stock_price_prediction_LSTM.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Stock Price Prediction of Microsoft"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "%matplotlib inline\n",
    "import matplotlib.pyplot as plt\n",
    "import matplotlib\n",
    "from sklearn.preprocessing import MinMaxScaler\n",
    "from keras.layers import LSTM,Dense,Dropout\n",
    "from sklearn.model_selection import TimeSeriesSplit\n",
    "from sklearn.metrics import mean_squared_error, r2_score\n",
    "import matplotlib.dates as mdates\n",
    "from sklearn import linear_model\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Reading data using parse date\n",
    "As it is timeseries data so we have to read the data by parsing date means making data column as index."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "df_final = pd.read_csv(\"MSFT.csv\",na_values=['null'],index_col='Date',parse_dates=True,infer_datetime_format=True)"
   ]
  },
  {
   "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>Open</th>\n",
       "      <th>High</th>\n",
       "      <th>Low</th>\n",
       "      <th>Close</th>\n",
       "      <th>Adj Close</th>\n",
       "      <th>Volume</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1990-01-02</th>\n",
       "      <td>0.605903</td>\n",
       "      <td>0.616319</td>\n",
       "      <td>0.598090</td>\n",
       "      <td>0.616319</td>\n",
       "      <td>0.447268</td>\n",
       "      <td>53033600</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1990-01-03</th>\n",
       "      <td>0.621528</td>\n",
       "      <td>0.626736</td>\n",
       "      <td>0.614583</td>\n",
       "      <td>0.619792</td>\n",
       "      <td>0.449788</td>\n",
       "      <td>113772800</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1990-01-04</th>\n",
       "      <td>0.619792</td>\n",
       "      <td>0.638889</td>\n",
       "      <td>0.616319</td>\n",
       "      <td>0.638021</td>\n",
       "      <td>0.463017</td>\n",
       "      <td>125740800</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1990-01-05</th>\n",
       "      <td>0.635417</td>\n",
       "      <td>0.638889</td>\n",
       "      <td>0.621528</td>\n",
       "      <td>0.622396</td>\n",
       "      <td>0.451678</td>\n",
       "      <td>69564800</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1990-01-08</th>\n",
       "      <td>0.621528</td>\n",
       "      <td>0.631944</td>\n",
       "      <td>0.614583</td>\n",
       "      <td>0.631944</td>\n",
       "      <td>0.458607</td>\n",
       "      <td>58982400</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                Open      High       Low     Close  Adj Close     Volume\n",
       "Date                                                                    \n",
       "1990-01-02  0.605903  0.616319  0.598090  0.616319   0.447268   53033600\n",
       "1990-01-03  0.621528  0.626736  0.614583  0.619792   0.449788  113772800\n",
       "1990-01-04  0.619792  0.638889  0.616319  0.638021   0.463017  125740800\n",
       "1990-01-05  0.635417  0.638889  0.621528  0.622396   0.451678   69564800\n",
       "1990-01-08  0.621528  0.631944  0.614583  0.631944   0.458607   58982400"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_final.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(7334, 6)"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\n",
    "df_final.shape"
   ]
  },
  {
   "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>Open</th>\n",
       "      <th>High</th>\n",
       "      <th>Low</th>\n",
       "      <th>Close</th>\n",
       "      <th>Adj Close</th>\n",
       "      <th>Volume</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>7334.000000</td>\n",
       "      <td>7334.000000</td>\n",
       "      <td>7334.000000</td>\n",
       "      <td>7334.000000</td>\n",
       "      <td>7334.000000</td>\n",
       "      <td>7.334000e+03</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>28.578362</td>\n",
       "      <td>28.891892</td>\n",
       "      <td>28.261203</td>\n",
       "      <td>28.582426</td>\n",
       "      <td>24.008891</td>\n",
       "      <td>6.022760e+07</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>22.279033</td>\n",
       "      <td>22.476021</td>\n",
       "      <td>22.054379</td>\n",
       "      <td>22.275692</td>\n",
       "      <td>21.941225</td>\n",
       "      <td>3.400345e+07</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>0.591146</td>\n",
       "      <td>0.604167</td>\n",
       "      <td>0.583333</td>\n",
       "      <td>0.598090</td>\n",
       "      <td>0.434039</td>\n",
       "      <td>3.705600e+06</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>12.457031</td>\n",
       "      <td>12.595703</td>\n",
       "      <td>12.261719</td>\n",
       "      <td>12.488281</td>\n",
       "      <td>9.062849</td>\n",
       "      <td>3.821840e+07</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>26.980000</td>\n",
       "      <td>27.235000</td>\n",
       "      <td>26.730000</td>\n",
       "      <td>26.990000</td>\n",
       "      <td>20.431164</td>\n",
       "      <td>5.429885e+07</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>34.716562</td>\n",
       "      <td>35.099998</td>\n",
       "      <td>34.250000</td>\n",
       "      <td>34.687500</td>\n",
       "      <td>26.993589</td>\n",
       "      <td>7.367888e+07</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>115.419998</td>\n",
       "      <td>116.180000</td>\n",
       "      <td>114.930000</td>\n",
       "      <td>115.610001</td>\n",
       "      <td>115.112709</td>\n",
       "      <td>7.645040e+08</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              Open         High          Low        Close    Adj Close  \\\n",
       "count  7334.000000  7334.000000  7334.000000  7334.000000  7334.000000   \n",
       "mean     28.578362    28.891892    28.261203    28.582426    24.008891   \n",
       "std      22.279033    22.476021    22.054379    22.275692    21.941225   \n",
       "min       0.591146     0.604167     0.583333     0.598090     0.434039   \n",
       "25%      12.457031    12.595703    12.261719    12.488281     9.062849   \n",
       "50%      26.980000    27.235000    26.730000    26.990000    20.431164   \n",
       "75%      34.716562    35.099998    34.250000    34.687500    26.993589   \n",
       "max     115.419998   116.180000   114.930000   115.610001   115.112709   \n",
       "\n",
       "             Volume  \n",
       "count  7.334000e+03  \n",
       "mean   6.022760e+07  \n",
       "std    3.400345e+07  \n",
       "min    3.705600e+06  \n",
       "25%    3.821840e+07  \n",
       "50%    5.429885e+07  \n",
       "75%    7.367888e+07  \n",
       "max    7.645040e+08  "
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_final.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "False"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_final.isnull().values.any()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x555def0>"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "df_final['Adj Close'].plot()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Correlation Analysis"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "X=df_final.drop(['Adj Close'],axis=1)\n",
    "X=X.drop(['Close'],axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0xb340978>"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x720 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "X.corrwith(df_final['Adj Close']).plot.bar(\n",
    "        figsize = (20, 10), title = \"Correlation with Adj Close\", fontsize = 20,\n",
    "        rot = 90, grid = True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "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>Open</th>\n",
       "      <th>High</th>\n",
       "      <th>Low</th>\n",
       "      <th>Close</th>\n",
       "      <th>Adj Close</th>\n",
       "      <th>Volume</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1990-01-02</th>\n",
       "      <td>0.605903</td>\n",
       "      <td>0.616319</td>\n",
       "      <td>0.598090</td>\n",
       "      <td>0.616319</td>\n",
       "      <td>0.447268</td>\n",
       "      <td>53033600</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1990-01-03</th>\n",
       "      <td>0.621528</td>\n",
       "      <td>0.626736</td>\n",
       "      <td>0.614583</td>\n",
       "      <td>0.619792</td>\n",
       "      <td>0.449788</td>\n",
       "      <td>113772800</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1990-01-04</th>\n",
       "      <td>0.619792</td>\n",
       "      <td>0.638889</td>\n",
       "      <td>0.616319</td>\n",
       "      <td>0.638021</td>\n",
       "      <td>0.463017</td>\n",
       "      <td>125740800</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1990-01-05</th>\n",
       "      <td>0.635417</td>\n",
       "      <td>0.638889</td>\n",
       "      <td>0.621528</td>\n",
       "      <td>0.622396</td>\n",
       "      <td>0.451678</td>\n",
       "      <td>69564800</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1990-01-08</th>\n",
       "      <td>0.621528</td>\n",
       "      <td>0.631944</td>\n",
       "      <td>0.614583</td>\n",
       "      <td>0.631944</td>\n",
       "      <td>0.458607</td>\n",
       "      <td>58982400</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                Open      High       Low     Close  Adj Close     Volume\n",
       "Date                                                                    \n",
       "1990-01-02  0.605903  0.616319  0.598090  0.616319   0.447268   53033600\n",
       "1990-01-03  0.621528  0.626736  0.614583  0.619792   0.449788  113772800\n",
       "1990-01-04  0.619792  0.638889  0.616319  0.638021   0.463017  125740800\n",
       "1990-01-05  0.635417  0.638889  0.621528  0.622396   0.451678   69564800\n",
       "1990-01-08  0.621528  0.631944  0.614583  0.631944   0.458607   58982400"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "test = df_final\n",
    "# Target column\n",
    "target_adj_close = pd.DataFrame(test['Adj Close'])\n",
    "display(test.head())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "# selecting Feature Columns\n",
    "feature_columns = ['Open', 'High', 'Low', 'Volume']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Normalizing the data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "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>Open</th>\n",
       "      <th>High</th>\n",
       "      <th>Low</th>\n",
       "      <th>Volume</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1990-01-02</th>\n",
       "      <td>0.000129</td>\n",
       "      <td>0.000105</td>\n",
       "      <td>0.000129</td>\n",
       "      <td>0.064837</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1990-01-03</th>\n",
       "      <td>0.000265</td>\n",
       "      <td>0.000195</td>\n",
       "      <td>0.000273</td>\n",
       "      <td>0.144673</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1990-01-04</th>\n",
       "      <td>0.000249</td>\n",
       "      <td>0.000300</td>\n",
       "      <td>0.000288</td>\n",
       "      <td>0.160404</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1990-01-05</th>\n",
       "      <td>0.000386</td>\n",
       "      <td>0.000300</td>\n",
       "      <td>0.000334</td>\n",
       "      <td>0.086566</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1990-01-08</th>\n",
       "      <td>0.000265</td>\n",
       "      <td>0.000240</td>\n",
       "      <td>0.000273</td>\n",
       "      <td>0.072656</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                Open      High       Low    Volume\n",
       "Date                                              \n",
       "1990-01-02  0.000129  0.000105  0.000129  0.064837\n",
       "1990-01-03  0.000265  0.000195  0.000273  0.144673\n",
       "1990-01-04  0.000249  0.000300  0.000288  0.160404\n",
       "1990-01-05  0.000386  0.000300  0.000334  0.086566\n",
       "1990-01-08  0.000265  0.000240  0.000273  0.072656"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.preprocessing import MinMaxScaler\n",
    "scaler = MinMaxScaler()\n",
    "feature_minmax_transform_data = scaler.fit_transform(test[feature_columns])\n",
    "feature_minmax_transform = pd.DataFrame(columns=feature_columns, data=feature_minmax_transform_data, index=test.index)\n",
    "feature_minmax_transform.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "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>Open</th>\n",
       "      <th>High</th>\n",
       "      <th>Low</th>\n",
       "      <th>Volume</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1990-01-02</th>\n",
       "      <td>0.000129</td>\n",
       "      <td>0.000105</td>\n",
       "      <td>0.000129</td>\n",
       "      <td>0.064837</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1990-01-03</th>\n",
       "      <td>0.000265</td>\n",
       "      <td>0.000195</td>\n",
       "      <td>0.000273</td>\n",
       "      <td>0.144673</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1990-01-04</th>\n",
       "      <td>0.000249</td>\n",
       "      <td>0.000300</td>\n",
       "      <td>0.000288</td>\n",
       "      <td>0.160404</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1990-01-05</th>\n",
       "      <td>0.000386</td>\n",
       "      <td>0.000300</td>\n",
       "      <td>0.000334</td>\n",
       "      <td>0.086566</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1990-01-08</th>\n",
       "      <td>0.000265</td>\n",
       "      <td>0.000240</td>\n",
       "      <td>0.000273</td>\n",
       "      <td>0.072656</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                Open      High       Low    Volume\n",
       "Date                                              \n",
       "1990-01-02  0.000129  0.000105  0.000129  0.064837\n",
       "1990-01-03  0.000265  0.000195  0.000273  0.144673\n",
       "1990-01-04  0.000249  0.000300  0.000288  0.160404\n",
       "1990-01-05  0.000386  0.000300  0.000334  0.086566\n",
       "1990-01-08  0.000265  0.000240  0.000273  0.072656"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Shape of features :  (7334, 4)\n",
      "Shape of target :  (7334, 1)\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>High</th>\n",
       "      <th>Low</th>\n",
       "      <th>Volume</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2019-02-01</th>\n",
       "      <td>0.898632</td>\n",
       "      <td>0.895480</td>\n",
       "      <td>0.889984</td>\n",
       "      <td>0.041838</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2019-02-04</th>\n",
       "      <td>0.890707</td>\n",
       "      <td>0.910189</td>\n",
       "      <td>0.893657</td>\n",
       "      <td>0.036290</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2019-02-05</th>\n",
       "      <td>0.918487</td>\n",
       "      <td>0.922908</td>\n",
       "      <td>0.921554</td>\n",
       "      <td>0.031046</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2019-02-06</th>\n",
       "      <td>0.926673</td>\n",
       "      <td>0.920572</td>\n",
       "      <td>0.917794</td>\n",
       "      <td>0.022219</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2019-02-07</th>\n",
       "      <td>0.910911</td>\n",
       "      <td>0.908372</td>\n",
       "      <td>0.906950</td>\n",
       "      <td>0.034247</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                Open      High       Low    Volume\n",
       "Date                                              \n",
       "2019-02-01  0.898632  0.895480  0.889984  0.041838\n",
       "2019-02-04  0.890707  0.910189  0.893657  0.036290\n",
       "2019-02-05  0.918487  0.922908  0.921554  0.031046\n",
       "2019-02-06  0.926673  0.920572  0.917794  0.022219\n",
       "2019-02-07  0.910911  0.908372  0.906950  0.034247"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "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>Adj Close</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Date</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2019-02-01</th>\n",
       "      <td>105.739998</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2019-02-04</th>\n",
       "      <td>107.220001</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2019-02-05</th>\n",
       "      <td>106.029999</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2019-02-06</th>\n",
       "      <td>105.269997</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2019-02-07</th>\n",
       "      <td>105.669998</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             Adj Close\n",
       "Date                  \n",
       "2019-02-01  105.739998\n",
       "2019-02-04  107.220001\n",
       "2019-02-05  106.029999\n",
       "2019-02-06  105.269997\n",
       "2019-02-07  105.669998"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " -----After process------ \n",
      "\n",
      "Shape of features :  (7244, 4)\n",
      "Shape of target :  (7244, 1)\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>Adj Close</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Date</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2018-09-24</th>\n",
       "      <td>113.957695</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-09-25</th>\n",
       "      <td>113.489723</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-09-26</th>\n",
       "      <td>113.917870</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-09-27</th>\n",
       "      <td>113.878044</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-09-28</th>\n",
       "      <td>115.112709</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             Adj Close\n",
       "Date                  \n",
       "2018-09-24  113.957695\n",
       "2018-09-25  113.489723\n",
       "2018-09-26  113.917870\n",
       "2018-09-27  113.878044\n",
       "2018-09-28  115.112709"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "display(feature_minmax_transform.head())\n",
    "print('Shape of features : ', feature_minmax_transform.shape)\n",
    "print('Shape of target : ', target_adj_close.shape)\n",
    "\n",
    "# Shift target array because we want to predict the n + 1 day value\n",
    "\n",
    "\n",
    "target_adj_close = target_adj_close.shift(-1)\n",
    "validation_y = target_adj_close[-90:-1]\n",
    "target_adj_close = target_adj_close[:-90]\n",
    "\n",
    "# Taking last 90 rows of data to be validation set\n",
    "validation_X = feature_minmax_transform[-90:-1]\n",
    "feature_minmax_transform = feature_minmax_transform[:-90]\n",
    "display(validation_X.tail())\n",
    "display(validation_y.tail())\n",
    "\n",
    "print(\"\\n -----After process------ \\n\")\n",
    "print('Shape of features : ', feature_minmax_transform.shape)\n",
    "print('Shape of target : ', target_adj_close.shape)\n",
    "display(target_adj_close.tail())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Train test Split using Timeseriessplit"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "ts_split= TimeSeriesSplit(n_splits=10)\n",
    "for train_index, test_index in ts_split.split(feature_minmax_transform):\n",
    "        X_train, X_test = feature_minmax_transform[:len(train_index)], feature_minmax_transform[len(train_index): (len(train_index)+len(test_index))]\n",
    "        y_train, y_test = target_adj_close[:len(train_index)].values.ravel(), target_adj_close[len(train_index): (len(train_index)+len(test_index))].values.ravel()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(6586, 4)"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\n",
    "X_train.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(658, 4)"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_test.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(6586,)"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_train.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(658,)"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_test.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "def validate_result(model, model_name):\n",
    "    predicted = model.predict(validation_X)\n",
    "    RSME_score = np.sqrt(mean_squared_error(validation_y, predicted))\n",
    "    print('RMSE: ', RSME_score)\n",
    "    \n",
    "    R2_score = r2_score(validation_y, predicted)\n",
    "    print('R2 score: ', R2_score)\n",
    "\n",
    "    plt.plot(validation_y.index, predicted,'r', label='Predict')\n",
    "    plt.plot(validation_y.index, validation_y,'b', label='Actual')\n",
    "    plt.ylabel('Price')\n",
    "    plt.gca().xaxis.set_major_formatter(mdates.DateFormatter('%Y-%m-%d'))\n",
    "    plt.gca().xaxis.set_major_locator(mdates.MonthLocator())\n",
    "    plt.title(model_name + ' Predict vs Actual')\n",
    "    plt.legend(loc='upper right')\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Benchmark Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RMSE:  64.22105109555618\n",
      "R2 score:  -304.3934023786507\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYgAAAEICAYAAABF82P+AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvhp/UCwAAIABJREFUeJzt3XeYVOX1wPHvgaU3BRYVkKIioCAIq4CiKEjEhmLvQIjYjTEW1CRqbBj9EU0UFVTAKNgRjKIixQoIooYuoAgrbelVyu75/XHuuLPL3d3ZMjuz7Pk8zzwzc+t779y5577lvldUFeeccy63ColOgHPOueTkAcI551woDxDOOedCeYBwzjkXygOEc865UB4gnHPOhfIAUUaJyAQR6RvDdNtE5LDSSJOLjYj8VUSeS3Q6CktEXhGR+4PPp4jIvAQnKSmJyBci0i/R6SgJHiDiSESWichOEdkqIptE5CsRuU5Eir3fVfUMVR0Vw3Q1VfXH4q4vWhB0Iq+sYBsj368oyXXlsf70qHWuFpEXRaRGvNdbUlT1QVW9rqSXKyJ/EJHMYL9sEZFvReTMkl4PgKpOVdWjY0zT1HikIY/1nSYiKiK3FWKe3wKfy8kDRPydo6q1gKbAYOAu4MXEJql4gqBTU1VrAsuxbYwMezX39CKSEodknBGsvwPQCbgzDutARCrGY7lx9HmwXw4EXgbeFJE6uSeK02+SDPoCG4J3V0weIEqJqm5W1fHAJUBfEWkDICJVROQJEVkuImtE5DkRqRaZT0TOFZHvgivCpSLSKxg+VUT+EHw+QkQ+FZHNIrJORF6Pml9F5Ijgcx0ReVlEMkTkZxH5SyQ3IyL9gqzxEyKyUUR+EpEzirKtIvKQiLwuImNEZCtwpYhUEJF7gm1YJyKviciBUfOcKCLTg5zWdyJycoz7dSXwMdA+allVRWSIiKwI9ulQEakaNf7uIOfxi4hcE+yjZsG4V0TkGRH5UES2AyfltzwRaSAiHwTp3iAin0Wt5x4RWRn8dgtF5JSo/TMyarrzRGResIzJItIyaly6iNwmInOC33eMiFSJYb9kAi8B1YHmwZX1siBNq4HhwfJ7i8j3wbq/iByXwbiOwW+xVUTGAFWixp0mIsuivjcVkXeDY2udiDwlIm2Bp4N9uE1E1uVOp4hcKSLTcw27Q0TeCT6fLSILgjSki8if8tpmEakJnA9cDxwlIu1zjT85OMY2B7/lVSJyA/afvCdI41gRSYk+JoJ5o4vX6gW/eUbwX3lPRBrl93uUVR4gSpmqfg2kAycFgx4DjsROcEcAjYC/AYjI8dhV4B3AAcDJwLKQxT6InSQPBBoD/85j9f8G6gCHAd2Aq4H+UeM7AYuA+sA/gBdFRAq/lQD0AUYH63sduA04K9iGxsB24F8AInIoMB64D6gLDALeEZF6Ba0kmLcXsCRq8BNAc+AYoAXQDLg3mP5s4GbgVGy/dw9Z7OXAA0AtYFp+y8N+mx+BVOBg4K/Beo4GrgU6qGpt4Awst5U7/a2BV4I0pQKfAO+JSKWoyS4GemK/W0fgqhj2SwowANgKLA0GNwZqAk2AG0TkOCxQ/AGohwWUcSJSOQhC44JhdYPP5+Wzrvex36AZcCjwhqrOAW4iyNWoav2Q2d8F2kjOerLLsWMHYAQwIMiFHwN8ms9mXwRsBN7C9uPVUWlsHqRxSLCtxwJzVHUodnw+EqSxTz7Lj6iA7bcmWMnAHuCpGOYre1TVX3F6YSfz00KGT8dOMIKdKA+PGtcF+Cn4/DzwzzyWPRX4Q/D5ZWAY0DhkOsUCT0VgF3BU1LhrganB537Akqhx1YN5Dy7sNgIPAZNzDVsMdIv6fmiQngrBvhiRa/pJwBV5rDMd2Iad/BQLjnWCcRWAX4GmUdOfBCyO2lcPRo1rFSyjWfD9FeClqPEFLe8R4J3o3zAY3hJYA/QAUkL2z8jg8wPA6FzrWw10jdrWS6PGDwGezmO//AHYC2wC1gFfAd2DcacF21E5avrhwH25lrEUOBELnCsAiRr3NXB/1PKWRe2P1UDFPNI0tYBj6DXgnqjfYzNQNfi+MlhGrRj+b1OBJ4LPVwX7PyX4/lfgzTzmeyWyXcH3lOhjImyaXPOnARlR378A+hWU3rLw8hxEYjTCyklTsRPxN0EWfxPwYTAc7CS6NHwROdyJBZuvg6KK34dMUx+oDPwcNeznIC0RqyMfVHVH8LFmDOsPsyLX9ybYlXFkO+dgf8IG2FXYZZFxwfjOQMN8ln+22lVlD+Bo7CoX7Cq+CvB91LL+G6yHYJnRacudztzDClreYGw/ThIrPrsDQFUXAX8G/g6sDYqGDg5ZV0OifhNVzcKCQujvAuwg/9/kC1U9QFXrq+oJqjo5atwaVd0d9b0pcFeu/X5IsO6GQLoGZ7xA9LET7VAsWGTmk678jAYuCz5fAbyjqr8G3/sAvYHlYsWqncIWEBQHnQxE6sDGYvupV1QaY/kvFUhEaojIC2LFwluAydj/a7/jAaKUBdn6RthVxjpgJ3B08Kc+QFXrqFUygp2oDi9omaq6WlWvUdWGWK5gqAT1DlHWYVnhplHDmgC/FG+L8k5Wru/pQM+o7TxAVauq6mpsO0fkGldDVR8vcCV2AnwFiEy7BtgNtMy1TyMVtauwopaIQwtIe77LU9UtqvonVW2GFcHcJSLdgnGvqOqJWPFUReDRkHWtJOo3EasTakx8fpfcv8kK4IFc+726qr7BvvsJ7HgJswJoKuEV+rF0F/0h0Cios7iM7OIlVHWGqvbGAvJ/sdxGmKuxi6QJQR3LEuyCKFLMlN9/KUcaVXUvlrutHjU4Orjfif2mx6sVH4YVU+4XPECUEhGpHZR/vwa8oqpzgqvF4cA/RaRBMF0jETk9mO1FoL+I9BCr5G0kIq1Cln2RiET+zBuxAz7H1VxwdfcG8LCI1BKRpli9wCtx2NwwzwGPiEiTIM0NRKR3MO4/QB8R6SkiFcUqhU8VkfxyENH+CZwpIm2C7XwBeFJEUsU0FpHfBdO+AQwQkZYiUp2gziAvBS1PRM4RkcODuprN2H7PFJHWwTZUwS4CdpLrN4lKT2+x+woqYXUaW4EZMW57cQwDbhSR44LtqhlsTw3sAqaCiNwUVNpehLUYCzMNWI/9vtVFpJqInBiMWwM0zlWnkkOQq3kbKz6riV2REyznchGprap7sP2SVy7laqzurn3U6xJs3x6IHee9ROSCYHvqi0i7qDTmvlfoe+CK4Hg8C+gaNa4WlpPbGNST/S2vbSvrPEDE33tiLXlWYGXtQ8hZMXwXdrUzPciufoKVX6NWod0fOwFuxiroonMAEccBM0RkG1bZ+0dV/SlkupuxOo8fsRPAaKwSsjQMwa4UJwX74yss3ajqMqwo4a9ABlaZ+2diPD6DXMirZJ/s/4wVh3yN7bePscplVPU94FngM6xe5Mtgnl35rCLP5WG/1WSsTuRL4ClV/QIrlvoHlnNbjTUg+EtI2udhTTKfDba9F9A7OCHGlarOwFr8PItdWPwAXBmM24X9JtcE487HKpTDlrMXOBtojR3ny4ELg9ETsf28Jriyz8torF7j9VxFVX2Bn4P/xgBCKuhFpCtWJPZMkJteHRwTY7E6skuC/8M52P9tAzAbaBss4gWgnViLpLeCYbcE278Jq/weH7XKIVjji/XYcTwhn+0q0yRnEaNz5UtQrDEbqBLk6JxzAc9BuHJHRPoETTnrYZXM4zw4OLcvDxCuPLoRK/pZjDX9vDGxyXEuOXkRk3POuVCeg3DOOReqTHfYVb9+fW3WrFmik+Gcc2XKN998s05VUwuarkwHiGbNmjFr1qxEJ8M558oUEcnrrvgcvIjJOedcKA8QzjnnQnmAcM45F6pM10E45xzAnj17SE9P59dffy144nKkatWqNG7cmEqV8uwKK18eIJxzZV56ejq1atWiWbNmSJGfcbV/UVXWr19Peno6zZs3L9IyvIjJOVfm/frrr9SrV8+DQxQRoV69esXKVXmAcM7tFzw47Ku4+6RcFjEtXAjjx8OePbB3L1xxBRyR+/E6zjlXzpXLHMScOXDXXfCXv8D998OgQYlOkXOurKtYsSLt27enTZs2XHTRRezYsaPgmfIwdepUzj77bADGjx/P4MGD85x206ZNDB06tMjryk+5DBDnnQfbtsHu3XDrrfDee7B+faJT5Zwry6pVq8Z3333H3LlzqVy5Ms8991yO8apKVlbhe5Xv3bs3g/K5ivUAUcIqVYIaNey9b18LFK/l9aRb55wrpJNOOoklS5awbNkyWrduzQ033ECHDh1YsWIFH3/8MV26dKFDhw5cdNFFbNu2DYAPP/yQVq1a0bVrV955553fljVy5EhuuukmANasWUOfPn1o164d7dq146uvvmLQoEEsXbqU9u3bc8cdd5TodpTLOoho7dvDMcfAqFFwoz8VwLmy79Zb4bvvSnaZ7dvDk0/GNOnevXuZMGECvXr1AmDRokWMGDGCoUOHsm7dOh566CE++eQTatSowWOPPcaQIUO48847ueaaa5g8eTJHHHEEl1xySeiyb7nlFrp168bYsWPJzMxk27ZtDB48mLlz5/JdSW8z5TQHkVvfvjBzJixYkOiUOOfKqp07d9K+fXvS0tJo0qQJAwYMAKBp06Z07twZgOnTpzN//nxOPPFE2rdvz6hRo/j5559ZuHAhzZs3p0WLFogIV155Zeg6Jk+ezPXXXw9YnUedOnXiuk3lPgcB1orpzjstF5FPXZBzriyI8Uq/pEXqIHKrUaPGb59VlZ49ezJmzJgc03z33XdJ2UzXcxDAQQfBmWfCM8/Am28mOjUuWb33Hnz4YaJT4cqyzp078+WXX7JkyRIAduzYwQ8//ECrVq346aefWLp0KcA+ASSiR48ePPvsswBkZmayZcsWatWqxdatW+OSXg8QgWeegaOPhosvhhtusHsk8rJ8Odx8M7z+eumlzyXW9Olw/vlw7rkwe3aiU+PKqtTUVEaOHMlll13GMcccQ+fOnVm4cCFVq1Zl2LBhnHXWWXTt2pWmTZuGzv/UU08xZcoU2rZtS8eOHZk3bx716tXjxBNPpE2bNiVeSV2mn0mdlpamJfnAoD174O674f/+z+6RePDBnOMzMuDRRy2Y7N4NXbvC55+X2OrLtKlTrS7nmGPg5JMtyEblrMu0TZvg2GPt8549UK2aBYlatRKbLpdtwYIFtG7dOtHJSEph+0ZEvlHVtILm9TqIKJUqwRNPwLp18MgjcNZZ0LkzbN0K//ynjdu+Hfr3h7VrYcaMRKc4eYwbB6tX28nzv/+Fn3+Gp59OdKoKr3dvO/GPGgUpKZCVBQMHQnq6XQzs2QOnnALXXQevvAJJWGzsXImJWxGTiLwkImtFZG7UsItEZJ6IZIlIWq7p7xaRJSKySEROj1e6YvHUU9C4MVx1ldV3HX443Hcf/O53MHcuvPACnHiiBYktW4q2jl274J137KS6P5g5E9LSrBuTq66Cl1+2mxHLkm3b4P33YfRo+P3vLedw/vlWL/Xww3axcNJJ8MADNs2IEYlOsXNxpqpxeQEnAx2AuVHDWgMtgalAWtTwo4DvgSpAc2ApULGgdXTs2FHjZcoUVRFVUO3eXXXGjJzj337bxs2aVbTlDxhg84uonnyy6jffFDvJCbNnj2q1aqq33mrfv/zStm348MSmq7CmTLF0n366vdeurVqxoupTT6lmZWVPt3evHRPVqqnOnZuw5Loo8+fPT3QSklbYvgFmaQzn8bjlIFT1M2BDrmELVHVRyOTnAq+p6i5V/QlYAhwfr7TF4pRT4O234ZNP7HV8rtS0aGHvixcXftkvvAAvvmg35t13H8yfb/f2lFXz5sHOnXDccfa9Sxdo2xZy9TSQ9KZNs/fRo+Gee+CAA2DSJLjllpxFSRUrWvFSrVpwySVQjC53nEtqydKKqRGwIup7ejBsHyIyUERmicisjIyMuCaqTx/o0SO8nPnww+29sAFi5kwLDD17WlHWfffZPRiff26Boiz6+mt7jwRRESuj/+YbKME2BHE3bRq0bAl161qR0rJl0K1b+LSHHAL/+Y8Fx3gEd1Wr+H/llZJftnOxSpYAEVbVF9q8SlWHqWqaqqalpqbGOVl5q17d6inCAsQnn8CECfsOX7cOLrwQDj7YrlIrVrTh/fpZBfmwYTmnz8qynmYfeii5r1JnzoQDD8wOmmA3H1avDkGT7aSnak1ZgxtegYIroH/3O2v1Nnx4dl9eY8daU9iidv6YmQkvvWQ5sFNPtfqcoMm8c6UuWQJEOnBo1PfGwMoEpSVmLVqEB4hbboHLL89ZSZuZaSfN1aut6Kp+/exxqalWGfryy1ZUA3bC+uMf4bHH4K9/hVat7CSUjK2SZ8604qXoE2qdOlb88uabViGf7H780Zoxd+lSuPn+/nc44QRr6XTttfY7jh9vJ/nCULWLivbtYcAAu2CINLOeMqVwy3KJM3bsWESEhQsX5jvdyJEjWbmy6Ke46O7A4ylZAsR44FIRqSIizYEWwNcJTlOBwgLE2rXWp9OmTTlbudx/P3z8sd1DkRbS+vjaa2HjRnjrLQsmDz5ozURvuw0+/dQCymWXWSuab76J62YVyo4d9nyNSP1DtIsvtibCEycWbdlZWdZ8tjSeQz99ur0XNkCkpMCYMfY+bBjcdJMFjOeft/QXZMUKK8468ki7m3/nTjsGZs+Ge++1oqzJkwu/PS4xxowZQ9euXXmtgO6hixsgSk0sNdlFeQFjgFXAHiyHMADoE3zeBawBPoqa/l6s9dIi4IxY1hHPVkyxeOIJa+2yfn32sEjrpgYNVJs3txYv48fbsAED8l5WVpZqixbWcqZqVZv+qqtUMzNt/N69qi+8YMsVUe3fX3XVqn2Xs2ZNzhY38RZpsTRu3L7jdu1SPeAA1b59i7bs22+3ZQ8bVqwkxuTGG1Vr1rT9XBTffKP68cf2+ZVXLN0TJ4ZPu2OH6ujRqj17ZreU69ZNdeRI22fRrrjCfvPS/E3LomRoxbR161Zt2LChLlq0SFu2bPnb8Mcee0zbtGmjxxxzjN5111365ptvao0aNfTII4/Udu3a6Y4dO7Rp06aakZGhqqozZ87Ubt26qarqjBkztEuXLtq+fXvt0qWLLly4UFVVp0yZomeddVZM6SpOK6a43SinqpflMWpsHtM/DDwcr/TEQ3RLpk6d7PNnn9nNYv/6F1x6qd1c9+ij0LFj/jeOidh0w4dblx9paVZfUSHI41WsaEUPF15odRJPPWXFNyNG2DCAzZuhaVMrmiqtTgdnzrT3sBxE5cpWHj9unN15Xrly7Mt95hnbd2A5qGuuKX5a8zNtmlWyR+qFCqtDh+zPF1xgv8Hzz8Npp2UP37UL7rjDihIjv9Vf/2p3oB92WPhyu3eHV1+1BgxHH120tJU3+fX2vXdv9t3whRFLb9/vvvsuvXr14sgjj6Ru3brMnj2bNWvW8O677zJjxgyqV6/Ohg0bqFu3Lk8//TRPPPEEaWHFCVFatWrFZ599RkpKCp988gn33HMPb7/9duESXwx+J3Ux5BUgunSxk/bhh1slc926VmxQtWr+y7vgAnvlp04dePxxK/M++2z7HAkQCxdacczjj1sLrEiaimvPHisKWbcu+5WRYe/vvWeV9YccEj7vhRfaXcmTJ0PQPX6BJk60epxzzrGy+Hh3Z5KeDt9/X3KPnq1a1U76//qX1TkdfLANHz0a/v1vu3C45hprSl2hgELe7t3tffJkDxDFpWr/D1ULFClFPPtt3WoBJvf8Y8aM4dagSdull17KmDFjyMrKon///lSvXh2AunXrxpzWNWtg69bN3HFHXxYvXoyIsCe/TuLiwANEMRx2mP3BI/UQmzfblct999mV6KBBcP31dmJo1qxk192ihV2dvvqqHUwi8MMPNq5mTbsTePZsqFKleOv59VcrU//2233HpaRY3ci11+Y9f8+edr/A22/HHiAeeQSaNLGy/REj7I7zn3+2K+6SkJFhTUinTLH3yHNA8mrSWhQDB8KQIZb+u++2Ye+8Y9swenTsXXQ0awbNm1uAuPnmkkvf/iyvK/1ffoFVq+y4rVbNmjQXZO9e+60iOcu1a62zznr17HeJWL9+PZMnT2bu3LmICJmZmYgIF1xwQUzdeKekpJCVlUVWFqxa9Svbt9tF2UMP/ZVTTz2VsWPHsmzZMk455ZSCE12CPEAUQ5UqdiKLBIgvv7ST9ckn2/c//AEuusiu+uOhZUsLSmvXWpflP/xgAeuVV6xPoYce2rfDwbxs2GAn8kqVcg7/+98tOAwebFew9evbKzUVatcu+ERXpYrlBMaOtWKzXbuyX7t323utWlYkU6uWtSaaOtXSXqOGVcqD5SKKEyC2bbOGAh99ZN2lQPby+/WzYBtdTFRcLVtaM9Vhw+Cuu6wPr4kT7YKhsP03de9uATYzs+hFYOXd7t2Wm6tb15pfp6dbA4vgwv43qtZQYPNme23bZvu8YUMLLMuX239swwbLOUf+L2+99RZXX301zz///G/L6tatG3Xr1uWFF17isssup0aN7CKm3F10N2vWjE8//YbWrc9gzJi3qVABDj0UNm/eTJUqdkvYyJEj472b9uEBopiiWzJ99pkdMNFFO/F84FPkCmjhwuwA0by5nZAvuMDuZH7ggYKLMTIzrYy1QQM7iR14oA2fPRv+8Q/Ljdx1V9HTeeWVdtV83XX7jqtSxYLEunV25TdypJ1A+/a18W3a2D787DNbTlE9/LBd0ffoYa3BTj3VAlbugFiSrr3WipM+/tj67Nq1y4r+Cqt7d7vz/vvvSzaIlReqdjUO0KiRnfBXrrQLq2bN7PjfujU7KOzebdNWr25Fpzt2ZM9fs6aduBcssGM2UrQ6ZswYBuUqo7zgggv43/8WkJbWmw4d0qhWrTJnnnkmjzzyCP369eO6666jWrVqTJo0jf797+P22wdQv/4jnHBCJ1autP/0rbfeyS239GXEiCH87nfdS2eHRYulJjtZX4luxaSqesMNqrVqqc6Zo9q5s+oJJ5Teun/6yVrAPP+8fW/fXvWMM+zzq6/auNx9SIWZNs2mBdW0NNWMDNWvvlJt21b1kENUN24sflpXrVJdudJafG3dqrp7d3bLnGuvtT6P5sxRPfRQ6wsp2llnqbZqVbx1V6+uevnlRV9GUezapZqaqnreeaqXXmqfi9JKauVK+20GDy75NO4v8mrFtHev6pIlqjNnqv7yS/bwZcusH7VFi+x95kxribZ4seratfu2Jtu0yebZs8e+L1yo+v33Bbcu++UXW/acOftOm5Vly5w5U/Xbb60FYqTVYkRmpur8+ZbGdeuy51u/XnX79gJ2SiAp+2IqL046ya4+2ra1tvSRIpHS0KSJVYguWmSn98WLrT092F2+IuF3dOc2YYLlMl5+2a5SDzrI6h0WLLBWVQccUPy0HnywXW3VrWtXYZUqZRe1PPSQFS+ddZZdqf3+9znnPekkyyWtXVu0dQ8ebFfv999frE0otMqVbVvee89e555btCKiQw6x52x89FHJp3F/pWq5tkWL7P6i3A0pDjrI3nftspzzkUdaLvqII6z4NHeLuzp1rIgzUjHdoIHlNDZtyj8dW7bYcf7rr5Y7ibZ1q9WHpaZaTrlBg31z+xUqWClFzZrw00/2mjPHimLXrSv8fiksDxDFdOmldlIbPtxOBsFzyktFhQp2YC9aZJVv27dnB4j69a3ZZiwB4oMPrIuJq66yR2pGnpa3erWdtOOtfn2r61i+3ALIuefmHB+p0/nii8Ive8UK6+6jX7/sVmel6ZprrAhj+/aiFS9F9Opl2x+nJ0vuV7ZssT6yfvjBTuJHHGEXKNF1P1Wr2kOg2ra1IqPatQsuio12wAEWRFatsorsMJHfvUEDm3bNmpzjN2+2NDVunH+LqpQUO3br1bMuXCpXtm069NC85ykpHiBKQOPGViH94oulfxJq2dICRKQFUyRAAJxxhnWkl1+/QGvWWId6Z5xh37t3t7qAiy+2A7K0XH+95RRuuWXfllcdO1qrk6J0ORGppP/b34qfxqI4/HDLzdWubfUfRdWrlzU39m438qaqbNuW3XdVs2aW88orB1yYgJBb5MS+c2d28/Lctm+3nEzt2hYktm61YRGbNlnOOZZcZYUK2dvTqpVtUyyNHbSYffN4gCjjWra07OacOfY9d4BQtUrSvESKLc48M35pjEVKilVE33ffvuMqV7Z7Pp57LrYcUcSSJdYn0sCBVhyXKCNGWMus4jQ5PvFEa3VVnoqZXn/dbjaM5RxXtWpVVq1az+LFSqVK9r+oX794QaAgdeva/23vXiuOzV2EFCleqlkzOy2RB4T9+qsVbxWm+FakcDebqirr16+nakE3YOXDWzGVcS1bWp8/EyZYtrlx4+xxHTtaLmDCBGu5E2bCBCuPbd++dNJbVC+8YHUsF15o9wTEchPgAw/YH+ree+Ofvvw0bGiv4qhc2XJ3EyZk3/cSoWp3mz//vOUWzz7birNKowgiXhYvhquvtiKiDz6w1m2R1nVhKlVqzPvvp9OkSQYNG5ZuD7gVK1pdwrRpdsKPtFyMBINI7n7XLtuubdssQGzcaBcNRe35NxZVq1alcfRJobBiqclO1lcytGJKtK+/thYuVaqotmmz7/jLL7fWMx99ZP0mffedtdRYtcpaZhx4oGq/fqWf7qJYtUr1sMNU69a1lh35mTfP+jm6887SSVtpeOYZ+61/+MG+b9ig+uST1sILrN+r1q2zj4fvv09seosqK8ta49Wqpfr3v6umpFi/ZgsWhE+/ebNqu3aqNWoU/QmPxbV9u7VUA9VLLrFjtWJF1XvvzZnOevWsD65TTrFWgolCjK2YEn6SL87LA4QddJEmquefv+/4N97IHp/X6/XXSz/dRbVkiXVed+ihqitW5By3bp3qgw/aSeWkk+wEE2kauD9YutR+r5tuss4aq1Wz7506qY4YYZ0AqlrwrFZNdeDAhCa3yMaOte0aMsS+T5tmv3m9evs22961S7VHDwsiH35Y+mmNlpWl+thjdmHSoIFtwyef5JxmyJDs/93ddycmnaoeIMqVQw6xX3LQoH3HZWXZ1fQXX1gu4u23VUeNUh3lTpgsAAAa0klEQVQ6VPUf/7AeaXO3+U52s2fbyf+oo3L2pNunT87A9/DDiUtjvLRoYdtWo4YFgNmzw6fr39+m2by5dNNXHEuWWC6pUSPVo4+2e2UiFi+23GP16qrvv2/DMjMthwx2TCeLDz5QrVPHemWOBO2InTtVmzSxNH/5ZWLSp+oBolw55RT7JV96KdEpKT2TJqlWrmw3Jm7frvruu/rbzWR79qhu2ZLoFMbHlCl2Y+SmTflPN2OG7Y9nny2VZBXJtm2q771nOaIjjsgO7IcdFn6D56pVqsceq1qhguq//pXdHfyjj5Z+2guybJnq9Onh4957T7V376J3LV8SYg0QYtOWTWlpaTqrLD30OE6uu84qKL/4wlq7lBdvvWXNcc84A/73P2tVMmtWfLvPKCtUrVsOVetLq7D9P8XD2rX2VMQffrD7FL76yiqhq1e3rk9OP92a8x5xRN7p3bbNnsw4frx9v/FG6yE3GbavLBGRb1Q1/77G8VZM+4UOHayVS6tWiU5J6brwQntuxA032Anirbc8OESIWF9Q119v98KUVNfvYCf3uXPhqKOsFV2sTS8feACGDrX7Alq2tHteevWCrl1jbwJcs6b1ivvAA9YK6MknPTjEk+cg9gN799odw9HdD5cnzz9vd63ecEOiU5Jctm61O4j79bNAWhQ//WQd251wgp2IV660JtEZGTY+cpdvmzbW22+bNnbne2rqvsvq1MmaYk+d6if1REt4DkJEXgLOBtaqaptgWF3gdaAZsAy4WFU3inWY/hRwJrAD6Keqs+OVtv1NSkr5DQ6Q//MoyrNatezeifxulAyzdq09rXD0aCsGAnvW9v/9nxXvbN9u92Ns3Jidm5g923Jwqnb/xQ8/5HxA1t69VgwYye25siGeRUwjgaeBl6OGDQImqepgERkUfL8LOANoEbw6Ac8G7865YjjtNPjvf2HZsoIfWrVjh/XHNW6c5cjatrXH4K5ebY+4/fBDuwFt1Kjwhz/t2GHzXn653dgW3b17pDsK7668bInbjeiq+hmwIdfgc4FRwedRwHlRw18OKtinAweISB4PsXTOxapnT3v/5JOCp331VSvfv/lmu9r/3//sqYhPPmm5hyVL7DkdV18dPn/16tZ55fHH23NEojuxmx2UB3iAKFtKuy+mg1R1FUDw3iAY3ghYETVdejDMOVcMrVtbNx8TJ+Y/nar1etu2rT1YqW3bnONvuw2WLrUOKfMjAvfcY3UXb7yRPXz2bAsg0X2FueSXLJ31hZVKhtaei8hAEZklIrMyIjVlzrlQIpaLmDTJ+uzKy9dfW3PY/B6JethhsfU8es451sLp0Uez1/ntt9CunT8ytawp7QCxJlJ0FLxHHgGTDkR3LdYYWBm2AFUdpqppqpqWGtZUwjmXw2mnWYdw336b9zTPPmtNSIvzWNeIChWsaGruXHj/fQsS337rxUtlUWkHiPFA8LRh+gLjooZfLaYzsDlSFOWcK57TTrP3vOoh1q+3G9iuvNJaPpWESy+1J7A98ojVXWzd6gGiLIpbgBCRMcA0oKWIpIvIAGAw0FNEFgM9g+8AHwA/AkuA4YC3aHeuhBx8sNUp5FUPMXKkdUV9/fUlt85KleDOO+0xvE8+acOOPbbklu9Kh98o51w58Kc/2QOXtmzJebd5Vpbd1XzQQUV7pGt+du60prVr19o6t20r3ANvXPzEeqNcslRSO+fiqHNnuw8h8uTBiEmTrAioJHMPEdWqWWACy8F4cCh7PEA4Vw5E+mKaMSPn8KFD7XGYF14Yn/Vef709Ca5Ll/gs38WXBwjnyoGmTaFBg5wBIj3dekUdMKB4z8vOT506dsPd4MEFT+uSj/fm6lw5IGK5iOgAMXy43SAX776sivNIZJdYnoNwrpzo1Mn6RNq0CfbssQDRq1f57ujR5c9zEM6VE5F6iJkzrTXTqlXWVbpzefEA4Vw5cdxxVtQ0YwZMmQJNmsCZZyY6VS6ZeYBwrpyoU8eeOjhmDMyfDw8/7H0jufx5HYRz5UinThYcKlWy1kvO5ccDhHPlSKQe4oIL7O5p5/LjAcK5cqRnT7sfInKHs3P58ToI58qRww+HNWsSnQpXVngOwjnnXCgPEM4550J5gHDOORfKA4RzzrlQHiCcc86F8gDhnHMuVEIChIj8UUTmisg8Ebk1GFZXRCaKyOLg/cBEpM0555wp9QAhIm2Aa4DjgXbA2SLSAhgETFLVFsCk4LtzzrkESUQOojUwXVV3qOpe4FOgD3AuMCqYZhRwXgLS5pxzLpCIADEXOFlE6olIdeBM4FDgIFVdBRC8NwibWUQGisgsEZmVkZFRaol2zrnyptQDhKouAB4DJgIfAt8Dewsx/zBVTVPVtNTU1Dil0jnnXEIqqVX1RVXtoKonAxuAxcAaETkEIHhfm4i0OeecM4lqxdQgeG8CnA+MAcYDfYNJ+gLjEpE255xzJlG9ub4tIvWAPcCNqrpRRAYDb4jIAGA5cFGC0uacc44EBQhVPSlk2HqgRwKS45xzLoTfSe2ccy6UBwjnnHOhPEA455wL5QHCOedcKA8QzjnnQnmAcM45F8oDhHPOuVAeIJxzzoXyAOGccy6UBwjnnHOhPEA455wL5QHCOedcKA8QzjnnQnmAcM45F8oDhHPOuVAeIJxzzoXyAOGccy5UoQKEiNSIV0Kcc84ll5gChIicICLzgQXB93YiMrSoKxWRP4nIPBGZKyJjRKSqiDQXkRkislhEXheRykVdvnPOueKLNQfxT+B0YD2Aqn4PnFyUFYpII+AWIE1V2wAVgUuBx4B/qmoLYCMwoCjLd845VzJiLmJS1RW5BmUWY70pQDURSQGqA6uA7sBbwfhRwHnFWL5zzrliijVArBCREwAVkcoicjtBcVNhqeovwBPAciwwbAa+ATap6t5gsnSgUdj8IjJQRGaJyKyMjIyiJME551wMYg0Q1wE3YiftdKB98L3QRORA4FygOdAQqAGcETKphs2vqsNUNU1V01JTU4uSBOecczFIiWUiVV0HXFFC6zwN+ElVMwBE5B3gBOAAEUkJchGNgZUltD7nnHNFEGsrplEickDU9wNF5KUirnM50FlEqouIAD2A+cAU4MJgmr7AuCIu3znnXAmItYjpGFXdFPmiqhuBY4uyQlWdgVVGzwbmBGkYBtwF3CYiS4B6wItFWb5zzrmSEVMRE1BBRA4MAgMiUrcQ8+5DVe8D7ss1+Efg+KIu0znnXMmK9ST/f8BXIhJphnoR8HB8kuSccy4ZxFpJ/bKIzMLuVRDgfFWdH9eUOeecS6h8A4SI1FbVLUGR0mpgdNS4uqq6Id4JdM45lxgF5SBGA2djN7JF35cgwffD4pQu55xzCZZvgFDVs4OmqN1UdXkppck551wSKLCZq6oqMLYU0uKccy6JxHofxHQROS6uKXHOOZdUYm3meipwnYgsA7YT1EGo6jHxSphzzrnEijVAhHWm55xzbj9WUDPXqlhPrkdg3WK8GNUlt3POuf1YQXUQo4A0LDicgd1R7ZxzrhwoqIjpKFVtCyAiLwJfxz9JzjnnkkFBOYg9kQ9etOScc+VLQTmIdiKyJfgs2HOkt5Ddiql2XFPnnHMuYQq6k7piaSXEOedccon1RjnnnHPljAcI55xzoTxAOOecC1XqAUJEWorId1GvLSJyq4jUFZGJIrI4eD+wtNPmnHMuW6kHCFVdpKrtVbU90BHYgfUWOwiYpKotgEnBd+eccwmS6CKmHsBSVf0ZOBe7c5vg/byEpco551zCA8SlwJjg80GqugogeG8QNoOIDBSRWSIyKyMjo5SS6Zxz5U/CAoSIVAZ6A28WZj5VHaaqaaqalpqaGp/EOeecS2gO4gxgtqquCb6vEZFDAIL3tQlLmXPOuYQGiMvILl4CGA/0DT73BcaVeoqcc879JiEBQkSqAz2Bd6IGDwZ6isjiYNzgRKTNOeecifWJciVKVXcA9XINW4+1anLOOZcEEt2KyTnnXJLyAOGccy6UBwjnnHOhPEA455wL5QHCOedcKA8QzjnnQnmAcM45F8oDhHPOuVAeIJxzzoXyAOGccy6UBwjnnHOhPEA455wL5QHCOedcKA8QzjnnQnmAcM45F8oDhHPOuVAeIJxzzoXyAOGccy5Uop5JfYCIvCUiC0VkgYh0EZG6IjJRRBYH7wcmIm3OOedMonIQTwEfqmoroB2wABgETFLVFsCk4LtzzrkEKfUAISK1gZOBFwFUdbeqbgLOBUYFk40CzivttDnnnMuWiBzEYUAGMEJEvhWRF0SkBnCQqq4CCN4bhM0sIgNFZJaIzMrIyCi9VDvnXDmTiACRAnQAnlXVY4HtFKI4SVWHqWqaqqalpqbGK43OOVfuJSJApAPpqjoj+P4WFjDWiMghAMH72gSkzTnnXKDUA4SqrgZWiEjLYFAPYD4wHugbDOsLjCvttDnnnMuWkqD13gy8KiKVgR+B/liwekNEBgDLgYsSlDbnnHMkKECo6ndAWsioHqWdFuecc+H8TmrnnHOhPEA455wL5QHCOedcKA8QzjnnQnmAcM45F8oDhHPOuVAeIJxzzoXyAOGccy6UBwjnnHOhPEA455wL5QHCOedcKA8QzjnnQnmAcM45F8oDhHPOuVAeIJxzzoXyAOGccy6UBwjnnHOhEvJEORFZBmwFMoG9qpomInWB14FmwDLgYlXdmIj0OeecS2wO4lRVba+qkUePDgImqWoLYFLw3TnnXIIkUxHTucCo4PMo4LwEpsU558q9RAUIBT4WkW9EZGAw7CBVXQUQvDcIm1FEBorILBGZlZGRUUrJdc658ichdRDAiaq6UkQaABNFZGGsM6rqMGAYQFpamsYrgc45V94lJAehqiuD97XAWOB4YI2IHAIQvK9NRNqcc86ZUg8QIlJDRGpFPgO/A+YC44G+wWR9gXGlnTbnnHPZElHEdBAwVkQi6x+tqh+KyEzgDREZACwHLkpA2pxzzgVKPUCo6o9Au5Dh64EepZ0e55xz4ZKpmatzzrkk4gHCOedcKA8QzjnnQnmAcM45F8oDhHPOuVAeIJxzzoXyAOGccy6UBwjnnHOhPEA455wL5QHCOedcKA8QzjnnQnmAcM45F8oDhHPOuVAeIJxzzoVK1CNHE+vDD+HWW0tueU2bwujRUK9ebNPv2QPPPw/Dh8OuXcVbd//+cNddxVtGmPR0uPxyWJtED/Y7/nh46CFo0iT/6TZvhiuugCVLiraem2+GG2+0z+vWwaWX2v4oyDnnwOOPF359qjBmjB0TTz4Jxx5rwydOhD//GXbvhqpV4emnoWvXwi+/pD3+OLz4YuzTi8Ddd8PVV+c/nSq8+y489hhs2lS8NCaT2rXhpZegTZucw1esgCuvhDVrirbca66x4yOORLXsPtY5LS1NZ82aVfgZp0+3P2JJiBzUXbta4KlUKf/pJ0yA226DhQuhc2cLLkX1008wcybMmgUdOhR9OWEuuQTGj4dzzy3Z5RbVnj3wwQf2+fbbLSjWrLnvdJmZdqKeOBH69IEKhcwk//ADzJ0Lc+ZAy5Zw7bX25z7/fDvR5WX1avj0U/t9e/WKfX3Tp9vFyowZ9v2442zYzp3QqhWkpECnTvDFF1C9uqWtcuXCbVNJ+vZb6NjR0tm8eWzzzJsHP/4IixZB48Z5L/e222DqVNvudvs8MqbsmjgRWreGzz/PPoa2b4eTTrKLmDPPLNpye/e2i7giEJFvVDWtwAlVtcy+OnbsqElh1ChVUL3hhrynmTtX9fTTbboWLVTHjVPNyireejduVE1NVT3hhOIvK9rkyZbOBx4ouWWWhJ9/Vr3sMkvbIYeovvSSamZmzmluv93GP/dc0daxerVq7dqqvXqpzpqlKqL6pz8VPN+vv9rveuSRqrt2FW5bDj5YdcSI7OPohRdU773XPn/xhU0/YYJ9f+yxom1XScjKUu3aVbV+fdUNG2Kf78cfVatWte3NbeVK1d//3vZzvXqqQ4eq7tlTcmlOBsOH22/36qv2PTNT9cILbZvffz8hSQJmaQzn2ISf5IvzSpoAoap6xx22O0VUK1RQbdtWdfRo1VWrVK+7zoYdcIDqkCGxnUBi9eKLtt5//1v1/vvtz1uhQtFe3bqpfvSRaps2qs2aqe7YUXLpLEnTpql27mzbfeyxqlOnqs6fr3r55TbsppuKt/whQ2w5jRqpNmigumlTbPN98IHN9/jjeU+zdaud/KtWtddf/mLDVO0EfMIJdqKsUkX1yitzztu7t2rNmqq//GKBbMoU1WeeUb3xRtXu3S3QtGxpAWbTJtUnn1Q99NDs37dxY9WHH1bNyCjSbtFXX7XtGz688PP+7W8276ef2vcdO1Qfeki1Rg3VSpVU//xnu+DZH2VmqqalqTZsqDppkl18FHScxFmsASJhRUwiUhGYBfyiqmeLSHPgNaAuMBu4SlV357eMIhcxxUNmJgwbBitXQlaWFTvNn29ZygoV4Prr4f77Y6+niFVWFnTpAl9/bd9794Zjjin8cnbtgldftfQDvPOOFdEkK1V47TUralqxwvZztWpWf/DggwUX9eVnzx4r4liwwIqX+vePfd5I8dbhh4ePX70aNmywooFHH923PmX2bEhLgxo1rEimYcPscUuXwtFH27bvjvpr1K4NRx1lxRhz5liRY4UKdmx062ZFGWDFWBMnQsWKtvyUFNtPud/zKkpbvtyKf2bMKHzR3Y4dlr4tW2yb1q61+p0+feAf/4Ajjijc8sqaGTOsSBmgbl2rk/nzn/MvtoyjWIuYEhkgbgPSgNpBgHgDeEdVXxOR54DvVfXZ/JaRVAEit6wsGDvWyo4HDrQ/R7zMnw/PPmuVVkUJDhG//gojR9of9957E3bwFsrOnTB0qJ2Arr8e6tcvmeV+/73Vwdx7b+FOhunp8Le/wdat4eOrVbMK8E6d8l7GSy9BgwZw9tn7jvvPf+DLL+14OuooezVsmP1bqVpd2AcfwMUXZweHiHnzLLBu22aBcO/enO979uSdripV4C9/sSBRFDNmwD//aRdTlSvb8XrKKUVbVln0zDP2Hxs4EGrVSmhSkjpAiEhjYBTwMHAbcA6QARysqntFpAtwv6qent9ykjpAOOdckoo1QCTqPogngTuBrOB7PWCTqu4NvqcDjcJmFJGBIjJLRGZlZGTEP6XOOVdOlXqAEJGzgbWq+k304JBJQ7M2qjpMVdNUNS01NTUuaXTOOZeYG+VOBHqLyJlAVaA2lqM4QERSglxEY2BlAtLmnHMuUOo5CFW9W1Ubq2oz4FJgsqpeAUwBLgwm6wuMK+20Oeecy5ZMfTHdBdwmIkuwOolC3MvvnHOupCW0LyZVnQpMDT7/CByfyPQ455zLlkw5COecc0nEA4RzzrlQZbo3VxHJAH4u4uz1gXUlmJz9le+ncL5fCub7KDaJ2E9NVbXA+wTKdIAoDhGZFcudhOWd76dwvl8K5vsoNsm8n7yIyTnnXCgPEM4550KV5wAxLNEJKCN8P4Xz/VIw30exSdr9VG7rIJxzzuWvPOcgnHPO5cMDhHPOuVBJEyBE5FARmSIiC0Rknoj8MRheV0Qmisji4P3AYHgrEZkmIrtE5PZcy/pTsIy5IjJGRKrmsc4PRWSTiPw31/DmIjIjWOfrIlI5j/k7isgcEVkiIv8Sscd6ichFwfqzRKTEmq8l2T66KdhuFZE8H+OW174UkZNFZLaI7BWRC/Oav4zum1dFZFEw/0siEvr809LaN1HrK8l99Mdg++aJyK35rLNXsC+WiMigqOFJdfxErS+Z9lHij6NYHlxdGi/gEKBD8LkW8ANwFPAPYFAwfBDwWPC5AXAc9lS626OW0wj4CagWfH8D6JfHOntgT7P7b67hbwCXBp+fA67PY/6vgS7Y8ywmAGcEw1sDLbF+ptL20310LNAMWAbUzyfNofsymPcY4GXgwv1s35wZHBMCjMnn+CmVfROHfdQGmAtUx/pz+wRoEbK+isBS4DCgMvA9cFQyHj9Juo8SfhwlTQ5CVVep6uzg81ZgAfZnPRd7PCnB+3nBNGtVdSYQ9hDdFKCaiKRgP1DosyVUdRKQ4+HBIiJAd+Ct3OvMNd0h2PO0p6n9Gi9HpW2Bqi6KcdNjliz7KBj+raouyy+9+e1LVV2mqv8j+6mCxZJk++YDDWAXEY1zT1Oa+yYqXSW1j1oD01V1h9rzWz4F+oSs8nhgiar+qKq7gdeCdSXd8RORZPso4cdR0gSIaCLSDLvCmAEcpKqrwH48LGLnSVV/AZ4AlgOrgM2q+nEhVh/r408bBeMoYLq4SPA+ilXMj5ItScmyb4IigauAD0NGJ2TfRKWtGUXcR9iV8ckiUk9EqmNXuoeGTNcIWBH1vbDb6PuIxB5HSRcgRKQm8DZwq6puKcL8B2IRuDnQEKghIlcWZhEhw8LaAsf8mNSSlgT7KOZVhQyL6z5Ksn0zFPhMVT8PW1XIsDJx/KjqAuAxYCJ20voe2BsyaXG30feRSdhxlFQBIoiUbwOvquo7weA1QXFOpFhnbQGLOQ34SVUzVHUP8A5wgoh0EpHvglfvfOZfR/D40+B7Y2CliFSMmv/vWKSOzvKVymNSk2Qf5Ze+j4L5XyCPfVmU5ca47qTZNyJyH5AK3BY1LGH7JioNJbGPUNUXVbWDqp4MbAAWBxW8kX10HfYfib5qLnAbfR/l3MZEH0cJfWBQtKAs7UVggaoOiRo1HnsE6WBiexTpcqBzkK3biVUkzlLVGUD7gtKhqioikcefvhZZp6pm5p5fRLaKSGcsC3o18O8CN7QYkmUf5UdVT8+V5n32ZXGWn5dk2jci8gfgdKCHqv5W/puofRO1vpLaR4hIA1VdKyJNgPOBLqq6kah9FJy0WohIc+AX7BHDl+e3XN9H2fsoKY4jLaHa/+K+gK5Y1uh/wHfB60ysjG0SsDh4rxtMfzAWfbcAm4LPtYNxDwALsXLA/wBV8ljn50AGdiJIB04Phh+GVQotAd7MZ/60YB1LgafJvjO9T7C8XcAa4KP9cB/dEnzfi12xvJDH/KH7Emv5kQ5sB9YD8/ajfbM3OCYi6fhbIvdNnPbR58B8rOikRz7rPBNrCbQUuDdqeFIdP0m6jxJ+HHlXG84550IlVR2Ec8655OEBwjnnXCgPEM4550J5gHDOORfKA4RzzrlQHiCcc86F8gDhnHMu1P8DWWmmg3BR+lUAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.tree import DecisionTreeRegressor\n",
    "\n",
    "dt = DecisionTreeRegressor(random_state=0)\n",
    "\n",
    "benchmark_dt=dt.fit(X_train, y_train)\n",
    "\n",
    "validate_result(benchmark_dt, 'Decision Tree Regression')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Process the data for LSTM"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train =np.array(X_train)\n",
    "X_test =np.array(X_test)\n",
    "\n",
    "X_tr_t = X_train.reshape(X_train.shape[0], 1, X_train.shape[1])\n",
    "X_tst_t = X_test.reshape(X_test.shape[0], 1, X_test.shape[1])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Model building : LSTM"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/200\n",
      "6586/6586 [==============================] - 9s 1ms/step - loss: 110.6465\n",
      "Epoch 2/200\n",
      "6586/6586 [==============================] - 5s 828us/step - loss: 21.8761\n",
      "Epoch 3/200\n",
      "6586/6586 [==============================] - 5s 797us/step - loss: 17.1062\n",
      "Epoch 4/200\n",
      "6586/6586 [==============================] - 12s 2ms/step - loss: 11.4689\n",
      "Epoch 5/200\n",
      "6586/6586 [==============================] - 25s 4ms/step - loss: 7.6143: 4s  -\n",
      "Epoch 6/200\n",
      "6586/6586 [==============================] - 10s 2ms/step - loss: 5.2283\n",
      "Epoch 7/200\n",
      "6586/6586 [==============================] - 5s 824us/step - loss: 3.7930\n",
      "Epoch 8/200\n",
      "6586/6586 [==============================] - 5s 811us/step - loss: 2.8919 0s - \n",
      "Epoch 9/200\n",
      "6586/6586 [==============================] - 5s 830us/step - loss: 2.2973\n",
      "Epoch 10/200\n",
      "6586/6586 [==============================] - 5s 795us/step - loss: 1.9021\n",
      "Epoch 11/200\n",
      "6586/6586 [==============================] - 5s 790us/step - loss: 1.6255\n",
      "Epoch 12/200\n",
      "6586/6586 [==============================] - 5s 829us/step - loss: 1.4193\n",
      "Epoch 13/200\n",
      "6586/6586 [==============================] - 5s 794us/step - loss: 1.2602\n",
      "Epoch 14/200\n",
      "6586/6586 [==============================] - 5s 802us/step - loss: 1.1448\n",
      "Epoch 15/200\n",
      "6586/6586 [==============================] - ETA: 0s - loss: 1.061 - 5s 808us/step - loss: 1.0628\n",
      "Epoch 16/200\n",
      "6586/6586 [==============================] - 5s 800us/step - loss: 1.0057\n",
      "Epoch 17/200\n",
      "6586/6586 [==============================] - 5s 812us/step - loss: 0.9648\n",
      "Epoch 18/200\n",
      "6586/6586 [==============================] - 5s 791us/step - loss: 0.9338\n",
      "Epoch 19/200\n",
      "6586/6586 [==============================] - 5s 813us/step - loss: 0.9087\n",
      "Epoch 20/200\n",
      "6586/6586 [==============================] - 5s 827us/step - loss: 0.8873\n",
      "Epoch 21/200\n",
      "6586/6586 [==============================] - 5s 782us/step - loss: 0.8709\n",
      "Epoch 22/200\n",
      "6586/6586 [==============================] - 5s 792us/step - loss: 0.8602\n",
      "Epoch 23/200\n",
      "6586/6586 [==============================] - 5s 786us/step - loss: 0.8541\n",
      "Epoch 24/200\n",
      "6586/6586 [==============================] - 5s 807us/step - loss: 0.8518\n",
      "Epoch 25/200\n",
      "6586/6586 [==============================] - ETA: 0s - loss: 0.849 - 5s 777us/step - loss: 0.8518\n",
      "Epoch 26/200\n",
      "6586/6586 [==============================] - 5s 763us/step - loss: 0.8539\n",
      "Epoch 27/200\n",
      "6586/6586 [==============================] - 5s 800us/step - loss: 0.8578\n",
      "Epoch 28/200\n",
      "6586/6586 [==============================] - 5s 823us/step - loss: 0.8629\n",
      "Epoch 29/200\n",
      "6586/6586 [==============================] - 5s 812us/step - loss: 0.8691\n",
      "Epoch 00029: early stopping\n"
     ]
    }
   ],
   "source": [
    "\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Dense\n",
    "import keras.backend as K\n",
    "from keras.callbacks import EarlyStopping\n",
    "from keras.optimizers import Adam\n",
    "from keras.models import load_model\n",
    "from keras.layers import LSTM\n",
    "K.clear_session()\n",
    "model_lstm = Sequential()\n",
    "model_lstm.add(LSTM(16, input_shape=(1, X_train.shape[1]), activation='relu', return_sequences=False))\n",
    "model_lstm.add(Dense(1))\n",
    "model_lstm.compile(loss='mean_squared_error', optimizer='adam')\n",
    "early_stop = EarlyStopping(monitor='loss', patience=5, verbose=1)\n",
    "history_model_lstm = model_lstm.fit(X_tr_t, y_train, epochs=200, batch_size=8, verbose=1, shuffle=False, callbacks=[early_stop])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Evaluation of Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The R2 score on the Train set is:\t0.887\n",
      "The R2 score on the Test set is:\t0.959\n"
     ]
    }
   ],
   "source": [
    "y_pred_test_lstm = model_lstm.predict(X_tst_t)\n",
    "y_train_pred_lstm = model_lstm.predict(X_tr_t)\n",
    "print(\"The R2 score on the Train set is:\\t{:0.3f}\".format(r2_score(y_train, y_train_pred_lstm)))\n",
    "r2_train = r2_score(y_train, y_train_pred_lstm)\n",
    "\n",
    "print(\"The R2 score on the Test set is:\\t{:0.3f}\".format(r2_score(y_test, y_pred_test_lstm)))\n",
    "r2_test = r2_score(y_test, y_pred_test_lstm)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Predictions made by LSTM"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "658/658 [==============================] - 2s 3ms/step\n"
     ]
    }
   ],
   "source": [
    "score_lstm= model_lstm.evaluate(X_tst_t, y_test, batch_size=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "LSTM: 14.690967\n"
     ]
    }
   ],
   "source": [
    "print('LSTM: %f'%score_lstm)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_pred_test_LSTM = model_lstm.predict(X_tst_t)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# LSTM's Prediction Visual"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(y_test, label='True')\n",
    "plt.plot(y_pred_test_LSTM, label='LSTM')\n",
    "plt.title(\"LSTM's_Prediction\")\n",
    "plt.xlabel('Observation')\n",
    "plt.ylabel('INR_Scaled')\n",
    "plt.legend()\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Converting Prediction data\n",
    "In this step I have made the prediction of test data and will convert the dataframe to csv so that we can see the price difference between actual and predicted price."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [],
   "source": [
    "col1 = pd.DataFrame(y_test, columns=['True'])\n",
    "\n",
    "col2 = pd.DataFrame(y_pred_test_LSTM, columns=['LSTM_prediction'])\n",
    "\n",
    "col3 = pd.DataFrame(history_model_lstm.history['loss'], columns=['Loss_LSTM'])\n",
    "results = pd.concat([col1, col2, col3], axis=1)\n",
    "results.to_excel('PredictionResults_LSTM_NonShift.xlsx')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Conclusion"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "It is impossible to  get a model that can 99% predict the price without any error, there are too many factors can affect the\n",
    "stock prices. So, we cannot hope there is a perfect model, but the general trend of predicted price is in line with the actual data, so the trader could have an indicator to reference, and makes trading decision by himself.\n",
    "\n",
    "\n",
    "Further, we can improve the model's accuracy by increasing the epochs, trying out different activation functions or even change the model's structure. As exact "
   ]
  },
  {
   "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": "3.6.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}