/250\n",
      "700/700 [==============================] - 0s 72us/sample - loss: 1033.1851\n",
      "Epoch 96/250\n",
      "700/700 [==============================] - 0s 70us/sample - loss: 1005.3119\n",
      "Epoch 97/250\n",
      "700/700 [==============================] - 0s 68us/sample - loss: 984.1996\n",
      "Epoch 98/250\n",
      "700/700 [==============================] - 0s 70us/sample - loss: 955.0385\n",
      "Epoch 99/250\n",
      "700/700 [==============================] - 0s 69us/sample - loss: 929.1960\n",
      "Epoch 100/250\n",
      "700/700 [==============================] - 0s 65us/sample - loss: 907.3498\n",
      "Epoch 101/250\n",
      "700/700 [==============================] - 0s 62us/sample - loss: 878.4089\n",
      "Epoch 102/250\n",
      "700/700 [==============================] - 0s 65us/sample - loss: 855.3708\n",
      "Epoch 103/250\n",
      "700/700 [==============================] - 0s 70us/sample - loss: 826.4969\n",
      "Epoch 104/250\n",
      "700/700 [==============================] - 0s 65us/sample - loss: 803.3701\n",
      "Epoch 105/250\n",
      "700/700 [==============================] - 0s 65us/sample - loss: 777.1073\n",
      "Epoch 106/250\n",
      "700/700 [==============================] - 0s 63us/sample - loss: 752.7016\n",
      "Epoch 107/250\n",
      "700/700 [==============================] - 0s 66us/sample - loss: 726.3103\n",
      "Epoch 108/250\n",
      "700/700 [==============================] - 0s 64us/sample - loss: 702.3781\n",
      "Epoch 109/250\n",
      "700/700 [==============================] - 0s 66us/sample - loss: 680.7890\n",
      "Epoch 110/250\n",
      "700/700 [==============================] - 0s 69us/sample - loss: 658.4421\n",
      "Epoch 111/250\n",
      "700/700 [==============================] - 0s 64us/sample - loss: 637.7595\n",
      "Epoch 112/250\n",
      "700/700 [==============================] - 0s 61us/sample - loss: 616.0853\n",
      "Epoch 113/250\n",
      "700/700 [==============================] - 0s 55us/sample - loss: 593.3358\n",
      "Epoch 114/250\n",
      "700/700 [==============================] - 0s 50us/sample - loss: 572.1539\n",
      "Epoch 115/250\n",
      "700/700 [==============================] - 0s 55us/sample - loss: 552.8028\n",
      "Epoch 116/250\n",
      "700/700 [==============================] - 0s 52us/sample - loss: 531.6440\n",
      "Epoch 117/250\n",
      "700/700 [==============================] - 0s 44us/sample - loss: 509.8523\n",
      "Epoch 118/250\n",
      "700/700 [==============================] - 0s 44us/sample - loss: 495.2358\n",
      "Epoch 119/250\n",
      "700/700 [==============================] - 0s 44us/sample - loss: 473.9457\n",
      "Epoch 120/250\n",
      "700/700 [==============================] - 0s 45us/sample - loss: 453.6896\n",
      "Epoch 121/250\n",
      "700/700 [==============================] - 0s 44us/sample - loss: 434.6598\n",
      "Epoch 122/250\n",
      "700/700 [==============================] - 0s 56us/sample - loss: 414.3410\n",
      "Epoch 123/250\n",
      "700/700 [==============================] - 0s 57us/sample - loss: 395.4955\n",
      "Epoch 124/250\n",
      "700/700 [==============================] - 0s 56us/sample - loss: 375.0707\n",
      "Epoch 125/250\n",
      "700/700 [==============================] - 0s 58us/sample - loss: 357.5712\n",
      "Epoch 126/250\n",
      "700/700 [==============================] - 0s 57us/sample - loss: 338.9402\n",
      "Epoch 127/250\n",
      "700/700 [==============================] - 0s 57us/sample - loss: 321.3305\n",
      "Epoch 128/250\n",
      "700/700 [==============================] - 0s 55us/sample - loss: 305.2194\n",
      "Epoch 129/250\n",
      "700/700 [==============================] - 0s 59us/sample - loss: 290.0184\n",
      "Epoch 130/250\n",
      "700/700 [==============================] - 0s 54us/sample - loss: 274.3552\n",
      "Epoch 131/250\n",
      "700/700 [==============================] - 0s 55us/sample - loss: 258.3452\n",
      "Epoch 132/250\n",
      "700/700 [==============================] - 0s 49us/sample - loss: 243.5834\n",
      "Epoch 133/250\n",
      "700/700 [==============================] - 0s 47us/sample - loss: 228.1444\n",
      "Epoch 134/250\n",
      "700/700 [==============================] - 0s 47us/sample - loss: 210.6655\n",
      "Epoch 135/250\n",
      "700/700 [==============================] - 0s 54us/sample - loss: 198.1901\n",
      "Epoch 136/250\n",
      "700/700 [==============================] - 0s 55us/sample - loss: 185.0087\n",
      "Epoch 137/250\n",
      "700/700 [==============================] - 0s 60us/sample - loss: 173.3965\n",
      "Epoch 138/250\n",
      "700/700 [==============================] - 0s 54us/sample - loss: 161.7362\n",
      "Epoch 139/250\n",
      "700/700 [==============================] - 0s 54us/sample - loss: 150.6162\n",
      "Epoch 140/250\n",
      "700/700 [==============================] - 0s 49us/sample - loss: 139.7221\n",
      "Epoch 141/250\n",
      "700/700 [==============================] - 0s 45us/sample - loss: 130.4378\n",
      "Epoch 142/250\n",
      "700/700 [==============================] - 0s 44us/sample - loss: 121.9848\n",
      "Epoch 143/250\n",
      "700/700 [==============================] - 0s 44us/sample - loss: 112.8443\n",
      "Epoch 144/250\n",
      "700/700 [==============================] - 0s 51us/sample - loss: 106.0970\n",
      "Epoch 145/250\n",
      "700/700 [==============================] - 0s 55us/sample - loss: 97.2701\n",
      "Epoch 146/250\n",
      "700/700 [==============================] - 0s 47us/sample - loss: 88.6131\n",
      "Epoch 147/250\n",
      "700/700 [==============================] - 0s 45us/sample - loss: 80.9536\n",
      "Epoch 148/250\n",
      "700/700 [==============================] - 0s 47us/sample - loss: 74.8259\n",
      "Epoch 149/250\n",
      "700/700 [==============================] - 0s 45us/sample - loss: 69.5459\n",
      "Epoch 150/250\n",
      "700/700 [==============================] - 0s 46us/sample - loss: 64.3472\n",
      "Epoch 151/250\n",
      "700/700 [==============================] - 0s 48us/sample - loss: 58.9562\n",
      "Epoch 152/250\n",
      "700/700 [==============================] - 0s 45us/sample - loss: 55.2090\n",
      "Epoch 153/250\n",
      "700/700 [==============================] - 0s 44us/sample - loss: 51.4039\n",
      "Epoch 154/250\n",
      "700/700 [==============================] - 0s 45us/sample - loss: 47.7305\n",
      "Epoch 155/250\n",
      "700/700 [==============================] - 0s 48us/sample - loss: 44.6452\n",
      "Epoch 156/250\n",
      "700/700 [==============================] - 0s 49us/sample - loss: 41.8011\n",
      "Epoch 157/250\n",
      "700/700 [==============================] - 0s 50us/sample - loss: 39.8495\n",
      "Epoch 158/250\n",
      "700/700 [==============================] - 0s 56us/sample - loss: 37.1591\n",
      "Epoch 159/250\n",
      "700/700 [==============================] - 0s 65us/sample - loss: 35.0836\n",
      "Epoch 160/250\n",
      "700/700 [==============================] - 0s 75us/sample - loss: 33.0623\n",
      "Epoch 161/250\n",
      "700/700 [==============================] - 0s 55us/sample - loss: 32.0438\n",
      "Epoch 162/250\n",
      "700/700 [==============================] - 0s 69us/sample - loss: 31.1198\n",
      "Epoch 163/250\n",
      "700/700 [==============================] - 0s 62us/sample - loss: 29.6004\n",
      "Epoch 164/250\n",
      "700/700 [==============================] - 0s 63us/sample - loss: 28.5741\n",
      "Epoch 165/250\n",
      "700/700 [==============================] - 0s 59us/sample - loss: 28.0328\n",
      "Epoch 166/250\n",
      "700/700 [==============================] - 0s 60us/sample - loss: 27.9162\n",
      "Epoch 167/250\n",
      "700/700 [==============================] - 0s 50us/sample - loss: 26.6109\n",
      "Epoch 168/250\n",
      "700/700 [==============================] - 0s 74us/sample - loss: 26.1990\n",
      "Epoch 169/250\n",
      "700/700 [==============================] - 0s 77us/sample - loss: 25.8745\n",
      "Epoch 170/250\n",
      "700/700 [==============================] - 0s 81us/sample - loss: 26.0709\n",
      "Epoch 171/250\n",
      "700/700 [==============================] - 0s 58us/sample - loss: 25.4341\n",
      "Epoch 172/250\n",
      "700/700 [==============================] - 0s 57us/sample - loss: 25.2585\n",
      "Epoch 173/250\n",
      "700/700 [==============================] - 0s 53us/sample - loss: 25.1435\n",
      "Epoch 174/250\n",
      "700/700 [==============================] - 0s 51us/sample - loss: 25.0415\n",
      "Epoch 175/250\n",
      "700/700 [==============================] - 0s 68us/sample - loss: 25.1966\n",
      "Epoch 176/250\n",
      "700/700 [==============================] - 0s 610us/sample - loss: 24.7159\n",
      "Epoch 177/250\n",
      "700/700 [==============================] - 0s 74us/sample - loss: 24.8331\n",
      "Epoch 178/250\n",
      "700/700 [==============================] - 0s 67us/sample - loss: 24.7384\n",
      "Epoch 179/250\n",
      "700/700 [==============================] - 0s 76us/sample - loss: 24.4352\n",
      "Epoch 180/250\n",
      "700/700 [==============================] - 0s 80us/sample - loss: 24.8853\n",
      "Epoch 181/250\n",
      "700/700 [==============================] - 0s 77us/sample - loss: 25.0524\n",
      "Epoch 182/250\n",
      "700/700 [==============================] - 0s 82us/sample - loss: 24.0645\n",
      "Epoch 183/250\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "700/700 [==============================] - 0s 77us/sample - loss: 24.4651\n",
      "Epoch 184/250\n",
      "700/700 [==============================] - 0s 68us/sample - loss: 24.4809\n",
      "Epoch 185/250\n",
      "700/700 [==============================] - 0s 66us/sample - loss: 24.1322\n",
      "Epoch 186/250\n",
      "700/700 [==============================] - 0s 65us/sample - loss: 24.5405\n",
      "Epoch 187/250\n",
      "700/700 [==============================] - 0s 69us/sample - loss: 24.3521\n",
      "Epoch 188/250\n",
      "700/700 [==============================] - 0s 64us/sample - loss: 24.6390\n",
      "Epoch 189/250\n",
      "700/700 [==============================] - 0s 66us/sample - loss: 24.4482\n",
      "Epoch 190/250\n",
      "700/700 [==============================] - 0s 67us/sample - loss: 24.1832\n",
      "Epoch 191/250\n",
      "700/700 [==============================] - 0s 66us/sample - loss: 24.2451\n",
      "Epoch 192/250\n",
      "700/700 [==============================] - 0s 68us/sample - loss: 24.1822\n",
      "Epoch 193/250\n",
      "700/700 [==============================] - 0s 68us/sample - loss: 24.3980\n",
      "Epoch 194/250\n",
      "700/700 [==============================] - 0s 67us/sample - loss: 24.4151\n",
      "Epoch 195/250\n",
      "700/700 [==============================] - 0s 59us/sample - loss: 24.3799\n",
      "Epoch 196/250\n",
      "700/700 [==============================] - 0s 58us/sample - loss: 24.5918\n",
      "Epoch 197/250\n",
      "700/700 [==============================] - 0s 56us/sample - loss: 24.1994\n",
      "Epoch 198/250\n",
      "700/700 [==============================] - 0s 63us/sample - loss: 24.3746\n",
      "Epoch 199/250\n",
      "700/700 [==============================] - 0s 61us/sample - loss: 24.2127\n",
      "Epoch 200/250\n",
      "700/700 [==============================] - 0s 60us/sample - loss: 24.1912\n",
      "Epoch 201/250\n",
      "700/700 [==============================] - 0s 59us/sample - loss: 24.6165\n",
      "Epoch 202/250\n",
      "700/700 [==============================] - 0s 62us/sample - loss: 24.4447\n",
      "Epoch 203/250\n",
      "700/700 [==============================] - 0s 59us/sample - loss: 23.9872\n",
      "Epoch 204/250\n",
      "700/700 [==============================] - 0s 59us/sample - loss: 24.3251\n",
      "Epoch 205/250\n",
      "700/700 [==============================] - 0s 63us/sample - loss: 24.4224\n",
      "Epoch 206/250\n",
      "700/700 [==============================] - 0s 59us/sample - loss: 24.5090\n",
      "Epoch 207/250\n",
      "700/700 [==============================] - 0s 59us/sample - loss: 24.8013\n",
      "Epoch 208/250\n",
      "700/700 [==============================] - 0s 59us/sample - loss: 24.1906\n",
      "Epoch 209/250\n",
      "700/700 [==============================] - 0s 63us/sample - loss: 24.3160\n",
      "Epoch 210/250\n",
      "700/700 [==============================] - 0s 62us/sample - loss: 23.9576\n",
      "Epoch 211/250\n",
      "700/700 [==============================] - 0s 59us/sample - loss: 24.5969\n",
      "Epoch 212/250\n",
      "700/700 [==============================] - 0s 60us/sample - loss: 24.4533\n",
      "Epoch 213/250\n",
      "700/700 [==============================] - 0s 59us/sample - loss: 24.5294\n",
      "Epoch 214/250\n",
      "700/700 [==============================] - 0s 59us/sample - loss: 24.4543\n",
      "Epoch 215/250\n",
      "700/700 [==============================] - 0s 59us/sample - loss: 24.3968\n",
      "Epoch 216/250\n",
      "700/700 [==============================] - 0s 63us/sample - loss: 24.4490\n",
      "Epoch 217/250\n",
      "700/700 [==============================] - 0s 61us/sample - loss: 24.1308\n",
      "Epoch 218/250\n",
      "700/700 [==============================] - 0s 59us/sample - loss: 24.5217\n",
      "Epoch 219/250\n",
      "700/700 [==============================] - 0s 57us/sample - loss: 24.3265\n",
      "Epoch 220/250\n",
      "700/700 [==============================] - 0s 60us/sample - loss: 24.4230\n",
      "Epoch 221/250\n",
      "700/700 [==============================] - 0s 62us/sample - loss: 24.2372\n",
      "Epoch 222/250\n",
      "700/700 [==============================] - 0s 60us/sample - loss: 24.3861\n",
      "Epoch 223/250\n",
      "700/700 [==============================] - 0s 63us/sample - loss: 23.9855\n",
      "Epoch 224/250\n",
      "700/700 [==============================] - 0s 61us/sample - loss: 24.4124\n",
      "Epoch 225/250\n",
      "700/700 [==============================] - 0s 60us/sample - loss: 24.4458\n",
      "Epoch 226/250\n",
      "700/700 [==============================] - 0s 61us/sample - loss: 24.5218\n",
      "Epoch 227/250\n",
      "700/700 [==============================] - 0s 59us/sample - loss: 24.6756\n",
      "Epoch 228/250\n",
      "700/700 [==============================] - 0s 62us/sample - loss: 24.4607\n",
      "Epoch 229/250\n",
      "700/700 [==============================] - 0s 61us/sample - loss: 24.0772\n",
      "Epoch 230/250\n",
      "700/700 [==============================] - 0s 57us/sample - loss: 24.3001\n",
      "Epoch 231/250\n",
      "700/700 [==============================] - 0s 62us/sample - loss: 24.1296\n",
      "Epoch 232/250\n",
      "700/700 [==============================] - 0s 62us/sample - loss: 24.1135\n",
      "Epoch 233/250\n",
      "700/700 [==============================] - 0s 63us/sample - loss: 24.6305\n",
      "Epoch 234/250\n",
      "700/700 [==============================] - 0s 61us/sample - loss: 24.1788\n",
      "Epoch 235/250\n",
      "700/700 [==============================] - 0s 60us/sample - loss: 24.4423\n",
      "Epoch 236/250\n",
      "700/700 [==============================] - 0s 61us/sample - loss: 24.4544\n",
      "Epoch 237/250\n",
      "700/700 [==============================] - 0s 59us/sample - loss: 24.2952\n",
      "Epoch 238/250\n",
      "700/700 [==============================] - 0s 64us/sample - loss: 24.3037\n",
      "Epoch 239/250\n",
      "700/700 [==============================] - 0s 61us/sample - loss: 24.1162\n",
      "Epoch 240/250\n",
      "700/700 [==============================] - 0s 58us/sample - loss: 24.1109\n",
      "Epoch 241/250\n",
      "700/700 [==============================] - 0s 57us/sample - loss: 24.3735\n",
      "Epoch 242/250\n",
      "700/700 [==============================] - 0s 57us/sample - loss: 24.3656\n",
      "Epoch 243/250\n",
      "700/700 [==============================] - 0s 61us/sample - loss: 24.3100\n",
      "Epoch 244/250\n",
      "700/700 [==============================] - 0s 61us/sample - loss: 24.2896\n",
      "Epoch 245/250\n",
      "700/700 [==============================] - 0s 60us/sample - loss: 24.2993\n",
      "Epoch 246/250\n",
      "700/700 [==============================] - 0s 60us/sample - loss: 24.5065\n",
      "Epoch 247/250\n",
      "700/700 [==============================] - 0s 63us/sample - loss: 24.4293\n",
      "Epoch 248/250\n",
      "700/700 [==============================] - 0s 61us/sample - loss: 24.2913\n",
      "Epoch 249/250\n",
      "700/700 [==============================] - 0s 58us/sample - loss: 24.0273\n",
      "Epoch 250/250\n",
      "700/700 [==============================] - 0s 61us/sample - loss: 24.5550\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<tensorflow.python.keras.callbacks.History at 0x7fb99761bdd0>"
      ]
     },
     "execution_count": 176,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Train the model\n",
    "model.fit(X_train, y_train, epochs = 250)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 177,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Look at the training history\n",
    "loss_df = pd.DataFrame(model.history.history['loss'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 178,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:>"
      ]
     },
     "execution_count": 178,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "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": [
    "# Plot the losses\n",
    "loss_df.plot()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 179,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "26.10526351928711"
      ]
     },
     "execution_count": 179,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Model Evaluation\n",
    "# Method 1\n",
    "model.evaluate(X_test, y_test, verbose = 0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 180,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "24.122437210083007"
      ]
     },
     "execution_count": 180,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.evaluate(X_train, y_train, verbose = 0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 181,
   "metadata": {},
   "outputs": [],
   "source": [
    "#\n",
    "test_predictions = model.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 182,
   "metadata": {},
   "outputs": [],
   "source": [
    "test_predictions = pd.Series(test_predictions.reshape(300,))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 183,
   "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>Test True Y</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>402.296319</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>624.156198</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>582.455066</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>578.588606</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>371.224104</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>295</th>\n",
       "      <td>525.704657</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>296</th>\n",
       "      <td>502.909473</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>297</th>\n",
       "      <td>612.727910</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>298</th>\n",
       "      <td>417.569725</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>299</th>\n",
       "      <td>410.538250</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>300 rows × 1 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "     Test True Y\n",
       "0     402.296319\n",
       "1     624.156198\n",
       "2     582.455066\n",
       "3     578.588606\n",
       "4     371.224104\n",
       "..           ...\n",
       "295   525.704657\n",
       "296   502.909473\n",
       "297   612.727910\n",
       "298   417.569725\n",
       "299   410.538250\n",
       "\n",
       "[300 rows x 1 columns]"
      ]
     },
     "execution_count": 183,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred_df = pd.DataFrame(y_test, columns=['Test True Y'])\n",
    "pred_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 184,
   "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>Test True Y</th>\n",
       "      <th>0</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>402.296319</td>\n",
       "      <td>406.024750</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>624.156198</td>\n",
       "      <td>624.744690</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>582.455066</td>\n",
       "      <td>593.277466</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>578.588606</td>\n",
       "      <td>573.301880</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>371.224104</td>\n",
       "      <td>367.232269</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>295</th>\n",
       "      <td>525.704657</td>\n",
       "      <td>530.030396</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>296</th>\n",
       "      <td>502.909473</td>\n",
       "      <td>516.339783</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>297</th>\n",
       "      <td>612.727910</td>\n",
       "      <td>610.830933</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>298</th>\n",
       "      <td>417.569725</td>\n",
       "      <td>417.732147</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>299</th>\n",
       "      <td>410.538250</td>\n",
       "      <td>411.609070</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>300 rows × 2 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "     Test True Y           0\n",
       "0     402.296319  406.024750\n",
       "1     624.156198  624.744690\n",
       "2     582.455066  593.277466\n",
       "3     578.588606  573.301880\n",
       "4     371.224104  367.232269\n",
       "..           ...         ...\n",
       "295   525.704657  530.030396\n",
       "296   502.909473  516.339783\n",
       "297   612.727910  610.830933\n",
       "298   417.569725  417.732147\n",
       "299   410.538250  411.609070\n",
       "\n",
       "[300 rows x 2 columns]"
      ]
     },
     "execution_count": 184,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred_df = pd.concat([pred_df, test_predictions], axis=1)\n",
    "pred_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 185,
   "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>Test True Y</th>\n",
       "      <th>Model Predictions</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>402.296319</td>\n",
       "      <td>406.024750</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>624.156198</td>\n",
       "      <td>624.744690</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>582.455066</td>\n",
       "      <td>593.277466</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>578.588606</td>\n",
       "      <td>573.301880</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>371.224104</td>\n",
       "      <td>367.232269</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>295</th>\n",
       "      <td>525.704657</td>\n",
       "      <td>530.030396</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>296</th>\n",
       "      <td>502.909473</td>\n",
       "      <td>516.339783</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>297</th>\n",
       "      <td>612.727910</td>\n",
       "      <td>610.830933</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>298</th>\n",
       "      <td>417.569725</td>\n",
       "      <td>417.732147</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>299</th>\n",
       "      <td>410.538250</td>\n",
       "      <td>411.609070</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>300 rows × 2 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "     Test True Y  Model Predictions\n",
       "0     402.296319         406.024750\n",
       "1     624.156198         624.744690\n",
       "2     582.455066         593.277466\n",
       "3     578.588606         573.301880\n",
       "4     371.224104         367.232269\n",
       "..           ...                ...\n",
       "295   525.704657         530.030396\n",
       "296   502.909473         516.339783\n",
       "297   612.727910         610.830933\n",
       "298   417.569725         417.732147\n",
       "299   410.538250         411.609070\n",
       "\n",
       "[300 rows x 2 columns]"
      ]
     },
     "execution_count": 185,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred_df.columns = ['Test True Y', 'Model Predictions']\n",
    "pred_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 186,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:xlabel='Test True Y', ylabel='Model Predictions'>"
      ]
     },
     "execution_count": 186,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "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": [
    "sns.scatterplot(x = 'Test True Y', y = 'Model Predictions', data=pred_df) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 187,
   "metadata": {},
   "outputs": [],
   "source": [
    "#\n",
    "from sklearn.metrics import mean_absolute_error,mean_squared_error"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 188,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "4.132913665414081"
      ]
     },
     "execution_count": 188,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mean_absolute_error(pred_df['Test True Y'], pred_df['Model Predictions'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 189,
   "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>price</th>\n",
       "      <th>feature1</th>\n",
       "      <th>feature2</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>1000.000000</td>\n",
       "      <td>1000.000000</td>\n",
       "      <td>1000.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>498.673029</td>\n",
       "      <td>1000.014171</td>\n",
       "      <td>999.979847</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>93.785431</td>\n",
       "      <td>0.974018</td>\n",
       "      <td>0.948330</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>223.346793</td>\n",
       "      <td>997.058347</td>\n",
       "      <td>996.995651</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>433.025732</td>\n",
       "      <td>999.332068</td>\n",
       "      <td>999.316106</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>502.382117</td>\n",
       "      <td>1000.009915</td>\n",
       "      <td>1000.002243</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>564.921588</td>\n",
       "      <td>1000.637580</td>\n",
       "      <td>1000.645380</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>774.407854</td>\n",
       "      <td>1003.207934</td>\n",
       "      <td>1002.666308</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             price     feature1     feature2\n",
       "count  1000.000000  1000.000000  1000.000000\n",
       "mean    498.673029  1000.014171   999.979847\n",
       "std      93.785431     0.974018     0.948330\n",
       "min     223.346793   997.058347   996.995651\n",
       "25%     433.025732   999.332068   999.316106\n",
       "50%     502.382117  1000.009915  1000.002243\n",
       "75%     564.921588  1000.637580  1000.645380\n",
       "max     774.407854  1003.207934  1002.666308"
      ]
     },
     "execution_count": 189,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 190,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "26.105265718830925"
      ]
     },
     "execution_count": 190,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mean_squared_error(pred_df['Test True Y'], pred_df['Model Predictions'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 191,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Predicting on brand new data\n",
    "new_gem = [[998,1000]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 192,
   "metadata": {},
   "outputs": [],
   "source": [
    "new_gem = scaler.transform(new_gem)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 193,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[420.41916]], dtype=float32)"
      ]
     },
     "execution_count": 193,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.predict(new_gem)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 194,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Save the model\n",
    "from tensorflow.keras.models import load_model\n",
    "model.save('my_gem_model.h5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 195,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:Sequential models without an `input_shape` passed to the first layer cannot reload their optimizer state. As a result, your model isstarting with a freshly initialized optimizer.\n"
     ]
    }
   ],
   "source": [
    "later_model = load_model('my_gem_model.h5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 196,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[420.41916]], dtype=float32)"
      ]
     },
     "execution_count": 196,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "later_model.predict(new_gem)"
   ]
  },
  {
   "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": "3.7.9"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}