Stock_Price_Predictor.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Final Year Major Project\n",
    "\n",
    "Stock Price Prediction\n",
    "\n",
    "The challenge of this project is to accurately predict the future closing value of a given stock across a given period of time in the future.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Use \n",
    "\n",
    "In the following cells we download and save the data set of the company we want.\n",
    "\n",
    "   **Step 1 :** Define a function to get historical data Alphavantage"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import datetime as dt\n",
    "from pandas_datareader import data\n",
    "import urllib.request, json \n",
    "import os"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "2. Download data\n",
    "You present code for using data from,\n",
    "\n",
    "Alpha Vantage. Before you start, however, you will first need an API key, which you can obtain for free here. After that, you can assigned that key to the api_key variable.\n",
    "Stock prices come in several different flavours. They are,\n",
    "\n",
    "Open: Opening stock price of the day\n",
    "Close: Closing stock price of the day\n",
    "High: Highest stock price of the data\n",
    "Low: Lowest stock price of the day\n",
    "Getting Data from Alphavantage\n",
    "You will first load in the data from Alpha Vantage. Since you're going to make use of the American Airlines Stock market prices to make your predictions, you set the ticker to \"AAL\". Additionally, you also define a url_string, which will return a JSON file with all the stock market data for American Airlines within last 20 years, and a file_to_save, which will be the file to which you save the data. You'll use the ticker variable that you defined beforehand to help name this file.\n",
    "\n",
    "Next, you're going to specify a condition: if you haven't already saved data, you will go ahead and grab the data from the URL that you set in url_string; You'll store the date, low, high, volume, close, open values to a pandas DataFrame df and you'll save it to file_to_save. However, if the data is already there, you'll just load it from the CSV."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "File already exists. Loading data from CSV\n"
     ]
    }
   ],
   "source": [
    "data_source = 'alphavantage'\n",
    "\n",
    "    # ====================== Loading Data from Alpha Vantage ==================================\n",
    "\n",
    "api_key = 'TFT45YBBY0XZUTOL'\n",
    "\n",
    "# American Airlines stock market prices\n",
    "ticker = \"MSFT\" \n",
    "\n",
    "# JSON file with all the stock market data for AAL within the last 20 years\n",
    "url_string = \"https://www.alphavantage.co/query?function=TIME_SERIES_DAILY&symbol=%s&outputsize=full&apikey=%s\"%(ticker,api_key)\n",
    "\n",
    "# Save data to this file\n",
    "file_to_save = 'stock_market_data-%s.csv'%ticker\n",
    "\n",
    "# If you haven't already saved data,\n",
    "# Go ahead and grab the data from the url\n",
    "# And store date, low, high, volume, close, open values to a Pandas dataframe\n",
    "if not os.path.exists(file_to_save):\n",
    "    with urllib.request.urlopen(url_string) as url:\n",
    "        data = json.loads(url.read().decode())\n",
    "        # extract stock market data\n",
    "        data = data['Time Series (Daily)']\n",
    "        df = pd.DataFrame(columns=['Date','Open','High','Low','Close','Volume'])\n",
    "        for k,v in data.items():\n",
    "            date = dt.datetime.strptime(k, '%Y-%m-%d')\n",
    "            data_row = [date.date(),float(v['3. low']),float(v['2. high']),\n",
    "                        float(v['4. close']),float(v['1. open']),float(v['5. volume'])]\n",
    "            df.loc[-1,:] = data_row\n",
    "            df.index = df.index + 1\n",
    "    print('Data saved to : %s'%file_to_save) \n",
    "    df = df.sort_values('Date')\n",
    "    df.index = df['Date']\n",
    "    del df['Date']\n",
    "    df.to_csv(file_to_save)\n",
    "\n",
    "# If the data is already there, just load it from the CSV\n",
    "else:\n",
    "    print('File already exists. Loading data from CSV')\n",
    "    df = pd.read_csv(file_to_save,index_col=\"Date\",parse_dates=True)\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Preprocess the data\n",
    "\n",
    "Now it is time to preprocess the data. In the following cells we will normalise it for better prediction of data.\n",
    "\n",
    "**Step 1 :** Get the data from csv file."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              Open    High     Low   Close      Volume\n",
      "Date                                                  \n",
      "1998-01-02  129.50  131.50  131.13  129.63   4968500.0\n",
      "1998-01-05  127.87  133.63  130.38  131.25  10047200.0\n",
      "1998-01-06  129.25  133.00  131.13  129.75   8479300.0\n",
      "1998-01-07  127.50  131.19  129.56  129.88   7686600.0\n",
      "1998-01-08  127.50  132.13  130.50  128.63   9702400.0\n",
      "1998-01-09  125.87  131.50  127.00  130.06  10941800.0\n",
      "1998-01-12  124.37  130.00  129.50  124.62   9799200.0\n",
      "1998-01-13  128.13  132.25  132.13  129.50   8368200.0\n",
      "1998-01-14  129.25  132.50  131.13  132.13   6846800.0\n",
      "1998-01-15  129.88  133.00  132.31  130.38   6327800.0\n",
      "--------------------------------------------------\n",
      "                Open    High     Low   Close      Volume\n",
      "Date                                                    \n",
      "2019-04-08  118.6400  120.02  119.93  119.81  15116186.0\n",
      "2019-04-09  118.5800  119.54  119.28  118.63  17611981.0\n",
      "2019-04-10  119.5400  120.35  120.19  119.76  16477169.0\n",
      "2019-04-11  119.9200  120.85  120.33  120.54  14209121.0\n",
      "2019-04-12  120.3700  120.98  120.95  120.64  19745143.0\n",
      "2019-04-15  120.5700  121.58  121.05  120.94  15792572.0\n",
      "2019-04-16  120.1000  121.65  120.77  121.64  14071787.0\n",
      "2019-04-17  120.5400  121.85  121.77  121.24  19300938.0\n",
      "2019-04-18  121.3018  123.52  123.37  122.19  27990998.0\n",
      "2019-04-22  122.5700  124.00  123.76  122.62  14823625.0\n"
     ]
    }
   ],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "print(df.head(10))\n",
    "print(\"--------------------------------------------------\")\n",
    "print(df.tail(10))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Step 2 :** Remove Unncessary data, i.e., Date and High value"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "   Item      Open   Close      Volume\n",
      "0     0  122.5700  122.62  14823625.0\n",
      "1     1  121.3018  122.19  27990998.0\n",
      "2     2  120.5400  121.24  19300938.0\n",
      "3     3  120.1000  121.64  14071787.0\n",
      "4     4  120.5700  120.94  15792572.0\n",
      "--------------------------------------------------\n",
      "      Item    Open   Close      Volume\n",
      "5354  5354  127.50  128.63   9702400.0\n",
      "5355  5355  127.50  129.88   7686600.0\n",
      "5356  5356  129.25  129.75   8479300.0\n",
      "5357  5357  127.87  131.25  10047200.0\n",
      "5358  5358  129.50  129.63   4968500.0\n"
     ]
    }
   ],
   "source": [
    "from Include import preprocess_data as ppd\n",
    "stocks = ppd.remove_data(df)\n",
    "\n",
    "#Print the dataframe head and tail\n",
    "print(stocks.head())\n",
    "print(\"--------------------------------------------------\")\n",
    "print(stocks.tail())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Step 2: ** Visualise raw data."
   ]
  },
  {
   "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": [
    "from Include import visualize\n",
    "visualize.plot_basic(stocks,ticker)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Step 3 :** Normalise the data using minmaxscaler function"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "   Item      Open     Close    Volume\n",
      "0     0  0.669443  0.656040  0.019342\n",
      "1     1  0.661560  0.653414  0.041751\n",
      "2     2  0.656825  0.647612  0.026962\n",
      "3     3  0.654090  0.650055  0.018063\n",
      "4     4  0.657011  0.645780  0.020991\n",
      "5     5  0.655768  0.643948  0.027718\n",
      "6     6  0.652971  0.643337  0.018297\n",
      "7     7  0.650609  0.638573  0.022157\n",
      "8     8  0.644642  0.631672  0.024088\n",
      "9     9  0.645015  0.638879  0.019840\n",
      "\n",
      "\n",
      "Open   --- mean : 0.213661154122   \t Std:  0.184139922502   \t Max:  1.0   \t Min:  0.0\n",
      "Close  --- mean : 0.211616618498   \t Std:  0.183773373128   \t Max:  1.0   \t Min:  0.0\n",
      "Volume --- mean : 0.0745209105311   \t Std:  0.0489334348103   \t Max:  1.0   \t Min:  0.0\n"
     ]
    }
   ],
   "source": [
    "stocks = ppd.get_normalised_data(stocks)\n",
    "print(stocks.head(10))\n",
    "\n",
    "print(\"\\n\")\n",
    "print(\"Open   --- mean :\", np.mean(stocks['Open']),  \"  \\t Std: \", np.std(stocks['Open']),  \"  \\t Max: \", np.max(stocks['Open']),  \"  \\t Min: \", np.min(stocks['Open']))\n",
    "print(\"Close  --- mean :\", np.mean(stocks['Close']), \"  \\t Std: \", np.std(stocks['Close']), \"  \\t Max: \", np.max(stocks['Close']), \"  \\t Min: \", np.min(stocks['Close']))\n",
    "print(\"Volume --- mean :\", np.mean(stocks['Volume']),\"  \\t Std: \", np.std(stocks['Volume']),\"  \\t Max: \", np.max(stocks['Volume']),\"  \\t Min: \", np.min(stocks['Volume']))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Step 4 :** Visualize the data again"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "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": [
    "visualize.plot_basic(stocks,ticker)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "**Step 5:** Log the normalised data for future resuablilty"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "file_to_save_preprocessed = 'stock_market_data-%s_preprocessed.csv'%ticker\n",
    "stocks.to_csv(file_to_save_preprocessed,index= False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Bench Mark Model\n",
    "\n",
    "In this section we will check our bench mark model. As is proposed in my proposal my bench mark model is a simple linear regressor model. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Step 1:** Load the preprocessed data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "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>Item</th>\n",
       "      <th>Open</th>\n",
       "      <th>Close</th>\n",
       "      <th>Volume</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0.669443</td>\n",
       "      <td>0.656040</td>\n",
       "      <td>0.019342</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>0.661560</td>\n",
       "      <td>0.653414</td>\n",
       "      <td>0.041751</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>0.656825</td>\n",
       "      <td>0.647612</td>\n",
       "      <td>0.026962</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3</td>\n",
       "      <td>0.654090</td>\n",
       "      <td>0.650055</td>\n",
       "      <td>0.018063</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4</td>\n",
       "      <td>0.657011</td>\n",
       "      <td>0.645780</td>\n",
       "      <td>0.020991</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Item      Open     Close    Volume\n",
       "0     0  0.669443  0.656040  0.019342\n",
       "1     1  0.661560  0.653414  0.041751\n",
       "2     2  0.656825  0.647612  0.026962\n",
       "3     3  0.654090  0.650055  0.018063\n",
       "4     4  0.657011  0.645780  0.020991"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import math\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "from IPython.display import display\n",
    "from sklearn import linear_model\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.metrics import mean_squared_error\n",
    "from sklearn.model_selection import TimeSeriesSplit\n",
    "\n",
    "from Include import visualize as vs\n",
    "from Include import stock_data as sd\n",
    "from Include import LinearRegressionModel\n",
    "\n",
    "stocks = pd.read_csv(file_to_save_preprocessed)\n",
    "display(stocks.head())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Step 2:** Split data into train and test pair"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "X_train (3671, 1)\n",
      "y_train (3671, 1)\n",
      "X_test (1688, 1)\n",
      "y_test (1688, 1)\n"
     ]
    }
   ],
   "source": [
    "X_train, X_test, y_train, y_test, label_range= sd.train_test_split_linear_regression(stocks)\n",
    "\n",
    "print(\"X_train\", X_train.shape)\n",
    "print(\"y_train\", y_train.shape)\n",
    "print(\"X_test\", X_test.shape)\n",
    "print(\"y_test\", y_test.shape)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Step 3:** Train a Linear regressor model on training set and get prediction"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = LinearRegressionModel.build_model(X_train,y_train)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Step 4:** Get prediction on test set"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([-0.0398898 , -0.0399992 , -0.0401086 , -0.04021799, -0.04032739,\n",
       "       -0.04043678, -0.04054618, -0.04065557, -0.04076497, -0.04087437])"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "predictions = LinearRegressionModel.predict_prices(model,X_test, label_range)\n",
    "predictions[:10]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Step 5:** Plot the predicted values against actual"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "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": [
    "vs.plot_prediction(y_test,predictions)         "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Step 6:** measure accuracy of the prediction"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train Score: 0.5563 MSE (0.7459 RMSE)\n",
      "Test Score: 0.26617422 MSE (0.51592075 RMSE)\n"
     ]
    }
   ],
   "source": [
    "trainScore = mean_squared_error(X_train, y_train)\n",
    "print('Train Score: %.4f MSE (%.4f RMSE)' % (trainScore, math.sqrt(trainScore)))\n",
    "\n",
    "testScore = mean_squared_error(predictions, y_test)\n",
    "print('Test Score: %.8f MSE (%.8f RMSE)' % (testScore, math.sqrt(testScore)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "\n",
    "\n",
    "\n",
    "## Long-Sort Term Memory Model\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Basic LSTM Model\n",
    "\n",
    "First lets make a basic LSTM model."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Step 1 :** import keras libraries for smooth implementaion of lstm "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    },
    {
     "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>Open</th>\n",
       "      <th>Close</th>\n",
       "      <th>Volume</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.669443</td>\n",
       "      <td>0.656040</td>\n",
       "      <td>0.019342</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.661560</td>\n",
       "      <td>0.653414</td>\n",
       "      <td>0.041751</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.656825</td>\n",
       "      <td>0.647612</td>\n",
       "      <td>0.026962</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.654090</td>\n",
       "      <td>0.650055</td>\n",
       "      <td>0.018063</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.657011</td>\n",
       "      <td>0.645780</td>\n",
       "      <td>0.020991</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       Open     Close    Volume\n",
       "0  0.669443  0.656040  0.019342\n",
       "1  0.661560  0.653414  0.041751\n",
       "2  0.656825  0.647612  0.026962\n",
       "3  0.654090  0.650055  0.018063\n",
       "4  0.657011  0.645780  0.020991"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import math\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "from IPython.display import display\n",
    "\n",
    "from keras.layers.core import Dense, Activation, Dropout\n",
    "from keras.layers.recurrent import LSTM\n",
    "from keras.models import Sequential\n",
    "from keras.metrics import mean_squared_error\n",
    "from sklearn.model_selection import StratifiedKFold\n",
    "\n",
    "import time #helper libraries\n",
    "\n",
    "from Include import visualize as vs\n",
    "from Include import stock_data as sd\n",
    "from Include import LinearRegressionModel\n",
    "from Include import lstm\n",
    "stocks = pd.read_csv(file_to_save_preprocessed)\n",
    "stocks_data = stocks.drop(['Item'], axis =1)\n",
    "display(stocks_data.head())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Step 2 :** Split train and test data sets and Unroll train and test data for lstm model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "x_train (4807, 50, 3)\n",
      "y_train (4807,)\n",
      "x_test (450, 50, 3)\n",
      "y_test (450,)\n"
     ]
    }
   ],
   "source": [
    "X_train, X_test,y_train, y_test = sd.train_test_split_lstm(stocks_data,1)\n",
    "\n",
    "unroll_length = 50\n",
    "X_train = sd.unroll(X_train, unroll_length)\n",
    "X_test = sd.unroll(X_test, unroll_length)\n",
    "y_train = y_train[-X_train.shape[0]:]\n",
    "y_test = y_test[-X_test.shape[0]:]\n",
    "\n",
    "print(\"x_train\", X_train.shape)\n",
    "print(\"y_train\", y_train.shape)\n",
    "print(\"x_test\", X_test.shape)\n",
    "print(\"y_test\", y_test.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Step 3 :** Build a basic Long-Short Term Memory model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "compilation time :  0.16766786575317383\n"
     ]
    }
   ],
   "source": [
    "# build basic lstm model\n",
    "model = lstm.build_basic_model(input_dim = X_train.shape[-1],output_dim = unroll_length, return_sequences=True)\n",
    "\n",
    "# Compile the model\n",
    "start = time.time()\n",
    "model.compile(loss='mean_squared_error', optimizer='adam')\n",
    "print('compilation time : ', time.time() - start)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Step 4:** Train the model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 4566 samples, validate on 241 samples\n",
      "Epoch 1/1\n",
      "4566/4566 [==============================] - 18s 4ms/step - loss: 0.0013 - val_loss: 0.0020\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.callbacks.History at 0x29ce8ae3710>"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.fit(\n",
    "    X_train,\n",
    "    y_train,\n",
    "    epochs=1,\n",
    "    verbose=1,\n",
    "    validation_split=0.05)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Step 5:** make prediction using test data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "predictions = model.predict(X_test)\n",
    "model.save('lstm_model.h5')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Step 6:** Plot the results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "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": [
    "vs.plot_lstm_prediction(y_test,predictions)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "** Step 7:** Get the test score."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train Score: 0.00025197 MSE (0.01587361 RMSE)\n",
      "Test Score: 0.00736598 MSE (0.08582531 RMSE)\n"
     ]
    }
   ],
   "source": [
    "trainScore = model.evaluate(X_train, y_train, verbose=0)\n",
    "print('Train Score: %.8f MSE (%.8f RMSE)' % (trainScore, math.sqrt(trainScore)))\n",
    "\n",
    "\n",
    "testScore = model.evaluate(X_test, y_test, verbose=0)\n",
    "print('Test Score: %.8f MSE (%.8f RMSE)' % (testScore, math.sqrt(testScore)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Improved LSTM Model\n",
    "\n",
    "**Step 1: ** Build an improved LSTM model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "compilation time :  0.015590190887451172\n"
     ]
    }
   ],
   "source": [
    "# Set up hyperparameters\n",
    "batch_size = 128\n",
    "\n",
    "# build improved lstm model\n",
    "model = lstm.build_improved_model( X_train.shape[-1],output_dim = unroll_length, return_sequences=True)\n",
    "\n",
    "start = time.time()\n",
    "\n",
    "#final_model.compile(loss='mean_squared_error', optimizer='adam')\n",
    "\n",
    "model.compile(loss='mean_squared_error', optimizer='adam')\n",
    "print('compilation time : ', time.time() - start)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Step 2: ** Train improved LSTM model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 716 samples, validate on 8 samples\n",
      "Epoch 1/30\n",
      "716/716 [==============================] - 3s 3ms/step - loss: 0.1505 - val_loss: 0.0250\n",
      "Epoch 2/30\n",
      "716/716 [==============================] - 2s 2ms/step - loss: 0.0266 - val_loss: 3.6007e-04\n",
      "Epoch 3/30\n",
      "716/716 [==============================] - 2s 2ms/step - loss: 0.0192 - val_loss: 9.7612e-05\n",
      "Epoch 4/30\n",
      "716/716 [==============================] - ETA: 0s - loss: 0.006 - 1s 2ms/step - loss: 0.0069 - val_loss: 0.0093\n",
      "Epoch 5/30\n",
      "716/716 [==============================] - 2s 2ms/step - loss: 0.0079 - val_loss: 0.0029\n",
      "Epoch 6/30\n",
      "716/716 [==============================] - 2s 2ms/step - loss: 0.0055 - val_loss: 7.2052e-04\n",
      "Epoch 7/30\n",
      "716/716 [==============================] - 2s 2ms/step - loss: 0.0049 - val_loss: 0.0031\n",
      "Epoch 8/30\n",
      "716/716 [==============================] - 2s 2ms/step - loss: 0.0048 - val_loss: 0.0029\n",
      "Epoch 9/30\n",
      "716/716 [==============================] - 2s 2ms/step - loss: 0.0045 - val_loss: 6.5802e-04\n",
      "Epoch 10/30\n",
      "716/716 [==============================] - 2s 2ms/step - loss: 0.0044 - val_loss: 0.0013\n",
      "Epoch 11/30\n",
      "716/716 [==============================] - 2s 2ms/step - loss: 0.0039 - val_loss: 8.9084e-04\n",
      "Epoch 12/30\n",
      "716/716 [==============================] - 2s 2ms/step - loss: 0.0040 - val_loss: 6.1070e-04\n",
      "Epoch 13/30\n",
      "716/716 [==============================] - 2s 2ms/step - loss: 0.0035 - val_loss: 5.1722e-04\n",
      "Epoch 14/30\n",
      "716/716 [==============================] - 2s 2ms/step - loss: 0.0035 - val_loss: 2.0833e-04\n",
      "Epoch 15/30\n",
      "716/716 [==============================] - 2s 2ms/step - loss: 0.0033 - val_loss: 3.9835e-04\n",
      "Epoch 16/30\n",
      "716/716 [==============================] - 2s 2ms/step - loss: 0.0033 - val_loss: 1.6807e-04\n",
      "Epoch 17/30\n",
      "716/716 [==============================] - 2s 2ms/step - loss: 0.0033 - val_loss: 1.2623e-04\n",
      "Epoch 18/30\n",
      "716/716 [==============================] - 2s 2ms/step - loss: 0.0034 - val_loss: 1.7712e-04\n",
      "Epoch 19/30\n",
      "716/716 [==============================] - 2s 2ms/step - loss: 0.0031 - val_loss: 9.1364e-05\n",
      "Epoch 20/30\n",
      "716/716 [==============================] - 2s 2ms/step - loss: 0.0030 - val_loss: 1.4275e-04\n",
      "Epoch 21/30\n",
      "716/716 [==============================] - 2s 2ms/step - loss: 0.0033 - val_loss: 8.2389e-05\n",
      "Epoch 22/30\n",
      "716/716 [==============================] - 2s 2ms/step - loss: 0.0035 - val_loss: 9.5690e-05\n",
      "Epoch 23/30\n",
      "716/716 [==============================] - 2s 2ms/step - loss: 0.0032 - val_loss: 1.0163e-04\n",
      "Epoch 24/30\n",
      "716/716 [==============================] - 2s 2ms/step - loss: 0.0030 - val_loss: 9.2132e-05\n",
      "Epoch 25/30\n",
      "716/716 [==============================] - 2s 2ms/step - loss: 0.0032 - val_loss: 8.9658e-05\n",
      "Epoch 26/30\n",
      "716/716 [==============================] - 1s 2ms/step - loss: 0.0033 - val_loss: 1.1084e-04\n",
      "Epoch 27/30\n",
      "716/716 [==============================] - 1s 2ms/step - loss: 0.0028 - val_loss: 7.9078e-05\n",
      "Epoch 28/30\n",
      "716/716 [==============================] - 2s 2ms/step - loss: 0.0029 - val_loss: 1.1161e-04\n",
      "Epoch 29/30\n",
      "716/716 [==============================] - 2s 2ms/step - loss: 0.0033 - val_loss: 8.0666e-05\n",
      "Epoch 30/30\n",
      "716/716 [==============================] - 2s 2ms/step - loss: 0.0031 - val_loss: 1.2491e-04\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.callbacks.History at 0x1ae9b666cc0>"
      ]
     },
     "execution_count": 56,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.fit(X_train, \n",
    "          y_train, \n",
    "          batch_size=batch_size,\n",
    "          epochs=30,\n",
    "          verbose=1,\n",
    "          validation_split=0.01\n",
    "         )"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Step 3:** Make prediction on improved LSTM model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Generate predictions \n",
    "model.save('lstm_model_improved.h5')\n",
    "predictions = model.predict(X_test, batch_size=batch_size)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Step 4:** plot the results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "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": [
    "vs.plot_lstm_prediction(y_test,predictions)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Step 5:** Get the test score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train Score: 0.00143191 MSE (0.03784054 RMSE)\n",
      "Test Score: 0.00099017 MSE (0.03146702 RMSE)\n"
     ]
    }
   ],
   "source": [
    "trainScore = model.evaluate(X_train, y_train, verbose=0)\n",
    "print('Train Score: %.8f MSE (%.8f RMSE)' % (trainScore, math.sqrt(trainScore)))\n",
    "\n",
    "testScore = model.evaluate(X_test, y_test, verbose=0)\n",
    "print('Test Score: %.8f MSE (%.8f RMSE)' % (testScore, math.sqrt(testScore)))\n",
    "#print(\"%s: %.2f%%\" % (model.metrics_names[1], testScore[1]*100))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Delta Price: 0.000990 - RMSE * Adjusted Close Range\n"
     ]
    }
   ],
   "source": [
    "range = [np.amin(stocks_data['Close']), np.amax(stocks_data['Close'])]\n",
    "\n",
    "#Calculate the stock price delta in $\n",
    "\n",
    "true_delta = testScore*(range[1]-range[0])\n",
    "print('Delta Price: %.6f - RMSE * Adjusted Close Range' % true_delta)    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1/1 [==============================] - 0s 328ms/step\n",
      "[[0.09404148161411285]]\n"
     ]
    }
   ],
   "source": [
    "# load model from single file\n",
    "from keras.models import load_model\n",
    "model = load_model('lstm_model_improved.h5')\n",
    "# make predictions\n",
    "yhat = model.predict(X_test[-1:], verbose=1)\n",
    "predict_nextDay=yhat.tolist()\n",
    "last_day=y_test[-1:].tolist()\n",
    "\n",
    "print(predict_nextDay)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0.09448058221163136]\n"
     ]
    }
   ],
   "source": [
    "print(last_day)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "If you buy it, it will be beneficial for you.\n"
     ]
    }
   ],
   "source": [
    "if(predict_nextDay[0] > last_day):\n",
    "    print('Either sell it or wait. We\\'ll assert you to WAIT. ')\n",
    "else:\n",
    "    print('If you buy it, it will be beneficial for you.')"
   ]
  }
 ],
 "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": "3.6.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}