Time series Forecasting using feed forward neural networks.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import pandas as pd\n",
    "import math\n",
    "\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Dense\n",
    "from sklearn.preprocessing import MinMaxScaler\n",
    "from sklearn.metrics import mean_squared_error"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "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>SUBDIVISION</th>\n",
       "      <th>YEAR</th>\n",
       "      <th>JAN</th>\n",
       "      <th>FEB</th>\n",
       "      <th>MAR</th>\n",
       "      <th>APR</th>\n",
       "      <th>MAY</th>\n",
       "      <th>JUN</th>\n",
       "      <th>JUL</th>\n",
       "      <th>AUG</th>\n",
       "      <th>SEP</th>\n",
       "      <th>OCT</th>\n",
       "      <th>NOV</th>\n",
       "      <th>DEC</th>\n",
       "      <th>ANNUAL</th>\n",
       "      <th>Jan-Feb</th>\n",
       "      <th>Mar-May</th>\n",
       "      <th>Jun-Sep</th>\n",
       "      <th>Oct-Dec</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>ANDAMAN &amp; NICOBAR ISLANDS</td>\n",
       "      <td>1901</td>\n",
       "      <td>49.2</td>\n",
       "      <td>87.1</td>\n",
       "      <td>29.2</td>\n",
       "      <td>2.3</td>\n",
       "      <td>528.8</td>\n",
       "      <td>517.5</td>\n",
       "      <td>365.1</td>\n",
       "      <td>481.1</td>\n",
       "      <td>332.6</td>\n",
       "      <td>388.5</td>\n",
       "      <td>558.2</td>\n",
       "      <td>33.6</td>\n",
       "      <td>3373.2</td>\n",
       "      <td>136.3</td>\n",
       "      <td>560.3</td>\n",
       "      <td>1696.3</td>\n",
       "      <td>980.3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>ANDAMAN &amp; NICOBAR ISLANDS</td>\n",
       "      <td>1902</td>\n",
       "      <td>0.0</td>\n",
       "      <td>159.8</td>\n",
       "      <td>12.2</td>\n",
       "      <td>0.0</td>\n",
       "      <td>446.1</td>\n",
       "      <td>537.1</td>\n",
       "      <td>228.9</td>\n",
       "      <td>753.7</td>\n",
       "      <td>666.2</td>\n",
       "      <td>197.2</td>\n",
       "      <td>359.0</td>\n",
       "      <td>160.5</td>\n",
       "      <td>3520.7</td>\n",
       "      <td>159.8</td>\n",
       "      <td>458.3</td>\n",
       "      <td>2185.9</td>\n",
       "      <td>716.7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>ANDAMAN &amp; NICOBAR ISLANDS</td>\n",
       "      <td>1903</td>\n",
       "      <td>12.7</td>\n",
       "      <td>144.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>235.1</td>\n",
       "      <td>479.9</td>\n",
       "      <td>728.4</td>\n",
       "      <td>326.7</td>\n",
       "      <td>339.0</td>\n",
       "      <td>181.2</td>\n",
       "      <td>284.4</td>\n",
       "      <td>225.0</td>\n",
       "      <td>2957.4</td>\n",
       "      <td>156.7</td>\n",
       "      <td>236.1</td>\n",
       "      <td>1874.0</td>\n",
       "      <td>690.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>ANDAMAN &amp; NICOBAR ISLANDS</td>\n",
       "      <td>1904</td>\n",
       "      <td>9.4</td>\n",
       "      <td>14.7</td>\n",
       "      <td>0.0</td>\n",
       "      <td>202.4</td>\n",
       "      <td>304.5</td>\n",
       "      <td>495.1</td>\n",
       "      <td>502.0</td>\n",
       "      <td>160.1</td>\n",
       "      <td>820.4</td>\n",
       "      <td>222.2</td>\n",
       "      <td>308.7</td>\n",
       "      <td>40.1</td>\n",
       "      <td>3079.6</td>\n",
       "      <td>24.1</td>\n",
       "      <td>506.9</td>\n",
       "      <td>1977.6</td>\n",
       "      <td>571.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>ANDAMAN &amp; NICOBAR ISLANDS</td>\n",
       "      <td>1905</td>\n",
       "      <td>1.3</td>\n",
       "      <td>0.0</td>\n",
       "      <td>3.3</td>\n",
       "      <td>26.9</td>\n",
       "      <td>279.5</td>\n",
       "      <td>628.7</td>\n",
       "      <td>368.7</td>\n",
       "      <td>330.5</td>\n",
       "      <td>297.0</td>\n",
       "      <td>260.7</td>\n",
       "      <td>25.4</td>\n",
       "      <td>344.7</td>\n",
       "      <td>2566.7</td>\n",
       "      <td>1.3</td>\n",
       "      <td>309.7</td>\n",
       "      <td>1624.9</td>\n",
       "      <td>630.8</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                 SUBDIVISION  YEAR   JAN    FEB   MAR    APR    MAY    JUN  \\\n",
       "0  ANDAMAN & NICOBAR ISLANDS  1901  49.2   87.1  29.2    2.3  528.8  517.5   \n",
       "1  ANDAMAN & NICOBAR ISLANDS  1902   0.0  159.8  12.2    0.0  446.1  537.1   \n",
       "2  ANDAMAN & NICOBAR ISLANDS  1903  12.7  144.0   0.0    1.0  235.1  479.9   \n",
       "3  ANDAMAN & NICOBAR ISLANDS  1904   9.4   14.7   0.0  202.4  304.5  495.1   \n",
       "4  ANDAMAN & NICOBAR ISLANDS  1905   1.3    0.0   3.3   26.9  279.5  628.7   \n",
       "\n",
       "     JUL    AUG    SEP    OCT    NOV    DEC  ANNUAL  Jan-Feb  Mar-May  \\\n",
       "0  365.1  481.1  332.6  388.5  558.2   33.6  3373.2    136.3    560.3   \n",
       "1  228.9  753.7  666.2  197.2  359.0  160.5  3520.7    159.8    458.3   \n",
       "2  728.4  326.7  339.0  181.2  284.4  225.0  2957.4    156.7    236.1   \n",
       "3  502.0  160.1  820.4  222.2  308.7   40.1  3079.6     24.1    506.9   \n",
       "4  368.7  330.5  297.0  260.7   25.4  344.7  2566.7      1.3    309.7   \n",
       "\n",
       "   Jun-Sep  Oct-Dec  \n",
       "0   1696.3    980.3  \n",
       "1   2185.9    716.7  \n",
       "2   1874.0    690.6  \n",
       "3   1977.6    571.0  \n",
       "4   1624.9    630.8  "
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "path = 'D:/Datasets/flood dataset - kaggle/flood prediction/datasets_1686_3183_rainfall in india 1901-2015.csv'\n",
    "dataframe = pd.read_csv(path, na_values=['na', '--'])\n",
    "dataframe.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Filling NULL with mean of each values of column, because nun values will cause error in the following way\n",
    "dataframe['JAN'].fillna((dataframe['JAN'].mean()), inplace=True)\n",
    "dataframe['FEB'].fillna((dataframe['FEB'].mean()), inplace=True)\n",
    "dataframe['MAR'].fillna((dataframe['MAR'].mean()), inplace=True)\n",
    "dataframe['APR'].fillna((dataframe['APR'].mean()), inplace=True)\n",
    "dataframe['MAY'].fillna((dataframe['MAY'].mean()), inplace=True)\n",
    "dataframe['JUN'].fillna((dataframe['JUN'].mean()), inplace=True)\n",
    "dataframe['JUL'].fillna((dataframe['JUL'].mean()), inplace=True)\n",
    "dataframe['AUG'].fillna((dataframe['AUG'].mean()), inplace=True)\n",
    "dataframe['SEP'].fillna((dataframe['SEP'].mean()), inplace=True)\n",
    "dataframe['OCT'].fillna((dataframe['OCT'].mean()), inplace=True)\n",
    "dataframe['DEC'].fillna((dataframe['DEC'].mean()), inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "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>Rainfall</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>49.2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>87.1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>29.2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2.3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>528.8</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Rainfall\n",
       "0      49.2\n",
       "1      87.1\n",
       "2      29.2\n",
       "3       2.3\n",
       "4     528.8"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "Rainfall    float64\n",
       "dtype: object"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Here, each month has been separated into a column with its value\n",
    "df_month = pd.DataFrame(dataframe[['JAN', 'FEB', 'MAR', 'APR', 'MAY', 'JUN', 'JUL','AUG', 'SEP', 'OCT']])\n",
    "df_month = df_month.T\n",
    "df_month = df_month.unstack(level=0)\n",
    "df_month = df_month.reset_index()\n",
    "\n",
    "# Creation of year/month in a separate column\n",
    "df_time = pd.Series(pd.period_range(start='1901', end='2016', freq='M', periods=None), dtype=str)\n",
    "df_time = df_time.reset_index()\n",
    "\n",
    "# To create a new datafarame which record the values of rainfall in the specific year and moth\n",
    "df = pd.DataFrame({'Rainfall':df_month[0]})\n",
    "display(df.head())\n",
    "display(df.dtypes)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "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"
    }
   ],
   "source": [
    "plt.plot(df)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(41160, 1)"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Convert pandas dataframe to numpy array\n",
    "dataset = df.values\n",
    "dataset = dataset.astype('float32')\n",
    "dataset.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Normalize the dataset\n",
    "scaler = MinMaxScaler(feature_range=(0,1))\n",
    "dataset = scaler.fit_transform(dataset)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Note: we cannot use random way to split training and test data as the sequence of events \n",
    "is important for time series. In this case, we take first 60% values for train and the remaining\n",
    "1/3 for testing split into train and test sets."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "27165\n",
      "13995\n"
     ]
    }
   ],
   "source": [
    "train_size = int(len(dataset) * 0.66)\n",
    "test_size = len(dataset) - train_size\n",
    "train, test = dataset[0:train_size,:], dataset[train_size:len(dataset),:]\n",
    "\n",
    "print(train_size)\n",
    "print(test_size)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Here, it is not possible to fit the model as normally do for image processing using X and Y. In this case, we should transform the data into something that looks like X and Y values. So, it can be trained on a sequence rather than indvidual datapoints. \n",
    "Therefore we need to convert into n number of columns for X where we feed sequence of numbers\n",
    "then the final column as Y where we provide the next number in the sequence as output.\n",
    "Here, we convert an array of values into a dataset matrix\n",
    "\n",
    "seq_size is the number of previous time steps to use as \n",
    "input variables to predict the next time period.\n",
    "\n",
    "creates a dataset where X is the amount of rainfall at a given time (t, t-1, t-2...) \n",
    "and Y is the amount of rainfall at the next time (t + 1)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "def to_sequences(dataset, se_size):\n",
    "    x = []\n",
    "    y = []\n",
    "    \n",
    "    for i in range(len(dataset)-seq_size-1):\n",
    "        window = dataset[i:(i+seq_size), 0]\n",
    "        x.append(window)\n",
    "        y.append(dataset[i+seq_size, 0])\n",
    "        \n",
    "    return np.array(x), np.array(y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Shape of training set: (27154, 10)\n",
      "Shape of test set: (13984, 10)\n"
     ]
    }
   ],
   "source": [
    "seq_size = 10    # Number of time steps to look back\n",
    "#Larger sequences may improve forecasting\n",
    "train_X, train_Y = to_sequences(train, seq_size)\n",
    "test_X, test_Y = to_sequences(test, seq_size)\n",
    "\n",
    "#Compare trainX and dataset. You can see that X= values at t, t+1 and t+2\n",
    "#whereas Y is the value that follows, t+3 (since our sequence size is 3)\n",
    "\n",
    "print(\"Shape of training set: {}\".format(train_X.shape))\n",
    "print(\"Shape of test set: {}\".format(test_X.shape))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Deep Model\n",
      "Model: \"sequential_1\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "dense_1 (Dense)              (None, 64)                704       \n",
      "_________________________________________________________________\n",
      "dense_2 (Dense)              (None, 32)                2080      \n",
      "_________________________________________________________________\n",
      "dense_3 (Dense)              (None, 1)                 33        \n",
      "=================================================================\n",
      "Total params: 2,817\n",
      "Trainable params: 2,817\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n",
      "None\n"
     ]
    }
   ],
   "source": [
    "#Create Model\n",
    "print(\"Deep Model\")\n",
    "#Fit dense model\n",
    "model = Sequential()\n",
    "model.add(Dense(64, input_dim=seq_size, activation='relu'))\n",
    "model.add(Dense(32, activation='relu'))\n",
    "model.add(Dense(1))\n",
    "model.compile(loss='mean_squared_error', optimizer='adam', metrics=['acc'])\n",
    "print(model.summary())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:From C:\\Users\\VHD\\anaconda3\\envs\\SI\\lib\\site-packages\\keras\\backend\\tensorflow_backend.py:422: The name tf.global_variables is deprecated. Please use tf.compat.v1.global_variables instead.\n",
      "\n",
      "Train on 27154 samples, validate on 13984 samples\n",
      "Epoch 1/100\n",
      " - 2s - loss: 0.0017 - acc: 0.0489 - val_loss: 0.0015 - val_acc: 0.0547\n",
      "Epoch 2/100\n",
      " - 2s - loss: 0.0015 - acc: 0.0489 - val_loss: 0.0016 - val_acc: 0.0547\n",
      "Epoch 3/100\n",
      " - 2s - loss: 0.0015 - acc: 0.0489 - val_loss: 0.0015 - val_acc: 0.0547\n",
      "Epoch 4/100\n",
      " - 2s - loss: 0.0015 - acc: 0.0489 - val_loss: 0.0015 - val_acc: 0.0547\n",
      "Epoch 5/100\n",
      " - 2s - loss: 0.0015 - acc: 0.0489 - val_loss: 0.0017 - val_acc: 0.0547\n",
      "Epoch 6/100\n",
      " - 2s - loss: 0.0015 - acc: 0.0489 - val_loss: 0.0015 - val_acc: 0.0547\n",
      "Epoch 7/100\n",
      " - 2s - loss: 0.0015 - acc: 0.0489 - val_loss: 0.0015 - val_acc: 0.0547\n",
      "Epoch 8/100\n",
      " - 2s - loss: 0.0015 - acc: 0.0489 - val_loss: 0.0015 - val_acc: 0.0547\n",
      "Epoch 9/100\n",
      " - 2s - loss: 0.0014 - acc: 0.0489 - val_loss: 0.0015 - val_acc: 0.0547\n",
      "Epoch 10/100\n",
      " - 2s - loss: 0.0014 - acc: 0.0489 - val_loss: 0.0015 - val_acc: 0.0547\n",
      "Epoch 11/100\n",
      " - 2s - loss: 0.0014 - acc: 0.0489 - val_loss: 0.0016 - val_acc: 0.0547\n",
      "Epoch 12/100\n",
      " - 2s - loss: 0.0014 - acc: 0.0489 - val_loss: 0.0016 - val_acc: 0.0547\n",
      "Epoch 13/100\n",
      " - 2s - loss: 0.0014 - acc: 0.0489 - val_loss: 0.0015 - val_acc: 0.0547\n",
      "Epoch 14/100\n",
      " - 2s - loss: 0.0014 - acc: 0.0489 - val_loss: 0.0014 - val_acc: 0.0547\n",
      "Epoch 15/100\n",
      " - 2s - loss: 0.0014 - acc: 0.0489 - val_loss: 0.0015 - val_acc: 0.0547\n",
      "Epoch 16/100\n",
      " - 2s - loss: 0.0014 - acc: 0.0489 - val_loss: 0.0015 - val_acc: 0.0547\n",
      "Epoch 17/100\n",
      " - 2s - loss: 0.0014 - acc: 0.0489 - val_loss: 0.0015 - val_acc: 0.0547\n",
      "Epoch 18/100\n",
      " - 2s - loss: 0.0014 - acc: 0.0489 - val_loss: 0.0015 - val_acc: 0.0547\n",
      "Epoch 19/100\n",
      " - 2s - loss: 0.0014 - acc: 0.0489 - val_loss: 0.0015 - val_acc: 0.0547\n",
      "Epoch 20/100\n",
      " - 2s - loss: 0.0014 - acc: 0.0489 - val_loss: 0.0015 - val_acc: 0.0547\n",
      "Epoch 21/100\n",
      " - 2s - loss: 0.0014 - acc: 0.0489 - val_loss: 0.0015 - val_acc: 0.0547\n",
      "Epoch 22/100\n",
      " - 2s - loss: 0.0014 - acc: 0.0489 - val_loss: 0.0015 - val_acc: 0.0547\n",
      "Epoch 23/100\n",
      " - 2s - loss: 0.0014 - acc: 0.0489 - val_loss: 0.0015 - val_acc: 0.0547\n",
      "Epoch 24/100\n",
      " - 2s - loss: 0.0014 - acc: 0.0489 - val_loss: 0.0015 - val_acc: 0.0547\n",
      "Epoch 25/100\n",
      " - 2s - loss: 0.0013 - acc: 0.0489 - val_loss: 0.0016 - val_acc: 0.0547\n",
      "Epoch 26/100\n",
      " - 2s - loss: 0.0013 - acc: 0.0489 - val_loss: 0.0016 - val_acc: 0.0547\n",
      "Epoch 27/100\n",
      " - 2s - loss: 0.0013 - acc: 0.0489 - val_loss: 0.0015 - val_acc: 0.0547\n",
      "Epoch 28/100\n",
      " - 2s - loss: 0.0013 - acc: 0.0489 - val_loss: 0.0015 - val_acc: 0.0547\n",
      "Epoch 29/100\n",
      " - 2s - loss: 0.0013 - acc: 0.0489 - val_loss: 0.0016 - val_acc: 0.0547\n",
      "Epoch 30/100\n",
      " - 2s - loss: 0.0013 - acc: 0.0489 - val_loss: 0.0015 - val_acc: 0.0547\n",
      "Epoch 31/100\n",
      " - 2s - loss: 0.0013 - acc: 0.0489 - val_loss: 0.0016 - val_acc: 0.0547\n",
      "Epoch 32/100\n",
      " - 2s - loss: 0.0013 - acc: 0.0489 - val_loss: 0.0016 - val_acc: 0.0547\n",
      "Epoch 33/100\n",
      " - 2s - loss: 0.0013 - acc: 0.0489 - val_loss: 0.0015 - val_acc: 0.0547\n",
      "Epoch 34/100\n",
      " - 2s - loss: 0.0013 - acc: 0.0489 - val_loss: 0.0016 - val_acc: 0.0547\n",
      "Epoch 35/100\n",
      " - 2s - loss: 0.0013 - acc: 0.0489 - val_loss: 0.0016 - val_acc: 0.0547\n",
      "Epoch 36/100\n",
      " - 2s - loss: 0.0013 - acc: 0.0489 - val_loss: 0.0015 - val_acc: 0.0547\n",
      "Epoch 37/100\n",
      " - 2s - loss: 0.0013 - acc: 0.0489 - val_loss: 0.0015 - val_acc: 0.0547\n",
      "Epoch 38/100\n",
      " - 2s - loss: 0.0013 - acc: 0.0489 - val_loss: 0.0016 - val_acc: 0.0547\n",
      "Epoch 39/100\n",
      " - 2s - loss: 0.0013 - acc: 0.0489 - val_loss: 0.0015 - val_acc: 0.0547\n",
      "Epoch 40/100\n",
      " - 2s - loss: 0.0013 - acc: 0.0489 - val_loss: 0.0015 - val_acc: 0.0547\n",
      "Epoch 41/100\n",
      " - 2s - loss: 0.0013 - acc: 0.0489 - val_loss: 0.0015 - val_acc: 0.0547\n",
      "Epoch 42/100\n",
      " - 2s - loss: 0.0013 - acc: 0.0489 - val_loss: 0.0016 - val_acc: 0.0547\n",
      "Epoch 43/100\n",
      " - 2s - loss: 0.0013 - acc: 0.0489 - val_loss: 0.0015 - val_acc: 0.0547\n",
      "Epoch 44/100\n",
      " - 2s - loss: 0.0013 - acc: 0.0489 - val_loss: 0.0017 - val_acc: 0.0547\n",
      "Epoch 45/100\n",
      " - 2s - loss: 0.0013 - acc: 0.0489 - val_loss: 0.0015 - val_acc: 0.0547\n",
      "Epoch 46/100\n",
      " - 2s - loss: 0.0013 - acc: 0.0489 - val_loss: 0.0016 - val_acc: 0.0547\n",
      "Epoch 47/100\n",
      " - 2s - loss: 0.0013 - acc: 0.0489 - val_loss: 0.0016 - val_acc: 0.0547\n",
      "Epoch 48/100\n",
      " - 2s - loss: 0.0013 - acc: 0.0489 - val_loss: 0.0015 - val_acc: 0.0547\n",
      "Epoch 49/100\n",
      " - 2s - loss: 0.0013 - acc: 0.0489 - val_loss: 0.0015 - val_acc: 0.0547\n",
      "Epoch 50/100\n",
      " - 2s - loss: 0.0013 - acc: 0.0489 - val_loss: 0.0016 - val_acc: 0.0547\n",
      "Epoch 51/100\n",
      " - 2s - loss: 0.0013 - acc: 0.0489 - val_loss: 0.0015 - val_acc: 0.0547\n",
      "Epoch 52/100\n",
      " - 2s - loss: 0.0013 - acc: 0.0489 - val_loss: 0.0016 - val_acc: 0.0547\n",
      "Epoch 53/100\n",
      " - 2s - loss: 0.0013 - acc: 0.0489 - val_loss: 0.0015 - val_acc: 0.0547\n",
      "Epoch 54/100\n",
      " - 2s - loss: 0.0012 - acc: 0.0489 - val_loss: 0.0015 - val_acc: 0.0547\n",
      "Epoch 55/100\n",
      " - 2s - loss: 0.0012 - acc: 0.0489 - val_loss: 0.0015 - val_acc: 0.0547\n",
      "Epoch 56/100\n",
      " - 2s - loss: 0.0012 - acc: 0.0489 - val_loss: 0.0015 - val_acc: 0.0547\n",
      "Epoch 57/100\n",
      " - 2s - loss: 0.0012 - acc: 0.0489 - val_loss: 0.0015 - val_acc: 0.0547\n",
      "Epoch 58/100\n",
      " - 2s - loss: 0.0012 - acc: 0.0489 - val_loss: 0.0015 - val_acc: 0.0547\n",
      "Epoch 59/100\n",
      " - 2s - loss: 0.0012 - acc: 0.0489 - val_loss: 0.0016 - val_acc: 0.0547\n",
      "Epoch 60/100\n",
      " - 2s - loss: 0.0012 - acc: 0.0489 - val_loss: 0.0016 - val_acc: 0.0547\n",
      "Epoch 61/100\n",
      " - 2s - loss: 0.0012 - acc: 0.0489 - val_loss: 0.0015 - val_acc: 0.0547\n",
      "Epoch 62/100\n",
      " - 2s - loss: 0.0012 - acc: 0.0489 - val_loss: 0.0015 - val_acc: 0.0547\n",
      "Epoch 63/100\n",
      " - 2s - loss: 0.0012 - acc: 0.0489 - val_loss: 0.0016 - val_acc: 0.0547\n",
      "Epoch 64/100\n",
      " - 2s - loss: 0.0012 - acc: 0.0489 - val_loss: 0.0016 - val_acc: 0.0547\n",
      "Epoch 65/100\n",
      " - 2s - loss: 0.0012 - acc: 0.0489 - val_loss: 0.0018 - val_acc: 0.0547\n",
      "Epoch 66/100\n",
      " - 2s - loss: 0.0012 - acc: 0.0489 - val_loss: 0.0016 - val_acc: 0.0547\n",
      "Epoch 67/100\n",
      " - 2s - loss: 0.0012 - acc: 0.0489 - val_loss: 0.0016 - val_acc: 0.0547\n",
      "Epoch 68/100\n",
      " - 2s - loss: 0.0012 - acc: 0.0489 - val_loss: 0.0015 - val_acc: 0.0547\n",
      "Epoch 69/100\n",
      " - 2s - loss: 0.0012 - acc: 0.0489 - val_loss: 0.0016 - val_acc: 0.0547\n",
      "Epoch 70/100\n",
      " - 2s - loss: 0.0012 - acc: 0.0489 - val_loss: 0.0016 - val_acc: 0.0547\n",
      "Epoch 71/100\n",
      " - 2s - loss: 0.0012 - acc: 0.0489 - val_loss: 0.0015 - val_acc: 0.0547\n",
      "Epoch 72/100\n",
      " - 2s - loss: 0.0012 - acc: 0.0489 - val_loss: 0.0017 - val_acc: 0.0547\n",
      "Epoch 73/100\n",
      " - 2s - loss: 0.0012 - acc: 0.0489 - val_loss: 0.0015 - val_acc: 0.0547\n",
      "Epoch 74/100\n",
      " - 2s - loss: 0.0012 - acc: 0.0489 - val_loss: 0.0015 - val_acc: 0.0547\n",
      "Epoch 75/100\n",
      " - 2s - loss: 0.0012 - acc: 0.0489 - val_loss: 0.0015 - val_acc: 0.0547\n",
      "Epoch 76/100\n",
      " - 2s - loss: 0.0012 - acc: 0.0489 - val_loss: 0.0017 - val_acc: 0.0547\n",
      "Epoch 77/100\n",
      " - 2s - loss: 0.0012 - acc: 0.0489 - val_loss: 0.0016 - val_acc: 0.0547\n",
      "Epoch 78/100\n",
      " - 2s - loss: 0.0012 - acc: 0.0489 - val_loss: 0.0016 - val_acc: 0.0547\n",
      "Epoch 79/100\n",
      " - 2s - loss: 0.0012 - acc: 0.0489 - val_loss: 0.0016 - val_acc: 0.0547\n",
      "Epoch 80/100\n",
      " - 2s - loss: 0.0012 - acc: 0.0489 - val_loss: 0.0016 - val_acc: 0.0547\n",
      "Epoch 81/100\n",
      " - 2s - loss: 0.0012 - acc: 0.0489 - val_loss: 0.0017 - val_acc: 0.0547\n",
      "Epoch 82/100\n",
      " - 2s - loss: 0.0012 - acc: 0.0489 - val_loss: 0.0016 - val_acc: 0.0547\n",
      "Epoch 83/100\n",
      " - 2s - loss: 0.0012 - acc: 0.0489 - val_loss: 0.0015 - val_acc: 0.0547\n",
      "Epoch 84/100\n",
      " - 2s - loss: 0.0012 - acc: 0.0489 - val_loss: 0.0017 - val_acc: 0.0547\n",
      "Epoch 85/100\n",
      " - 2s - loss: 0.0012 - acc: 0.0489 - val_loss: 0.0017 - val_acc: 0.0547\n",
      "Epoch 86/100\n",
      " - 2s - loss: 0.0012 - acc: 0.0489 - val_loss: 0.0016 - val_acc: 0.0547\n",
      "Epoch 87/100\n",
      " - 2s - loss: 0.0012 - acc: 0.0489 - val_loss: 0.0017 - val_acc: 0.0547\n",
      "Epoch 88/100\n",
      " - 2s - loss: 0.0012 - acc: 0.0489 - val_loss: 0.0016 - val_acc: 0.0547\n",
      "Epoch 89/100\n",
      " - 2s - loss: 0.0012 - acc: 0.0489 - val_loss: 0.0016 - val_acc: 0.0547\n",
      "Epoch 90/100\n",
      " - 2s - loss: 0.0012 - acc: 0.0489 - val_loss: 0.0017 - val_acc: 0.0547\n",
      "Epoch 91/100\n",
      " - 2s - loss: 0.0012 - acc: 0.0489 - val_loss: 0.0016 - val_acc: 0.0547\n",
      "Epoch 92/100\n",
      " - 2s - loss: 0.0012 - acc: 0.0489 - val_loss: 0.0016 - val_acc: 0.0547\n",
      "Epoch 93/100\n",
      " - 2s - loss: 0.0012 - acc: 0.0489 - val_loss: 0.0016 - val_acc: 0.0547\n",
      "Epoch 94/100\n",
      " - 2s - loss: 0.0012 - acc: 0.0489 - val_loss: 0.0015 - val_acc: 0.0547\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 95/100\n",
      " - 2s - loss: 0.0012 - acc: 0.0489 - val_loss: 0.0017 - val_acc: 0.0547\n",
      "Epoch 96/100\n",
      " - 2s - loss: 0.0012 - acc: 0.0489 - val_loss: 0.0016 - val_acc: 0.0547\n",
      "Epoch 97/100\n",
      " - 2s - loss: 0.0012 - acc: 0.0489 - val_loss: 0.0015 - val_acc: 0.0547\n",
      "Epoch 98/100\n",
      " - 2s - loss: 0.0012 - acc: 0.0489 - val_loss: 0.0016 - val_acc: 0.0547\n",
      "Epoch 99/100\n",
      " - 2s - loss: 0.0012 - acc: 0.0489 - val_loss: 0.0016 - val_acc: 0.0547\n",
      "Epoch 100/100\n",
      " - 2s - loss: 0.0012 - acc: 0.0489 - val_loss: 0.0017 - val_acc: 0.0547\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.callbacks.callbacks.History at 0x21dc5b08>"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.fit(train_X, train_Y, validation_data=(test_X, test_Y), verbose= 2, epochs = 100)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[-2.4616878]\n",
      " [39.587326 ]\n",
      " [72.50304  ]\n",
      " ...\n",
      " [17.740303 ]\n",
      " [16.35677  ]\n",
      " [21.934618 ]]\n",
      "[[  0.         159.80000272  12.20000003 ...   0.20000001   5.40000019\n",
      "    6.69999955]]\n"
     ]
    }
   ],
   "source": [
    "#Make prediction\n",
    "predict_train = model.predict(train_X)\n",
    "predict_test = model.predict(test_X)\n",
    "\n",
    "# Estimate model performance\n",
    "#SInce we used minmaxscaler we can now use scaler.inverse_transform\n",
    "#to invert the transformation.\n",
    "\n",
    "predict_train = scaler.inverse_transform(predict_train)\n",
    "trainY_inverse = scaler.inverse_transform([train_Y])\n",
    "predict_test = scaler.inverse_transform(predict_test)\n",
    "testY_inverse = scaler.inverse_transform([test_Y])\n",
    "\n",
    "print(predict_train)\n",
    "print(trainY_inverse)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train score: 80.30 RMSE\n",
      "Test score: 97.88 RMSE\n"
     ]
    }
   ],
   "source": [
    "# calculate root mean squared error\n",
    "trainScore = math.sqrt(mean_squared_error(trainY_inverse[0], predict_train[:,0]))\n",
    "print(\"Train score: %.2f RMSE\" % (trainScore))\n",
    "\n",
    "testScore = math.sqrt(mean_squared_error(testY_inverse[0], predict_test[:,0]))\n",
    "print(\"Test score: %.2f RMSE\" % (testScore))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "# shift train predictions for plotting\n",
    "#we must shift the predictions so that they align on the x-axis with the original dataset. \n",
    "predict_train_plot = np.empty_like(dataset)\n",
    "predict_train_plot[:,:] = np.nan\n",
    "predict_train_plot[seq_size:len(predict_train)+seq_size, :] = predict_train\n",
    "\n",
    "# shift test predictions for plotting\n",
    "predict_test_plot = np.empty_like(dataset)\n",
    "predict_test_plot[:,:] = np.nan\n",
    "predict_test_plot[len(predict_train)+(seq_size*2)+1:len(dataset)-1, :] = predict_test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "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"
    }
   ],
   "source": [
    "# plot baseline and predictions\n",
    "plt.plot(scaler.inverse_transform(dataset))\n",
    "plt.plot(predict_train_plot)\n",
    "plt.plot(predict_test_plot)\n",
    "plt.show()"
   ]
  },
  {
   "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
}