keras_time_series.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {
    "collapsed": true,
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "%matplotlib inline\n",
    "import matplotlib.pyplot as plt\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import keras\n",
    "import sklearn\n",
    "import sklearn.preprocessing\n",
    "import sklearn.metrics"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "url = 'https://github.com/raw/blue-yonder/pydse/master/pydse/data/international-airline-passengers.csv'\n",
    "data = pd.read_csv(url, sep=';')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Month</th>\n",
       "      <th>Passengers</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1949-01</td>\n",
       "      <td>112</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1949-02</td>\n",
       "      <td>118</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1949-03</td>\n",
       "      <td>132</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1949-04</td>\n",
       "      <td>129</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1949-05</td>\n",
       "      <td>121</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     Month  Passengers\n",
       "0  1949-01         112\n",
       "1  1949-02         118\n",
       "2  1949-03         132\n",
       "3  1949-04         129\n",
       "4  1949-05         121"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x7f4260f96a58>]"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXoAAAD8CAYAAAB5Pm/hAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\ncBLohRBihZNAL4QQK9z/Dxb9B3vY4hpdAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f4261012dd8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(data.Passengers)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "np.random.seed(7)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "ts = data.iloc[:, 1].astype('float').values.reshape(-1, 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(144, 1)"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ts.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "scaler = sklearn.preprocessing.MinMaxScaler(feature_range=(0, 1))\n",
    "ts = scaler.fit_transform(ts)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x7f4260f2f550>]"
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXcAAAD8CAYAAACMwORRAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n2UmUPYCglQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f4260f3a2b0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(ts)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "train_size = int(len(ts) * 0.67)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "ts_train = ts[:train_size]\n",
    "ts_test = ts[train_size:]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(96, 1)"
      ]
     },
     "execution_count": 54,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ts_train.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(48, 1)"
      ]
     },
     "execution_count": 55,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ts_test.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "look_back=1\n",
    "def make_xy(ts):\n",
    "    Y = ts[look_back:]\n",
    "    X = ts[:-look_back]\n",
    "    return X, Y"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "X_train, Y_train = make_xy(ts_train)\n",
    "X_test, Y_test = make_xy(ts_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {
    "collapsed": true,
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "X_train = X_train.reshape(X_train.shape[0], 1, 1)\n",
    "X_test = X_test.reshape(X_test.shape[0], 1, 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(95, 1, 1)"
      ]
     },
     "execution_count": 85,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(95, 1)"
      ]
     },
     "execution_count": 86,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Y_train.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "model = keras.models.Sequential([\n",
    "    keras.layers.LSTM(4, input_shape=(1, look_back)),\n",
    "    keras.layers.Dense(1)\n",
    "])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "model.compile(optimizer=keras.optimizers.Adam(), loss=keras.losses.mean_squared_error)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {
    "collapsed": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 95 samples, validate on 47 samples\n",
      "Epoch 1/100\n",
      "95/95 [==============================] - 2s 19ms/step - loss: 0.0406 - val_loss: 0.2212\n",
      "Epoch 2/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0199 - val_loss: 0.1526\n",
      "Epoch 3/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0147 - val_loss: 0.1246\n",
      "Epoch 4/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0130 - val_loss: 0.1090\n",
      "Epoch 5/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0120 - val_loss: 0.0966\n",
      "Epoch 6/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0111 - val_loss: 0.0835\n",
      "Epoch 7/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0101 - val_loss: 0.0753\n",
      "Epoch 8/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0090 - val_loss: 0.0690\n",
      "Epoch 9/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0079 - val_loss: 0.0564\n",
      "Epoch 10/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0069 - val_loss: 0.0472\n",
      "Epoch 11/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0060 - val_loss: 0.0388\n",
      "Epoch 12/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0051 - val_loss: 0.0355\n",
      "Epoch 13/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0044 - val_loss: 0.0247\n",
      "Epoch 14/100\n",
      "95/95 [==============================] - ETA: 0s - loss: 0.003 - 0s 3ms/step - loss: 0.0037 - val_loss: 0.0220\n",
      "Epoch 15/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0033 - val_loss: 0.0172\n",
      "Epoch 16/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0029 - val_loss: 0.0146\n",
      "Epoch 17/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0026 - val_loss: 0.0125\n",
      "Epoch 18/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0024 - val_loss: 0.0112\n",
      "Epoch 19/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0023 - val_loss: 0.0097\n",
      "Epoch 20/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0022 - val_loss: 0.0093\n",
      "Epoch 21/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0021 - val_loss: 0.0088\n",
      "Epoch 22/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0021 - val_loss: 0.0091\n",
      "Epoch 23/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0022 - val_loss: 0.0096\n",
      "Epoch 24/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0022 - val_loss: 0.0088\n",
      "Epoch 25/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0021 - val_loss: 0.0086\n",
      "Epoch 26/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0021 - val_loss: 0.0086\n",
      "Epoch 27/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0021 - val_loss: 0.0087\n",
      "Epoch 28/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0021 - val_loss: 0.0088\n",
      "Epoch 29/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0021 - val_loss: 0.0090\n",
      "Epoch 30/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0021 - val_loss: 0.0085\n",
      "Epoch 31/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0021 - val_loss: 0.0087\n",
      "Epoch 32/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0021 - val_loss: 0.0084\n",
      "Epoch 33/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0021 - val_loss: 0.0088\n",
      "Epoch 34/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0021 - val_loss: 0.0085\n",
      "Epoch 35/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0021 - val_loss: 0.0086\n",
      "Epoch 36/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0021 - val_loss: 0.0092\n",
      "Epoch 37/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0021 - val_loss: 0.0086\n",
      "Epoch 38/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0021 - val_loss: 0.0085\n",
      "Epoch 39/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0021 - val_loss: 0.0085\n",
      "Epoch 40/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0021 - val_loss: 0.0085\n",
      "Epoch 41/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0021 - val_loss: 0.0086\n",
      "Epoch 42/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0020 - val_loss: 0.0082\n",
      "Epoch 43/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0020 - val_loss: 0.0088\n",
      "Epoch 44/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0021 - val_loss: 0.0087\n",
      "Epoch 45/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0021 - val_loss: 0.0083\n",
      "Epoch 46/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0021 - val_loss: 0.0083\n",
      "Epoch 47/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0021 - val_loss: 0.0092\n",
      "Epoch 48/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0021 - val_loss: 0.0084\n",
      "Epoch 49/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0022 - val_loss: 0.0085\n",
      "Epoch 50/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0020 - val_loss: 0.0082\n",
      "Epoch 51/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0021 - val_loss: 0.0085\n",
      "Epoch 52/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0021 - val_loss: 0.0083\n",
      "Epoch 53/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0021 - val_loss: 0.0085\n",
      "Epoch 54/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0021 - val_loss: 0.0086\n",
      "Epoch 55/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0021 - val_loss: 0.0084\n",
      "Epoch 56/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0021 - val_loss: 0.0082\n",
      "Epoch 57/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0020 - val_loss: 0.0082\n",
      "Epoch 58/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0020 - val_loss: 0.0089\n",
      "Epoch 59/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0020 - val_loss: 0.0089\n",
      "Epoch 60/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0021 - val_loss: 0.0083\n",
      "Epoch 61/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0021 - val_loss: 0.0087\n",
      "Epoch 62/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0021 - val_loss: 0.0086\n",
      "Epoch 63/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0021 - val_loss: 0.0085\n",
      "Epoch 64/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0021 - val_loss: 0.0083\n",
      "Epoch 65/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0020 - val_loss: 0.0085\n",
      "Epoch 66/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0021 - val_loss: 0.0085\n",
      "Epoch 67/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0021 - val_loss: 0.0088\n",
      "Epoch 68/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0020 - val_loss: 0.0090\n",
      "Epoch 69/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0021 - val_loss: 0.0084\n",
      "Epoch 70/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0020 - val_loss: 0.0082\n",
      "Epoch 71/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0021 - val_loss: 0.0082\n",
      "Epoch 72/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0020 - val_loss: 0.0096\n",
      "Epoch 73/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0020 - val_loss: 0.0083\n",
      "Epoch 74/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0021 - val_loss: 0.0089\n",
      "Epoch 75/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0020 - val_loss: 0.0085\n",
      "Epoch 76/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0021 - val_loss: 0.0082\n",
      "Epoch 77/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0021 - val_loss: 0.0085\n",
      "Epoch 78/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0021 - val_loss: 0.0092\n",
      "Epoch 79/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0020 - val_loss: 0.0095\n",
      "Epoch 80/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0021 - val_loss: 0.0093\n",
      "Epoch 81/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0021 - val_loss: 0.0085\n",
      "Epoch 82/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0020 - val_loss: 0.0081\n",
      "Epoch 83/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0021 - val_loss: 0.0087\n",
      "Epoch 84/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0021 - val_loss: 0.0084\n",
      "Epoch 85/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0021 - val_loss: 0.0093\n",
      "Epoch 86/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0020 - val_loss: 0.0087\n",
      "Epoch 87/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0020 - val_loss: 0.0081\n",
      "Epoch 88/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0020 - val_loss: 0.0082\n",
      "Epoch 89/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0020 - val_loss: 0.0095\n",
      "Epoch 90/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0020 - val_loss: 0.0088\n",
      "Epoch 91/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0020 - val_loss: 0.0086\n",
      "Epoch 92/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0020 - val_loss: 0.0092\n",
      "Epoch 93/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0021 - val_loss: 0.0089\n",
      "Epoch 94/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0020 - val_loss: 0.0089\n",
      "Epoch 95/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0021 - val_loss: 0.0091\n",
      "Epoch 96/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0020 - val_loss: 0.0089\n",
      "Epoch 97/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0021 - val_loss: 0.0086\n",
      "Epoch 98/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0021 - val_loss: 0.0085\n",
      "Epoch 99/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0021 - val_loss: 0.0096\n",
      "Epoch 100/100\n",
      "95/95 [==============================] - 0s 3ms/step - loss: 0.0020 - val_loss: 0.0087\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.callbacks.History at 0x7f4260b616a0>"
      ]
     },
     "execution_count": 63,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.fit(X_train, Y_train, epochs=100, batch_size=1, validation_data=[X_test, Y_test])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "pred_train = model.predict(X_train)\n",
    "pred_test = model.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "pred_train_original_scale = scaler.inverse_transform(pred_train)\n",
    "pred_test_original_scale = scaler.inverse_transform(pred_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "Y_train_original_scale = scaler.inverse_transform(Y_train)\n",
    "Y_test_original_scale = scaler.inverse_transform(Y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "22.927175101435921"
      ]
     },
     "execution_count": 71,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_rmse = np.sqrt(sklearn.metrics.mean_squared_error(Y_train_original_scale, pred_train_original_scale))\n",
    "train_rmse"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "48.279758992641824"
      ]
     },
     "execution_count": 72,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_rmse = np.sqrt(sklearn.metrics.mean_squared_error(Y_test_original_scale, pred_test_original_scale))\n",
    "test_rmse"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(47, 1)"
      ]
     },
     "execution_count": 73,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred_test.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(95, 1)"
      ]
     },
     "execution_count": 74,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred_train.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(144, 1)"
      ]
     },
     "execution_count": 75,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ts.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x7f425c130c88>]"
      ]
     },
     "execution_count": 80,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXoAAAD8CAYAAAB5Pm/hAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n3P8DnLYNNIhv8mcAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f425c130d68>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(data.Passengers)\n",
    "\n",
    "plt.plot(np.vstack([pred_train_original_scale, pred_test_original_scale]))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "http://machinelearningmastery.com/time-series-prediction-lstm-recurrent-neural-networks-python-keras/"
   ]
  }
 ],
 "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": "3.6.1"
  },
  "toc": {
   "nav_menu": {
    "height": "12px",
    "width": "252px"
   },
   "number_sections": true,
   "sideBar": true,
   "skip_h1_title": false,
   "toc_cell": false,
   "toc_position": {},
   "toc_section_display": "block",
   "toc_window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}