stock_price_predictor.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "# Predict Stock Price\n",
    "\n",
    "1. Get historic stock price data\n",
    "2. Extract price to predict\n",
    "3. Normalize\n",
    "4. Create Time Series\n",
    "5. Build model\n",
    "6. Fit model\n",
    "7. Test and evaluate"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n",
      "Using CNTK backend\n",
      "/home/nbuser/anaconda3_420/lib/python3.5/site-packages/keras/backend/cntk_backend.py:19: UserWarning: CNTK backend warning: GPU is not detected. CNTK's CPU version is not fully optimized,please run with GPU to get better performance.\n",
      "  'CNTK backend warning: GPU is not detected. '\n"
     ]
    }
   ],
   "source": [
    "from keras import backend as K\n",
    "import os\n",
    "from importlib import reload\n",
    "\n",
    "def set_keras_backend(backend):\n",
    "    if K.backend() != backend:\n",
    "        os.environ['KERAS_BACKEND'] = backend\n",
    "        reload(K)\n",
    "        assert K.backend() == backend\n",
    "\n",
    "set_keras_backend(\"cntk\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/nbuser/anaconda3_420/lib/python3.5/site-packages/sklearn/cross_validation.py:44: DeprecationWarning: This module was deprecated in version 0.18 in favor of the model_selection module into which all the refactored classes and functions are moved. Also note that the interface of the new CV iterators are different from that of this module. This module will be removed in 0.20.\n",
      "  \"This module will be removed in 0.20.\", DeprecationWarning)\n"
     ]
    }
   ],
   "source": [
    "import numpy as np # linear algebra\n",
    "import pandas as pd # data processing, CSV file I/O (e.g. pd.read_csv)\n",
    "from keras.layers.core import Dense, Activation, Dropout\n",
    "from keras.layers.recurrent import LSTM\n",
    "from keras.models import Sequential\n",
    "from sklearn.cross_validation import  train_test_split\n",
    "from sklearn.preprocessing import MinMaxScaler\n",
    "from sklearn.metrics import mean_squared_error\n",
    "import matplotlib.pyplot as plt\n",
    "import math"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "np.random.seed(7)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>date</th>\n",
       "      <th>close</th>\n",
       "      <th>volume</th>\n",
       "      <th>open</th>\n",
       "      <th>high</th>\n",
       "      <th>low</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>16:00</td>\n",
       "      <td>91.33</td>\n",
       "      <td>67,998,445</td>\n",
       "      <td>86.82</td>\n",
       "      <td>91.475</td>\n",
       "      <td>85.2500</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2018/02/05</td>\n",
       "      <td>88.00</td>\n",
       "      <td>50588980.0000</td>\n",
       "      <td>90.56</td>\n",
       "      <td>93.240</td>\n",
       "      <td>88.0000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2018/02/02</td>\n",
       "      <td>91.78</td>\n",
       "      <td>47484680.0000</td>\n",
       "      <td>93.64</td>\n",
       "      <td>93.970</td>\n",
       "      <td>91.5000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2018/02/01</td>\n",
       "      <td>94.26</td>\n",
       "      <td>46866010.0000</td>\n",
       "      <td>94.79</td>\n",
       "      <td>96.070</td>\n",
       "      <td>93.5813</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2018/01/31</td>\n",
       "      <td>95.01</td>\n",
       "      <td>47448270.0000</td>\n",
       "      <td>93.75</td>\n",
       "      <td>95.400</td>\n",
       "      <td>93.5100</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         date  close         volume   open    high      low\n",
       "0       16:00  91.33     67,998,445  86.82  91.475  85.2500\n",
       "1  2018/02/05  88.00  50588980.0000  90.56  93.240  88.0000\n",
       "2  2018/02/02  91.78  47484680.0000  93.64  93.970  91.5000\n",
       "3  2018/02/01  94.26  46866010.0000  94.79  96.070  93.5813\n",
       "4  2018/01/31  95.01  47448270.0000  93.75  95.400  93.5100"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#load the dataset\n",
    "msft_dataset = pd.read_csv('~/library/HistoricalQuotes.csv')\n",
    "msft_dataset.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "date       object\n",
       "close     float64\n",
       "volume     object\n",
       "open      float64\n",
       "high      float64\n",
       "low       float64\n",
       "dtype: object"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "msft_dataset.dtypes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "msft_dataset['date'] = pd.to_datetime(msft_dataset['date'])\n",
    "msft_dataset['close'] = pd.to_numeric(msft_dataset['close'], downcast='float')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "DatetimeIndex: 2520 entries, 2018-02-07 16:00:00 to 2008-02-05 00:00:00\n",
      "Data columns (total 5 columns):\n",
      "close     2520 non-null float32\n",
      "volume    2520 non-null object\n",
      "open      2520 non-null float64\n",
      "high      2520 non-null float64\n",
      "low       2520 non-null float64\n",
      "dtypes: float32(1), float64(3), object(1)\n",
      "memory usage: 108.3+ KB\n"
     ]
    }
   ],
   "source": [
    "msft_dataset.set_index('date',inplace=True)\n",
    "msft_dataset.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "msft_dataset.sort_index(inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f051a7e9080>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#extract just close prices as that is what we want to predict\n",
    "msft_close = msft_dataset['close']\n",
    "msft_close = msft_close.values.reshape(len(msft_close), 1)\n",
    "plt.plot(msft_close)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "#normalize data\n",
    "scaler = MinMaxScaler(feature_range=(0,1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[ 0.17430502],\n",
       "       [ 0.167418  ],\n",
       "       [ 0.16240922],\n",
       "       ..., \n",
       "       [ 0.95955426],\n",
       "       [ 0.91222137],\n",
       "       [ 0.95391935]], dtype=float32)"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "msft_close = scaler.fit_transform(msft_close)\n",
    "msft_close"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Split data into train and test:  1764 756\n"
     ]
    }
   ],
   "source": [
    "#split data into train and test\n",
    "train_size = int(len(msft_close)* 0.7)\n",
    "test_size = len(msft_close) - train_size\n",
    "\n",
    "msft_train, msft_test = msft_close[0:train_size, :], msft_close[train_size:len(msft_close), :]\n",
    "\n",
    "print('Split data into train and test: ', len(msft_train), len(msft_test))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "#need to now convert the data into time series looking back over a period of days...e.g. use last 7 days to predict price\n",
    "\n",
    "def create_ts(ds, series):\n",
    "    X, Y =[], []\n",
    "    for i in range(len(ds)-series - 1):\n",
    "        item = ds[i:(i+series), 0]\n",
    "        X.append(item)\n",
    "        Y.append(ds[i+series, 0])\n",
    "    return np.array(X), np.array(Y)\n",
    "\n",
    "series = 7\n",
    "\n",
    "trainX, trainY = create_ts(msft_train, series)\n",
    "testX, testY = create_ts(msft_test, series)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[ 0.17430502,  0.167418  ,  0.16240922, ...,  0.16353619,\n",
       "         0.16516405,  0.17292762],\n",
       "       [ 0.167418  ,  0.16240922,  0.16791886, ...,  0.16516405,\n",
       "         0.17292762,  0.16716754],\n",
       "       [ 0.16240922,  0.16791886,  0.16353619, ...,  0.17292762,\n",
       "         0.16716754,  0.1661658 ],\n",
       "       ..., \n",
       "       [ 0.40045083,  0.40107691,  0.39894813, ...,  0.32607061,\n",
       "         0.33633858,  0.31617832],\n",
       "       [ 0.40107691,  0.39894813,  0.34447783, ...,  0.33633858,\n",
       "         0.31617832,  0.32719761],\n",
       "       [ 0.39894813,  0.34447783,  0.32607061, ...,  0.31617832,\n",
       "         0.32719761,  0.33120459]], dtype=float32)"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "trainX"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 0.17430502,  0.167418  ,  0.16240922,  0.16791886,  0.16353619,\n",
       "        0.16516405,  0.17292762], dtype=float32)"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "trainX[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "#reshape into  LSTM format - samples, steps, features\n",
    "trainX = np.reshape(trainX, (trainX.shape[0], trainX.shape[1], 1))\n",
    "testX = np.reshape(testX, (testX.shape[0], testX.shape[1], 1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/100\n",
      "1756/1756 [==============================] - 0s - loss: 0.0116     \n",
      "Epoch 2/100\n",
      "1756/1756 [==============================] - 0s - loss: 0.0026     \n",
      "Epoch 3/100\n",
      "1756/1756 [==============================] - 0s - loss: 0.0016     \n",
      "Epoch 4/100\n",
      "1756/1756 [==============================] - 0s - loss: 9.2738e-04     \n",
      "Epoch 5/100\n",
      "1756/1756 [==============================] - 0s - loss: 4.6664e-04     \n",
      "Epoch 6/100\n",
      "1756/1756 [==============================] - 0s - loss: 2.4888e-04     \n",
      "Epoch 7/100\n",
      "1756/1756 [==============================] - 0s - loss: 1.7485e-04     \n",
      "Epoch 8/100\n",
      "1756/1756 [==============================] - 0s - loss: 1.5534e-04     \n",
      "Epoch 9/100\n",
      "1756/1756 [==============================] - 0s - loss: 1.5097e-04     \n",
      "Epoch 10/100\n",
      "1756/1756 [==============================] - 0s - loss: 1.4955e-04     \n",
      "Epoch 11/100\n",
      "1756/1756 [==============================] - 0s - loss: 1.4849e-04     \n",
      "Epoch 12/100\n",
      "1756/1756 [==============================] - 0s - loss: 1.4429e-04     \n",
      "Epoch 13/100\n",
      "1756/1756 [==============================] - 0s - loss: 1.4249e-04     \n",
      "Epoch 14/100\n",
      "1756/1756 [==============================] - 0s - loss: 1.4013e-04     \n",
      "Epoch 15/100\n",
      "1756/1756 [==============================] - 0s - loss: 1.3803e-04     \n",
      "Epoch 16/100\n",
      "1756/1756 [==============================] - 0s - loss: 1.3585e-04     \n",
      "Epoch 17/100\n",
      "1756/1756 [==============================] - 0s - loss: 1.3390e-04     \n",
      "Epoch 18/100\n",
      "1756/1756 [==============================] - 0s - loss: 1.3172e-04     \n",
      "Epoch 19/100\n",
      "1756/1756 [==============================] - 0s - loss: 1.3100e-04     \n",
      "Epoch 20/100\n",
      "1756/1756 [==============================] - 0s - loss: 1.2700e-04     \n",
      "Epoch 21/100\n",
      "1756/1756 [==============================] - 0s - loss: 1.2615e-04     \n",
      "Epoch 22/100\n",
      "1756/1756 [==============================] - 0s - loss: 1.2385e-04     \n",
      "Epoch 23/100\n",
      "1756/1756 [==============================] - 0s - loss: 1.2148e-04     \n",
      "Epoch 24/100\n",
      "1756/1756 [==============================] - 0s - loss: 1.2155e-04     \n",
      "Epoch 25/100\n",
      "1756/1756 [==============================] - 0s - loss: 1.1795e-04     \n",
      "Epoch 26/100\n",
      "1756/1756 [==============================] - 0s - loss: 1.1788e-04     \n",
      "Epoch 27/100\n",
      "1756/1756 [==============================] - 0s - loss: 1.1539e-04     \n",
      "Epoch 28/100\n",
      "1756/1756 [==============================] - 0s - loss: 1.1321e-04     \n",
      "Epoch 29/100\n",
      "1756/1756 [==============================] - 0s - loss: 1.1164e-04     \n",
      "Epoch 30/100\n",
      "1756/1756 [==============================] - 0s - loss: 1.1127e-04     \n",
      "Epoch 31/100\n",
      "1756/1756 [==============================] - 0s - loss: 1.0902e-04     \n",
      "Epoch 32/100\n",
      "1756/1756 [==============================] - 0s - loss: 1.0835e-04     \n",
      "Epoch 33/100\n",
      "1756/1756 [==============================] - 0s - loss: 1.0673e-04     \n",
      "Epoch 34/100\n",
      "1756/1756 [==============================] - 0s - loss: 1.0602e-04     \n",
      "Epoch 35/100\n",
      "1756/1756 [==============================] - 0s - loss: 1.0460e-04     \n",
      "Epoch 36/100\n",
      "1756/1756 [==============================] - 0s - loss: 1.0408e-04     \n",
      "Epoch 37/100\n",
      "1756/1756 [==============================] - 0s - loss: 1.0271e-04     \n",
      "Epoch 38/100\n",
      "1756/1756 [==============================] - 0s - loss: 1.0225e-04     \n",
      "Epoch 39/100\n",
      "1756/1756 [==============================] - 0s - loss: 1.0206e-04     \n",
      "Epoch 40/100\n",
      "1756/1756 [==============================] - 0s - loss: 1.0156e-04     \n",
      "Epoch 41/100\n",
      "1756/1756 [==============================] - 0s - loss: 9.9167e-05     \n",
      "Epoch 42/100\n",
      "1756/1756 [==============================] - 0s - loss: 9.9631e-05     \n",
      "Epoch 43/100\n",
      "1756/1756 [==============================] - 0s - loss: 9.8027e-05     \n",
      "Epoch 44/100\n",
      "1756/1756 [==============================] - 0s - loss: 9.7440e-05     \n",
      "Epoch 45/100\n",
      "1756/1756 [==============================] - 0s - loss: 9.7561e-05     \n",
      "Epoch 46/100\n",
      "1756/1756 [==============================] - 0s - loss: 9.6687e-05     \n",
      "Epoch 47/100\n",
      "1756/1756 [==============================] - 0s - loss: 9.7849e-05     \n",
      "Epoch 48/100\n",
      "1756/1756 [==============================] - 0s - loss: 9.7156e-05     \n",
      "Epoch 49/100\n",
      "1756/1756 [==============================] - 0s - loss: 9.5147e-05     \n",
      "Epoch 50/100\n",
      "1756/1756 [==============================] - 0s - loss: 9.5845e-05     \n",
      "Epoch 51/100\n",
      "1756/1756 [==============================] - 0s - loss: 9.6749e-05     \n",
      "Epoch 52/100\n",
      "1756/1756 [==============================] - 0s - loss: 9.6565e-05     \n",
      "Epoch 53/100\n",
      "1756/1756 [==============================] - 0s - loss: 9.4321e-05     \n",
      "Epoch 54/100\n",
      "1756/1756 [==============================] - 0s - loss: 9.4437e-05     \n",
      "Epoch 55/100\n",
      "1756/1756 [==============================] - 0s - loss: 9.4141e-05     \n",
      "Epoch 56/100\n",
      "1756/1756 [==============================] - 0s - loss: 9.3355e-05     \n",
      "Epoch 57/100\n",
      "1756/1756 [==============================] - 0s - loss: 9.2018e-05     \n",
      "Epoch 58/100\n",
      "1756/1756 [==============================] - 0s - loss: 9.2348e-05     \n",
      "Epoch 59/100\n",
      "1756/1756 [==============================] - 0s - loss: 9.4049e-05     \n",
      "Epoch 60/100\n",
      "1756/1756 [==============================] - 0s - loss: 9.1578e-05     \n",
      "Epoch 61/100\n",
      "1756/1756 [==============================] - 0s - loss: 9.1822e-05     \n",
      "Epoch 62/100\n",
      "1756/1756 [==============================] - 0s - loss: 9.3079e-05     \n",
      "Epoch 63/100\n",
      "1756/1756 [==============================] - 0s - loss: 9.1892e-05     \n",
      "Epoch 64/100\n",
      "1756/1756 [==============================] - 0s - loss: 9.1110e-05     \n",
      "Epoch 65/100\n",
      "1756/1756 [==============================] - 0s - loss: 9.2623e-05     \n",
      "Epoch 66/100\n",
      "1756/1756 [==============================] - 0s - loss: 9.1186e-05     \n",
      "Epoch 67/100\n",
      "1756/1756 [==============================] - 0s - loss: 9.0814e-05     \n",
      "Epoch 68/100\n",
      "1756/1756 [==============================] - 0s - loss: 9.0554e-05     \n",
      "Epoch 69/100\n",
      "1756/1756 [==============================] - 0s - loss: 9.0855e-05     \n",
      "Epoch 70/100\n",
      "1756/1756 [==============================] - 0s - loss: 9.0454e-05     \n",
      "Epoch 71/100\n",
      "1756/1756 [==============================] - 0s - loss: 9.0162e-05     \n",
      "Epoch 72/100\n",
      "1756/1756 [==============================] - 0s - loss: 8.7955e-05     \n",
      "Epoch 73/100\n",
      "1756/1756 [==============================] - 0s - loss: 9.0543e-05     \n",
      "Epoch 74/100\n",
      "1756/1756 [==============================] - 0s - loss: 8.9320e-05     \n",
      "Epoch 75/100\n",
      "1756/1756 [==============================] - 0s - loss: 9.0249e-05     \n",
      "Epoch 76/100\n",
      "1756/1756 [==============================] - 0s - loss: 8.6588e-05     \n",
      "Epoch 77/100\n",
      "1756/1756 [==============================] - 0s - loss: 8.7728e-05     \n",
      "Epoch 78/100\n",
      "1756/1756 [==============================] - 0s - loss: 8.9248e-05     \n",
      "Epoch 79/100\n",
      "1756/1756 [==============================] - 0s - loss: 8.7452e-05     \n",
      "Epoch 80/100\n",
      "1756/1756 [==============================] - 0s - loss: 8.8095e-05     \n",
      "Epoch 81/100\n",
      "1756/1756 [==============================] - 0s - loss: 8.7210e-05     \n",
      "Epoch 82/100\n",
      "1756/1756 [==============================] - 0s - loss: 8.6662e-05     \n",
      "Epoch 83/100\n",
      "1756/1756 [==============================] - 0s - loss: 8.5006e-05     \n",
      "Epoch 84/100\n",
      "1756/1756 [==============================] - 0s - loss: 8.6366e-05     \n",
      "Epoch 85/100\n",
      "1756/1756 [==============================] - 0s - loss: 8.5888e-05     \n",
      "Epoch 86/100\n",
      "1756/1756 [==============================] - 0s - loss: 8.5586e-05     \n",
      "Epoch 87/100\n",
      "1756/1756 [==============================] - 0s - loss: 8.5922e-05     \n",
      "Epoch 88/100\n",
      "1756/1756 [==============================] - 0s - loss: 8.3764e-05     \n",
      "Epoch 89/100\n",
      "1756/1756 [==============================] - 0s - loss: 8.6982e-05     \n",
      "Epoch 90/100\n",
      "1756/1756 [==============================] - 0s - loss: 8.4982e-05     \n",
      "Epoch 91/100\n",
      "1756/1756 [==============================] - 0s - loss: 8.3283e-05     \n",
      "Epoch 92/100\n",
      "1756/1756 [==============================] - 0s - loss: 8.6153e-05     \n",
      "Epoch 93/100\n",
      "1756/1756 [==============================] - 0s - loss: 8.3700e-05     \n",
      "Epoch 94/100\n",
      "1756/1756 [==============================] - 0s - loss: 8.1712e-05     \n",
      "Epoch 95/100\n",
      "1756/1756 [==============================] - 0s - loss: 8.2159e-05     \n",
      "Epoch 96/100\n",
      "1756/1756 [==============================] - 0s - loss: 8.0967e-05     \n",
      "Epoch 97/100\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1756/1756 [==============================] - 0s - loss: 8.1732e-05     \n",
      "Epoch 98/100\n",
      "1756/1756 [==============================] - 0s - loss: 8.1559e-05     \n",
      "Epoch 99/100\n",
      "1756/1756 [==============================] - 0s - loss: 8.0210e-05     \n",
      "Epoch 100/100\n",
      "1756/1756 [==============================] - 0s - loss: 7.9433e-05     \n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.callbacks.History at 0x7f051a27ed30>"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#build the model\n",
    "model = Sequential()\n",
    "model.add(LSTM(4, input_shape=(series, 1)))\n",
    "model.add(Dense(1))\n",
    "model.compile(loss='mse', optimizer='adam')\n",
    "#fit the model\n",
    "model.fit(trainX, trainY, epochs=100, batch_size=32)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "#test this model out\n",
    "trainPredictions = model.predict(trainX)\n",
    "testPredictions = model.predict(testX)\n",
    "#unscale predictions\n",
    "trainPredictions = scaler.inverse_transform(trainPredictions)\n",
    "testPredictions = scaler.inverse_transform(testPredictions)\n",
    "trainY = scaler.inverse_transform([trainY])\n",
    "testY = scaler.inverse_transform([testY])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train score: %.2f rmse 0.7209341979280773\n",
      "Test score: %.2f rmse 4.938673008857972\n"
     ]
    }
   ],
   "source": [
    "#lets calculate the root mean squared error\n",
    "trainScore = math.sqrt(mean_squared_error(trainY[0], trainPredictions[:, 0]))\n",
    "testScore = math.sqrt(mean_squared_error(testY[0], testPredictions[:, 0]))\n",
    "print('Train score: %.2f rmse', trainScore)\n",
    "print('Test score: %.2f rmse', testScore)\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f051a695dd8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#lets plot the predictions on a graph and see how well it did\n",
    "train_plot = np.empty_like(msft_close)\n",
    "train_plot[:,:] = np.nan\n",
    "train_plot[series:len(trainPredictions)+series, :] = trainPredictions\n",
    "\n",
    "test_plot = np.empty_like(msft_close)\n",
    "test_plot[:,:] = np.nan\n",
    "test_plot[len(trainPredictions)+(series*2)+1:len(msft_close)-1, :] = testPredictions\n",
    "\n",
    "#plot on graph\n",
    "plt.plot(scaler.inverse_transform(msft_close))\n",
    "plt.plot(train_plot)\n",
    "plt.plot(test_plot)\n",
    "plt.show()\n"
   ]
  },
  {
   "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": "3.5.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}