Notebook 2_ Stock Market Analysis using LSTM(Vanilla)-how good is a LSTM at predicting price movement.ipynb

{"nbformat":4,"nbformat_minor":0,"metadata":{"colab":{"name":"Notebook 2: Stock Market Analysis using LSTM(Vanilla)-how good is a LSTM at predicting price movement.ipynb","version":"0.3.2","provenance":[],"collapsed_sections":[]},"kernelspec":{"name":"python3","display_name":"Python 3"},"accelerator":"GPU"},"cells":[{"cell_type":"code","metadata":{"id":"2zUjhYQHfTmA","colab_type":"code","outputId":"14bb64aa-1d66-4a5b-d37a-7f3cdf550440","executionInfo":{"status":"ok","timestamp":1565631952227,"user_tz":240,"elapsed":1756,"user":{"displayName":"Nupur Deshpande","photoUrl":"","userId":"13787757531456301957"}},"colab":{"base_uri":"https://localhost:8080/","height":33}},"source":["from keras.layers.core import Dense, Activation, Dropout\n","from keras.layers.recurrent import LSTM\n","from keras.models import Sequential\n","import time #helper libraries"],"execution_count":0,"outputs":[{"output_type":"stream","text":["Using TensorFlow backend.\n"],"name":"stderr"}]},{"cell_type":"markdown","metadata":{"id":"UJ0jimOSaLx0","colab_type":"text"},"source":["**WHY USE LSTM? Is it good to predict re-occuring data?I didn't get great results from the 1st notebook and decided to try to predict stock price movements by giving in stock price movements instead of just simple stock prices as input**\n","\n","\n","**In this notebook I normalized only the closing prices and then broke it into train and test datasets. I just kept a simple lookback window to give all similar data same movement(movement is just giving the data points the same kind of normalization) and put them in price points as input**\n","\n"]},{"cell_type":"code","metadata":{"id":"Uwk3PeSUGDk3","colab_type":"code","colab":{}},"source":["# Code to read csv file into Colaboratory:\n","!pip install -U -q PyDrive\n","from pydrive.auth import GoogleAuth\n","from pydrive.drive import GoogleDrive\n","from google.colab import auth\n","from oauth2client.client import GoogleCredentials\n","# Authenticate and create the PyDrive client.\n","auth.authenticate_user()\n","gauth = GoogleAuth()\n","gauth.credentials = GoogleCredentials.get_application_default()\n","drive = GoogleDrive(gauth)"],"execution_count":0,"outputs":[]},{"cell_type":"code","metadata":{"id":"ctysdd5cJzV4","colab_type":"code","colab":{}},"source":["link = 'https://drive.google.com/open?id=1y394o-wiPqYPaWzVYTPRr8xpE_aM9621'"],"execution_count":0,"outputs":[]},{"cell_type":"code","metadata":{"id":"_RCOZpTdND1z","colab_type":"code","outputId":"bfbfde48-e1ef-4d96-9ce5-501bf8a86e8c","executionInfo":{"status":"ok","timestamp":1565631982249,"user_tz":240,"elapsed":262,"user":{"displayName":"Nupur Deshpande","photoUrl":"","userId":"13787757531456301957"}},"colab":{"base_uri":"https://localhost:8080/","height":35}},"source":["fluff, id = link.split('=')\n","print (id) # Verify that you have everything after '='"],"execution_count":0,"outputs":[{"output_type":"stream","text":["1y394o-wiPqYPaWzVYTPRr8xpE_aM9621\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"id":"hyxINN1tNEqW","colab_type":"code","outputId":"f7af8116-ef78-4e2f-fe41-46e886e800b8","executionInfo":{"status":"ok","timestamp":1565631983792,"user_tz":240,"elapsed":776,"user":{"displayName":"Nupur Deshpande","photoUrl":"","userId":"13787757531456301957"}},"colab":{"base_uri":"https://localhost:8080/","height":532}},"source":["import pandas as pd\n","downloaded = drive.CreateFile({'id':id}) \n","downloaded.GetContentFile('AMZN.csv')  \n","df3 = pd.read_csv('AMZN.csv')\n","# Dataset is now stored in a Pandas Dataframe"],"execution_count":0,"outputs":[{"output_type":"stream","text":["WARNING: Logging before flag parsing goes to stderr.\n","W0812 17:46:22.876577 140683407243136 __init__.py:44] file_cache is unavailable when using oauth2client >= 4.0.0 or google-auth\n","Traceback (most recent call last):\n","  File \"/usr/local/lib/python3.6/dist-packages/googleapiclient/discovery_cache/__init__.py\", line 36, in autodetect\n","    from google.appengine.api import memcache\n","ModuleNotFoundError: No module named 'google.appengine'\n","\n","During handling of the above exception, another exception occurred:\n","\n","Traceback (most recent call last):\n","  File \"/usr/local/lib/python3.6/dist-packages/googleapiclient/discovery_cache/file_cache.py\", line 33, in <module>\n","    from oauth2client.contrib.locked_file import LockedFile\n","ModuleNotFoundError: No module named 'oauth2client.contrib.locked_file'\n","\n","During handling of the above exception, another exception occurred:\n","\n","Traceback (most recent call last):\n","  File \"/usr/local/lib/python3.6/dist-packages/googleapiclient/discovery_cache/file_cache.py\", line 37, in <module>\n","    from oauth2client.locked_file import LockedFile\n","ModuleNotFoundError: No module named 'oauth2client.locked_file'\n","\n","During handling of the above exception, another exception occurred:\n","\n","Traceback (most recent call last):\n","  File \"/usr/local/lib/python3.6/dist-packages/googleapiclient/discovery_cache/__init__.py\", line 41, in autodetect\n","    from . import file_cache\n","  File \"/usr/local/lib/python3.6/dist-packages/googleapiclient/discovery_cache/file_cache.py\", line 41, in <module>\n","    'file_cache is unavailable when using oauth2client >= 4.0.0 or google-auth')\n","ImportError: file_cache is unavailable when using oauth2client >= 4.0.0 or google-auth\n"],"name":"stderr"}]},{"cell_type":"code","metadata":{"id":"TKqVdqkHNPEq","colab_type":"code","colab":{}},"source":["#VANILLA LSTM NETWORK CREATED:"],"execution_count":0,"outputs":[]},{"cell_type":"code","metadata":{"id":"biH6rwhyOAAm","colab_type":"code","colab":{}},"source":["from keras.layers.core import Dense, Activation, Dropout\n","from keras.layers.recurrent import LSTM\n","from keras.models import Sequential\n","import time #helper libraries"],"execution_count":0,"outputs":[]},{"cell_type":"code","metadata":{"id":"OtRzd9-IOB8L","colab_type":"code","colab":{}},"source":["df3 = df3.dropna(axis = \"columns\")"],"execution_count":0,"outputs":[]},{"cell_type":"code","metadata":{"id":"77Me2RVdPMy6","colab_type":"code","outputId":"23fb53a8-458a-4d08-fe65-692c429362a1","executionInfo":{"status":"ok","timestamp":1565631995984,"user_tz":240,"elapsed":486,"user":{"displayName":"Nupur Deshpande","photoUrl":"","userId":"13787757531456301957"}},"colab":{"base_uri":"https://localhost:8080/","height":194}},"source":["df3.head()"],"execution_count":0,"outputs":[{"output_type":"execute_result","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>Date</th>\n","      <th>Open</th>\n","      <th>High</th>\n","      <th>Low</th>\n","      <th>Adj Close</th>\n","      <th>Volume</th>\n","      <th>Close</th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","    <tr>\n","      <th>0</th>\n","      <td>1/4/2010</td>\n","      <td>136.250000</td>\n","      <td>136.610001</td>\n","      <td>133.139999</td>\n","      <td>133.899994</td>\n","      <td>7599900</td>\n","      <td>133.899994</td>\n","    </tr>\n","    <tr>\n","      <th>1</th>\n","      <td>1/5/2010</td>\n","      <td>133.429993</td>\n","      <td>135.479996</td>\n","      <td>131.809998</td>\n","      <td>134.690002</td>\n","      <td>8851900</td>\n","      <td>134.690002</td>\n","    </tr>\n","    <tr>\n","      <th>2</th>\n","      <td>1/6/2010</td>\n","      <td>134.600006</td>\n","      <td>134.729996</td>\n","      <td>131.649994</td>\n","      <td>132.250000</td>\n","      <td>7178800</td>\n","      <td>132.250000</td>\n","    </tr>\n","    <tr>\n","      <th>3</th>\n","      <td>1/7/2010</td>\n","      <td>132.009995</td>\n","      <td>132.320007</td>\n","      <td>128.800003</td>\n","      <td>130.000000</td>\n","      <td>11030200</td>\n","      <td>130.000000</td>\n","    </tr>\n","    <tr>\n","      <th>4</th>\n","      <td>1/8/2010</td>\n","      <td>130.559998</td>\n","      <td>133.679993</td>\n","      <td>129.029999</td>\n","      <td>133.520004</td>\n","      <td>9830500</td>\n","      <td>133.520004</td>\n","    </tr>\n","  </tbody>\n","</table>\n","</div>"],"text/plain":["       Date        Open        High  ...   Adj Close    Volume       Close\n","0  1/4/2010  136.250000  136.610001  ...  133.899994   7599900  133.899994\n","1  1/5/2010  133.429993  135.479996  ...  134.690002   8851900  134.690002\n","2  1/6/2010  134.600006  134.729996  ...  132.250000   7178800  132.250000\n","3  1/7/2010  132.009995  132.320007  ...  130.000000  11030200  130.000000\n","4  1/8/2010  130.559998  133.679993  ...  133.520004   9830500  133.520004\n","\n","[5 rows x 7 columns]"]},"metadata":{"tags":[]},"execution_count":9}]},{"cell_type":"code","metadata":{"id":"1LArM8rQPU6Y","colab_type":"code","colab":{}},"source":["#working with a conventional OHCLV dataset"],"execution_count":0,"outputs":[]},{"cell_type":"code","metadata":{"id":"AUI-7-jpPgdJ","colab_type":"code","colab":{}},"source":["import numpy as np"],"execution_count":0,"outputs":[]},{"cell_type":"code","metadata":{"id":"gqkI7vdpQB3d","colab_type":"code","colab":{}},"source":["#FOCUS ON PREDICTING CLOSING PRICES AS BEFORE"],"execution_count":0,"outputs":[]},{"cell_type":"code","metadata":{"id":"4_eLiOTqSt1p","colab_type":"code","colab":{}},"source":["df_close = df3['Close']"],"execution_count":0,"outputs":[]},{"cell_type":"code","metadata":{"id":"CpmyiqPMS6lv","colab_type":"code","outputId":"61308885-7d53-4fa0-e080-a41f2341b926","executionInfo":{"status":"ok","timestamp":1565632000476,"user_tz":240,"elapsed":254,"user":{"displayName":"Nupur Deshpande","photoUrl":"","userId":"13787757531456301957"}},"colab":{"base_uri":"https://localhost:8080/","height":35}},"source":["df_close.shape"],"execution_count":0,"outputs":[{"output_type":"execute_result","data":{"text/plain":["(2265,)"]},"metadata":{"tags":[]},"execution_count":14}]},{"cell_type":"code","metadata":{"id":"cFX_SbARTFbK","colab_type":"code","colab":{}},"source":["#Reshaping it\n","df_close = df_close.values.reshape(2265,1)"],"execution_count":0,"outputs":[]},{"cell_type":"code","metadata":{"id":"-kd1EeAMTShd","colab_type":"code","outputId":"4aa1e06e-5ee1-451b-f30b-9af5bd8d4825","executionInfo":{"status":"ok","timestamp":1565632002915,"user_tz":240,"elapsed":930,"user":{"displayName":"Nupur Deshpande","photoUrl":"","userId":"13787757531456301957"}},"colab":{"base_uri":"https://localhost:8080/","height":269}},"source":["import matplotlib.pyplot as plt\n","from sklearn.preprocessing import MinMaxScaler\n","\n","#Ploting it\n","plt.plot(df_close)\n","plt.show()\n","\n","scaler = MinMaxScaler(feature_range=(0, 1))\n","df_close = scaler.fit_transform(df_close)\n"],"execution_count":0,"outputs":[{"output_type":"display_data","data":{"image/png":"iVBORw0KGgoAAAANSUhEUgAAAYEAAAD8CAYAAACRkhiPAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4zLCBo\ndHRwOi8vbWF0cGxvdGxpYi5vcmcvnQurowAAIABJREFUeJzt3Xd8leX9//HXJ4OEEXaYIYYlCIKA\nkaE4UKuI/VbtVFu3Ur/VavuzQ6uto9X6tVM7XK1Vq6JWHNSBUjdVwIBMWZEZDIQZQva4fn+c+5yc\nk5xAxklOcs77+Xjkkftc93Xf5zqHcD7n2uacQ0RE4lNCtAsgIiLRoyAgIhLHFAREROKYgoCISBxT\nEBARiWMKAiIicUxBQEQkjikIiIjEMQUBEZE4lhTtAhxJ3759XVZWVrSLISLSYSxdunSPcy69MXnb\nfRDIysoiJycn2sUQEekwzGxrY/OqOUhEJI4pCIiIxDEFARGROKYgICISxxQERETimIKAiEgcUxAQ\nEYljCgIiIm1g0+5DfJS7J9rFqKfdTxYTEYkFp//ufQC23HtulEsS6og1ATMbYmbvmtlnZrbGzG70\n0nub2QIz2+j97uWlm5k9YGa5ZrbSzCYF3esyL/9GM7us9V6WiIg0RmOag6qAm5xzY4CpwHVmNga4\nGXjbOTcSeNt7DHAOMNL7mQ08CL6gAdwOTAEmA7f7A4eISLzYcaCU655ZRklFVbSLAjQiCDjn8p1z\ny7zjImAtMBg4D3jCy/YEcL53fB7wpPNZBPQ0s4HA2cAC59w+59x+YAEwM6KvRkSknbv/Pxt4bWU+\nv3z1s2gXBWhix7CZZQETgcVAf+dcvndqJ9DfOx4MbA+6LM9Layg93PPMNrMcM8vZvXt3U4ooItLu\nVFXXBI4LisoBqKx20SpOiEYHATPrBswFfuCcOxh8zjnngIi9IufcI865bOdcdnp6o1ZDFRFptyqC\ngsB7631fbJMT28fgzEaVwsyS8QWAp51zL3rJu7xmHrzfBV76DmBI0OUZXlpD6SIiMS3ct/6amg5S\nEzAzA/4OrHXO/T7o1DzAP8LnMuCVoPRLvVFCU4FCr9noTeAsM+vldQif5aWJiMS04Oag9qYx8wRO\nAi4BVpnZci/tZ8C9wPNmdhWwFfimd+51YBaQC5QAVwA45/aZ2S+BT7x8dznn9kXkVYiItGPhagLF\n7WR00BGDgHNuIWANnD4jTH4HXNfAvR4DHmtKAUVEOrrKMDWBkorqKJSkvvbRMyEiEsOqwrT/F5e3\nj5qAgoCISCuqqKphy97ieuntpSagtYNERFrRD59fzmsr8+ult5c+AdUERERaUXAAmDCkZ+C4pFw1\nARGRuPLElZNZunUfH2zYw9xledEuDqCagIhIqwmeH9AtJYkenZM5fXR/uqYkUlJRjW8wZXQpCIiI\ntJKDZbXt/oeCRgN16ZREdY2jvCr6k8gUBEREWsGh8iqeWrQVgNED0ph3/UmBcx9s8K0ftHBj9Hca\nUxAQEWkFt7y4it8v2ADAj84axfiM2k7h00b1A6CqRjUBEZGYU13j+PeKLwKPu3dODjk/eahvP60u\nnaI/NkdBQEQkwlbkHQh53L1z6Id9YoLvo/eLA6VtVqaGKAiIiERYUVnoRLC01NCaQFKCbzm2m19c\nFdJhHA0KAiIiEVZYWhnyuGed5qCkxNo1OcsqoztpTEFARCTC8us083RNCW0OSkqo/eitifJcAQUB\nEZEIO9IKof7mIICqKO81rCAgIhJhxUdYITS4OSh4FFE0NGZ7ycfMrMDMVgelPWdmy72fLf4dx8ws\ny8xKg849FHTN8Wa2ysxyzewBb9tKEZGYU1JRTd9uKQ2eD24O+vUb6wLHxeVVbb6URGNqAo8DM4MT\nnHPfcs5NcM5NwLcB/YtBpz/3n3POXRuU/iBwDTDS+wm5p4hIrCipqKJrSiKXn5jFfV8fX+98cE3A\nb/OeYsbe/ibPfbK9LYpYW5YjZXDOfWBmWeHOed/mvwmcfrh7mNlAoLtzbpH3+EngfOCNJpZXRKTd\nK6mopkunJO74ytiw54P7BADeXLMzsIfvgs92ceHkzFYuYa2W9gmcDOxyzm0MShtqZp+a2ftmdrKX\nNhgIXjc1z0sTEYkZTy3ayo3PfsrB0kq6pSQ2mC8pMfSj97v/XMrSbfsBaOuG8pbOWb4ImBP0OB/I\ndM7tNbPjgZfNLHwoPAwzmw3MBsjMbLuIKCLSEre97Os6HdgjlROyejeYr25NAOBgqX9EUdtGgWbX\nBMwsCfgq8Jw/zTlX7pzb6x0vBT4HjgZ2ABlBl2d4aWE55x5xzmU757LT09ObW0QRkajILyxjRL9u\nDZ5PDBMEKqujs5hcS5qDzgTWOecCzTxmlm5mid7xMHwdwJucc/nAQTOb6vUjXAq80oLnFhFpVzbv\nCd1MvltKww0t4WoCFd7eAm3dHNSYIaJzgI+BUWaWZ2ZXeacuJLQpCOAUYKU3ZPQF4Frn3D7v3PeA\nvwG5+GoI6hQWkZixdW9oEOjSqeE+ATPjmpOHhqQFgkDki3ZYjRkddFED6ZeHSZuLb8houPw5wLFN\nLJ+ISIcwf/XOkMd1F42r69Zzx/Doh5sDj/3NQTlb90e+cIehGcMiIi20r7iCZ+uM78/q26VJ9/Bv\nNbmvuCJi5WoMBQERkRaqu2oowNhBPZp0j7X5ByNVnCZREBARaSH/gnEZvTo3+x5727gG4KcgICLS\nQiXegnHTR/SNyP227yuJyH0aQ0FARKSFiit8NYFvZA8hLSWJRy/NbtR1544bGDZ9yeZ9YdNbQ/R3\nORYR6eD8zUFpqUmsuvPsRl83PL1r2PS6m9C0JtUERERaaMd+305i/dIaXj46nJoGVo2uqmm72cMK\nAiIiLbRh1yH6d0+hZ5dOTbquoa0l23K3MTUHiYg0U2FJJTlb97F+10GO7p/W5Ovr1gTMwLm2XUdI\nQUBEpBHWfFHI0L5d6dKp9mPzW498zLqdRQBcNX1oQ5c2qG5NIDUpkdLKaqoaaidqBWoOEhE5goqq\nGs59YCFjfvFmSLo/AAAcO7h7k+9bU+fDPiXZ95Fc1YY1AQUBEZEjKCqrnRH86krfxvB1l3dISzn8\nWkHh+GNAry6+axO9JUQr27BPQEFAROQIisqqAscvf+rbCuUbD30UkmdiZs8m39ffHJSa7FtxdF+J\nL7BUqzlIRKT9OFReGwTKKmtwzvH5bt/S0UP7duWnM0fTp1vThocCuDpB4PjMXgBUtuEQUXUMi4gc\nQXBNYGHuHuYsqV0x9N0fndbs+1Z7QSAlyfd9PMFrDlq29UCz79lUqgmIiBxBfmFpyOPncrY3kLNp\n/K0+/pqA33/W7orI/RtDQUBE5AjW7yoKTWhgkldT+XeZ9O853NDksdbUmO0lHzOzAjNbHZR2h5nt\nMLPl3s+soHO3mFmuma03s7OD0md6ablmdnPkX4qISOQdKq/i4fc3MahHaiAtUh/VXx4/CIDso3x9\nAe0yCACPAzPDpP/BOTfB+3kdwMzG4Nt7eKx3zV/NLNHbfP4vwDnAGOAiL6+ISLv269fXAvBFYVkg\nLVL7AE8d1ofcu8/h2MG+DWiWbWu7vgC/IwYB59wHQGPXNT0PeNY5V+6c24xvU/nJ3k+uc26Tc64C\neNbLKyLSrpVWVtdLS0iI3HbwSYkJ7DlUHnh8xuh+APzjv5sbuiSiWtIncL2ZrfSai3p5aYOB4B6T\nPC+tofSwzGy2meWYWc7u3btbUEQRkZbpG2bop39SV6QkJdZ+FHfyRgo9/P6miD5HQ5obBB4EhgMT\ngHzgdxErEeCce8Q5l+2cy05PT4/krUVEmsQ/M/i33ziOgV6/QCRrAgDfOD6Dy0/MYuUdZ5HsBYT+\n3Zs+76A5mhUEnHO7nHPVzrka4FF8zT0AO4AhQVkzvLSG0kVE2rWNu4qYPLQ3Xz8+g69NyiDBIl8T\nSE1O5I6vjKV7anKgJhCuBtIamhUEzCx4T7QLAP/IoXnAhWaWYmZDgZHAEuATYKSZDTWzTvg6j+c1\nv9giIm1jRV4hA7r7agDJiQnUOHARGx9U306vA/rtdQWt9hzBGjNEdA7wMTDKzPLM7CrgPjNbZWYr\ngRnADwGcc2uA54HPgPnAdV6NoQq4HngTWAs87+UVEWmXDpVXMX91PgDzVvgWjUtK9NUAtu/zTR47\nIatX+IsjoHfXpm1Q01xHXDbCOXdRmOS/Hyb/3cDdYdJfB15vUulERKLk9N++R0GRb9ROqrfE844D\npSG/H76kcRvKN4W/OegXX26bUfSaMSwiEoY/AAA8eeUUAJ5ZvC0kT2t8W/fPHvbXOlqbgoCIyBH4\nRwVNzurd6s+V5AUBi9iUtMNTEBAROQL/Am8/OntUIO2Fa6e1ynP5awJVbbSctIKAiMgR+PsE/O31\nACObsbF8Y7T1YnIKAiIiddTd2ctfE0gKmiSWltI627EEagJttMWkgoCISB3BewoDgVm8/k1fRvbr\nFvFZw37+QNNWW0wqCIiI1HGgpDJsun8xua6tVAsASEzwfSxXKQiIiETH3uLysOndU30f/lOH9Wm1\n527rmoD2GBYRqeNXr60Nmz6yfxqvfn86xwzs3mrPXTs6SDUBEZGoGNa3W4Pnjh3cI/BB3RpqawIa\nIioiEhXpaSkhw0Hb0hSvqcm/21hrU3OQiEgdZZXVdE5O5JFLjqcszM5irelLY/qz9LYz6dNGS0kr\nCIiI1FFa4QsCp43qF5Xnb6sAAGoOEhEJ4ZzjvQ0Frdru356oJiAiEuTjTXvZdTD8ENFYpJqAiEiQ\nwgYmisWqxuws9piZFZjZ6qC035jZOjNbaWYvmVlPLz3LzErNbLn381DQNcd7u5HlmtkDZhHepFNE\nJAIOlVcB8NzsqVEuSdtoTE3gcWBmnbQFwLHOufHABuCWoHOfO+cmeD/XBqU/CFyDb9/hkWHuKSIS\ndcVeEBjRr+G5ArHkiEHAOfcBsK9O2lvevsEAi4CMw93D25i+u3NukXPOAU8C5zevyCIirae4wjck\ntFtqfHSZRqJP4ErgjaDHQ83sUzN738xO9tIGA3lBefK8NBGRdiVvfwmpyQmkJCVGuyhtokWhzsxu\nBaqAp72kfCDTObfXzI4HXjazsc2472xgNkBmZmZLiigi0iTb9pUwakDrrQ3U3jS7JmBmlwNfBr7t\nNfHgnCt3zu31jpcCnwNHAzsIbTLK8NLCcs494pzLds5lp6enN7eIIiJNtqeogv5pbTdZK9qaFQTM\nbCbwE+ArzrmSoPR0M0v0jofh6wDe5JzLBw6a2VRvVNClwCstLr2ISITtPlROehwFgSM2B5nZHOA0\noK+Z5QG34xsNlAIs8EZ6LvJGAp0C3GVmlUANcK1zzt+p/D18I4064+tDCO5HEBGJusrqGvaXVNC3\nDZdtiLYjBgHn3EVhkv/eQN65wNwGzuUAxzapdCIiraymxlFeVUPnTols21eCc5DRq3O0i9VmNGNY\nROLa7xds4JhfzGdVXiG/mb8egMzeXaJcqrYTHwNhRUQa8Od3cwG4+NFFFHkTxXp17RTNIrUp1QRE\nRCAQAAB6dkmOYknaloKAiEgdPTvHT01AzUEiEtcSE4xqb1P3v357EmePHRA3ewmAagIiEsdqahzJ\nibUf+LPGDYyrAACqCYhInCqrrGb0z+dHuxhRp5qAiMSlRZv2hjx++uopUSpJdKkmICJx6c/v+IaG\nDundmdduOJnuqfEzIiiYgoCIxJ3cgiJytu4H4MOfnB7l0kSXmoNEJO7sLqqIdhHaDQUBEYk7/iGh\n/3PcoCiXJPoUBEQk7lT7tkDh8hOzoluQdkBBQETiTo0XBOJsSkBYCgIiElOcc7z0aR6HgtYCqqvG\naw6Kt4lh4SgIiEhMWbeziB8+t4Kb565sMI+/TyDBFAQUBEQkpqzfWQRAfmFZg3m8GKAgQCODgJk9\nZmYFZrY6KK23mS0ws43e715eupnZA2aWa2YrzWxS0DWXefk3mtllkX85IhLvfvzCCgBGpHdrMI+/\nT0DNQY2vCTwOzKyTdjPwtnNuJPC29xjgHHwbzI8EZgMPgi9o4NufeAowGbjdHzhERJpr+74S9hfX\njvuvrPZ9wFf5v+6HUdsc1Lpl6wgaNWPYOfeBmWXVST4P3wb0AE8A7wE/9dKfdM45YJGZ9TSzgV7e\nBf6N581sAb7AMqdFr0BE4lZ1jePk+94F4PoZIxif0SNwrrQyfMfw/NU7+f6cTwFIUBRo0bIR/Z1z\n+d7xTqC/dzwY2B6UL89LayhdRKRZ5i7LCxz7t4n0e33VTsqrqklJSuQ3b67jr+99zqZ7ZnHtU0sD\nedQnEKGOYe9bf8N1ryYys9lmlmNmObt3747UbUUkxjy9aOthz4+6bT7OOf7y7uc4B+9vCP08SVQQ\naFEQ2OU18+D9LvDSdwBDgvJleGkNpdfjnHvEOZftnMtOT09vQRFFJJbtLiqvl3b+hEHcc8G4wONj\nb38zcPzYf7eE5E3Q+MgWBYF5gH+Ez2XAK0Hpl3qjhKYChV6z0ZvAWWbWy+sQPstLExFpsqcWbeWL\nMMNAX17+BRdPyQw8Lq6oDhx/UKcm0LWTFlJu7BDROcDHwCgzyzOzq4B7gS+Z2UbgTO8xwOvAJiAX\neBT4HoDXIfxL4BPv5y5/J7GISFNUVddw28u+Eeu3nDM65Fxw53BDrp8xgm9PyaRX1/jZUL4h5lzE\nmvJbRXZ2tsvJyYl2MUSknThYVsn4O94KPF7/q5mkJCXy7roCrnj8E1763olMzOzFy5/u4AfPLQdg\nclZvEhOMjzftZVCPVD665YxoFb9NmNlS51x2Y/KqRUxEOpQN3oxgv5SkRABmjO7H6jvPZmKmb/rR\n+RMH88GPZ9A9NYmrTx5KcpLv4+6Biya2bYHbOTWIiUiHErww3JZ7zw051y0l9CMts08XVt5xNgAj\n+nXjuU+2Mz6jZ+sXsgNREBCRDmXvId/s4NdvOLlJ1w1L78Yts45pjSJ1aAoCItIh5GzZx9cf+pgh\nvTuTnGiM7N/w2kDSeAoCItIhfP2hjwHYvq8UgOREdWlGgt5FEekQUpJqP67+cvGkw+SUplAQEJEO\noUfn5MDxueMHRrEksUXNQSLS7uXtL6GgqJz/96WjuUybw0eUgoCItEs7C8vYc6icYeld+fvCzQCM\ny+gRUiOQllMQEJE2UVha2aQP8J+9tIp31hWEpPXRMg8RpyAgIq3uXznb+fELK5k+oi8XT8lk1riG\n2/Qf+eBzXvr0C9bmH6x3bmjfrq1ZzLikICAire7HL6wEYGHuHhbm7mHJrWfQLy21Xr7yqmrueX1d\n4PH5EwZx01mj2LK3mKw+XUlLVVNQpCkIiEirqgmz12/urkNhg8DLn4ZuMXLv18aTmpzIkN5dWq18\n8U5DREWkVRWWVgIwbVifQFregdJ6+Uorqvnp3FUhaanJia1bOFEQEJHWdf/bGwGYdFTtwm3lVTX1\n8q3MOxDy+KczR9fLI5Gn5iARaTWvrvyCzXuKATh33CBmnzKc4+58i799uInzJwyic3IiSd7yD4/9\n1zcMdP4PTmZo36500rIQbUJBQERaxfZ9JVz/zKcA9O3WiTGDulNW6dvqceveEsZ5G8Ns/vUszIw9\nhyrol5bC6AHdo1bmeNTsIGBmo4DngpKGAb8AegLXAP7NPH/mnHvdu+YW4CqgGrjBOac9hkViSFll\nNe+uK+D5nO0E9wfv8ZZ/Dl7/x2/oLa8Hjs88pl+rl1FCNTsIOOfWAxMAzCwR2AG8BFwB/ME599vg\n/GY2BrgQGAsMAv5jZkc756oRkZjwsxdX8WKdET4AP//yGADM7LDXl1To46CtRarR7Qzgc+fc1sPk\nOQ941jlX7pzbjG8j+skRen4RiaL3N+zmin8s4dVV+WHPnzdhUOD4nGMHNHifqur2ved5LIpUn8CF\nwJygx9eb2aVADnCTc24/MBhYFJQnz0urx8xmA7MBMjMzI1REEWkt1zyZQ0XQiJ+lt53JZ/kHmT6i\nb71v/3/41gTeWD0/JO2l753IBX/9iB98aWSblFdqtbgmYGadgK8A//KSHgSG42sqygd+19R7Ouce\ncc5lO+ey09PTW1pEEWkFW/cWk1twCCAkAIwd1J0+3VI4eWR62Oaf4LH/xw7uzjezM5iY2Yst957L\nicP7tn7BJUQkagLnAMucc7sA/L8BzOxR4FXv4Q5gSNB1GV6aiHQgB0oqKK2s5tTfvAfA+IweAAzu\n2ZnLT8zinHENN/fU9cK1J2pCWJRFIghcRFBTkJkNdM75GwYvAFZ7x/OAZ8zs9/g6hkcCSyLw/CLS\nhibctSDk8cq8QgDu+eo4Tj26cTX3r04azIvLdoQdLSRtq0VBwMy6Al8CvhuUfJ+ZTQAcsMV/zjm3\nxsyeBz4DqoDrNDJIpGMpLq9q8NyUob0bfZ/7vjaeu8479oijhaT1tSgMO+eKnXN9nHOFQWmXOOfG\nOefGO+e+ElQrwDl3t3NuuHNulHPujZY8t4i03La9Jdz0/ApKGzk0s6CoPOTxklvPCBw3pVknKTGB\nbimaq9oeqC4mEmfKKqtZt/MgZZXV/Omdjcxdlseba3Y26to7/70GgE6JCXz4kxn0S0vlvAmDmDCk\n5xGulPZKoVgkjnz0+R4ufnRxvfTSysbVBN5b71sI4N/fnx5Y3vn+CydGroDS5lQTEIkj4QIAEFjT\np7H6paVEojjSDigIiMSJg2WVDZ4rq6y/tHNdVdW1eXp20Q5fsUJBQCRObN9XEjg+d3zoHr+NqQn4\n9wC4aHKmRvXEEAUBkTiRf6AscPy1Sb4VWwb37EynpAT2Fpfzxqp8/uGt6V/X7qJyxt7uW/R39IC0\n1i+stBl1DIvEiaufzAEIbPK+9LYz6ZSUwNVP5JCzZT9PLdoGwBUnDa137Z/e2Rg41gSv2KIgIBIH\nKoPa89O7+Tp1+3i/D5ZVsW5n0WGvT0ywsMfS8Smki8SBz744CMAfvzWhXnv+tr3Fh722qrqGF5fV\nLvO162DZYXJLR6MgIBIHvjhQCsCoMO35j16W3eB1ldU1jLj1DQpLa0cWjR3cI/IFlKhRc5BIHCgq\n8635k5Za/7/8uDof6hVVNSQnGhXVNWzcdSiQftu5x3Dh5Ewt9xBj9K8pEgfe3+ib6ZuWWn98f+c6\na/6UVlTzqwXrefLj2o0CExOMq6YP1dDQGKTmIJEYV1hayWsrfes4hvsWn5QY+jFwqKIqJAAA5N59\njgJAjFIQEIlx/v4AaHhkz/wfnMw3szMAePC93HrnFQBil5qDRGLUc59sY21+EY9/tAWAYwZ2bzDv\n6AHd+Z/jBvF8Tl5gvoCfVgiNbQoCIjFgzReFnPvAQq44KYtzxw0kO6s3P527KiTPM1dPOew9utfp\nL1h710wKisro3z014uWV9iMSG81vMbNVZrbczHK8tN5mtsDMNnq/e3npZmYPmFmuma00s0ktfX6R\neFdT4zj3gYUA/OO/W/j6Qx/zzrpdIXlSkhLo1bXTYe/Trc7Ioc6dEjmqT1ftARzjItUnMMM5N8E5\n5x9wfDPwtnNuJPC29xh8m9KP9H5mAw9G6PlF4pZ/YbdgVz6eE/J4xe1nHfE+w9O7RaxM0nG0Vsfw\necAT3vETwPlB6U86n0VATzMbGO4GItI4FV4Q+P7pI7h02lFh8zT12/yfL9ZGMfEiEkHAAW+Z2VIz\nm+2l9Q/aW3gn0N87HgxsD7o2z0sTkWYqr/YtA92ve+phO3+b4qjeXSNyH2n/ItExPN05t8PM+gEL\nzGxd8EnnnDMz15QbesFkNkBmZmYEiigSu/6buwfw1Qj69KjtxH366incN39d2KUijiTT2zpSYl+L\ng4Bzbof3u8DMXgImA7vMbKBzLt9r7inwsu8AhgRdnuGl1b3nI8AjANnZ2U0KICLx5ndvbQCgtKKK\nzp18zT5ThvbmpBF9eeX66U26V2pyAmWVNfTQzmFxo0XNQWbW1czS/MfAWcBqYB5wmZftMuAV73ge\ncKk3SmgqUBjUbCQizZC33zcZ7OqTh9HJm/2b0MzJXe/cdBpz//fEiJVN2r+W1gT6Ay95swmTgGec\nc/PN7BPgeTO7CtgKfNPL/zowC8gFSoArWvj8InEtv7B2NnBqciI9Ovu+wR/XzAleg3p2ZlDPzhEp\nm3QMLQoCzrlNwHFh0vcCZ4RJd8B1LXlOkXhTVllNwcFyMvuEttOvyisMBIE7/mcMAMcO7sG/rp3G\n8Zm92ryc0jFpxrBIO3fnv9cwZ8n2wLaQq3cUcrC0kov/tjiQ59RR/QLHJ2T1jkYxpYNSEBBp5z7L\n9239OPnutxvMM6inlnaQ5tEqoiLt3NhBhx/7f+VJQ0lJ0tIO0jyqCUjcKausJjU5kY9y9zDpqF7t\nfm2cVO8DPqtPF6YN78uXxw9kRd4BzhrTnxH9mj4HQCSYgoDElW889BGfbNkfkvbc7KlMGdYnSiU6\nssrqGnp37cR7P54RSDtpRN8olkhiiZqDJC785s11ZN38Wr0AAPCX9z6PQonC211Uzpwl26is9q0H\n5Jzjn4u2sq+4Isolk1ilmoDEhb+8W/tB3zk5ke6dk9h1sBzwfdD6vb9hNwN7pHJ0/7ZtZnHOMfme\nt9ld5CvTwo17uHTaUXzrkUUAdErS9zVpHQoCEvOeXly7X+7UYb15+uqpJCYYu4vKOeHu/zBteB8q\nq2uYdf+HbCw4REavziz86eltWsbz//LfQAAAeG1VPq+tqp1Mf+MZI9u0PBI/FAQkZmXd/FrI4wU/\nPIWRQd/w+3bzbbJy3/z1LN60j40FhwDfMgzOuSPuq3ugpILb563h6unDGJfRo9nlLCypZEVeIQAP\nfWcS1z61LOT8C9dOY6Imf0krURCQuNAtJYkR/UI3TQn+kH9/w24ALp6SyTOLt/HOugLOOKY/DXHO\nMeGuBYBv1677vu6bOF9YWklKUkKTRhzlHSgB4Mdnj2LmsQNZ8YuzKK2sJinR6NstpdH3EWkONTRK\nzEoI+iL/2g3Tj/jN/iczR/HtKb6ly696Iofyqmq+OFAaNu/WvSWB45IK33r+d/57Dcfd+RbXPJkT\n9hqA1TsK2XOoPCStuNx3/XEZvvV+enRJZkCPVAUAaROqCUhMWpl3gBoH4wb34HunDeeoPuE3SZlz\nzVTMoLSymlNHppMQFDlG3TaENlayAAAOKElEQVQfgPd+dBpZfbtSXF7F2NvfrHeP8qoabnlxFXOW\nbAPgw417+M7fFvOPK04gObH2e9Y/F23l5y+vJqtPl5DhnsXlVQB0TWnf8xUkNqkmIDElt+AQ33z4\nY77y5/8CMG14H84Z1/AOptOG92HqsD7MGNUvEACevHJySJ5dB8sAeGFpXth7LPhsVyAAjBvs6xtY\nmLuHkbe+wV3//gyALXuK+fnLq33HQbUIgENeEOiWou9k0vb0VycdTsHBMt76bBcHyyr53mkjAumv\nLN/Bjc8uD8l788zRTb7/icNDJ469suILqp3j9nlrQtK/lT2EXUVlvLfe15/w2OXZzBjVj+n/9y47\nvGakx/67mWeWbA0pJ8DfF24m/0Apl52Yxb1v+Dbj66ogIFFgwWOk26Ps7GyXk9NwG6vEn2N+Pp/S\nSl87+llj+vPwJcdTUlEdtqlmy73nNus5Xlm+g5ueX0FVTf3/H8H33HGglJPufQeATffMCtQmqqpr\n+OqDH7HSG/Xj9/Alx/Pdfy4N+5xr75oZ2BlMpCXMbKlzLrsxedUcJB3KgZKKQAAAeOuzXVzx+Cdc\n+fgnAPTqksyXx/uafz66uflj/c+bMLje9aMHpPHaDaHbNQ7u2ZmXrzuJ9b+aGdKfkJSYwLOzp4bk\nnTCkJ2ePHRC2w/f4o3opAEhUKAhIm1u9o9D3LdtbGqEx3l67izvmreGnc1cCvk3Ul952Jp2TE3lv\n/W4Wb95HgsHTV0/lt984jo9uPr3FO2T1657KH781IfD4+WunMXZQ/fkAE4b0DLuKZ5dOSbxz06mB\nxw9+ZxIA7/7oVM6fMIjnvzstcO7PF09sUVlFmqvZzUFmNgR4Et8Wkw54xDl3v5ndAVwD7Pay/sw5\n97p3zS3AVUA1cINzrn79vQ41B8WW9zfs5rLHlgBwzwXjSEzwjZpZveMgr35/Ov27p3LC3f9haN+u\n/OmiiRQUldE5OYmLHl0UuMfZY/vz8CW+mu7a/IOcc/+HgC8wtMbCalv3FjOgR2qzl2t+f8Nupg3r\nE3bphx0HStmyp1gLwklENaU5qCVBYCAw0Dm3zNtsfilwPr79hA85535bJ/8YYA4wGRgE/Ac42jlX\nzWEoCLR/q/IK+dM7G/nuqcPZWVjGuePrj8b5YMNunl68lTfX7GrRcyUmGDm3nkmvrp1adB+RWNaU\nINDs4QjOuXwg3zsuMrO1wODDXHIe8KxzrhzYbGa5+ALCx80tg0TfzsIy/ufPCwFf+zzAgB4ncvxR\nvmUO9hwq56R736G8qrbpZ/SANNbtLAo8TktJosgbJtmQKUN789TVU0gwIzHh8JO+RKTxIjImzcyy\ngInAYuAk4HozuxTIAW5yzu3HFyAWBV2WRwNBw8xmA7MBMjMzI1HEmPXe+gLGDe7B4s37+N7Ty/jw\nJzMY0rvLkS9sAf+6Ojc++ymvLP+i3vmvPfgRb/3wFI7un8a3H10cEgAA5v/gFA6WVfJR7h5G9OtG\n987JLNm8j+uf+ZQzRvdj9inDAqtnLvzpDDJ6te7rEYlnLR4iambdgPeBu51zL5pZf2APvn6CX+Jr\nMrrSzP4MLHLOPeVd93fgDefcC4e7v5qDGlZYUslxd71VL725wyLBt+vWqyvzOefYASHj1mtqHDsO\nlJLRqzNjfvFmyAidCUN6Mvd/T2TL3mJumbuKJVv2AdA9NYmDZVWMHpDGlScNJe9AKROH9GTG6H71\nnreuTbsPsa+4gmxtmi7SZG3SHOQ9UTIwF3jaOfcigHNuV9D5R4FXvYc7gCFBl2d4aXIENTUOM3h5\n+Q56du4U+BBdsDZ8+3pldQ2Gb5hiU81dlsetL63mR/9awdePz+Der47jg427ufLx8IH4tRumB0bM\nDE/vxvPXTgus3nmwzNfE89jlJzR5pM6w9G4MS29y8UWkiZodBMy3GtffgbXOud8HpQ/0+gsALgBW\ne8fzgGfM7Pf4OoZHAkua+/zxorrGMev+D1m/q7YNPffuc3hq0Vbu8JYk8Buf0YOVeYWMvPUNAD6/\nZ1aT28/3B+1g9cLSvAaXSrhuxnB+fHb42bi/PG8s97+dy55D5Qzr27XFQzVFpPW0pCZwEnAJsMrM\n/HP1fwZcZGYT8DUHbQG+C+CcW2NmzwOfAVXAdUcaGRRvFm/ay98WbubiKZnMGOX7tv/Wmp0hAQBg\n5G1v4G/F65aSxPkTBwV2xAqeobp5T3G95ZPDWb2jkC//aSGv3TCdkopqEhOMh79zPFcHrYZ5wxkj\nOS6jB1c9kcMFEwdz3YwRDd7vkmlZXDIti6KyyiOu3Cki0aVlI9pAQVEZzkH/7qnkF5bSPy01ZHYp\nwLvrC7jiH5+EpD30neO59infEgP/uPwEMvt04YzfvR84P2FIT57/7jSSE40aBz9/ZTXPLN7GTV86\nmt8t2BDIt+6XM0lNTgw8xznHDuCKk4ZyQlYvzIxTf/NuYGnk0QPSyNtfyuo7z6awpJI/vr2Bn806\nJmQ1TBFp39pknkBb6YhBoKKqBofjR/9ayb9X1I6euXr6UP62cDMXTBzMH7yZqB/l7uHivy0+7P0u\nPzGLO74yFgjdLetvl2Zz5pjajU8OlVfx4YbdnHFMf46+7Y1A+pDenXnkkuzApCq/9LSUkC0N/cZn\n9GDe9dPrpYtIx6Ag0MaccxRXVPP0oq1s2VvMnCXbj3jNj88exXemHFVvdM/mX8+iqsZx92trefyj\nLYzqn8abPzwlJI8/EGz+9awGm1tO/PXbfFFY1ujX8M5Np/LNhxdRXlXN/B+cwmC144t0WAoCbcA5\nx77iCkoqqpnx2/fCrjZ56tHpnD9xEKeP7s9xd/o+7G88YyT3v70xJN+1pw4nMQGuOXkYPbsceSbs\n0q37WbfzIN+eclSDeWpqHDXO8eHGPVzxeG0z0xs3nsyBkkr2l1Tw3CfbufHMkYwd1L3ZSyKISPuj\nINAE1TWOBKPJHZjPLtnGzS+uCnvu5JF9+edVU0LSCorK+LygmGnD+7B4097AZCho3iiepqqpcfX6\nIUQkNrXZPIGO7rWV+Vz3zDLg8E0rda3MO1AvALxy3Un0TUvhqUVbufGMkfWu6ZeWSr+0VACmDOvD\nlnvP5dNt+xkzqHubLIOgACAi4cRkTaCmxvHSpzvYW1zO7FOGh5zLLShiwWcFzFmyjW37arf5G5be\nledmTyM97fCbe2/bW8Ipv3kXgIe+M4nTRvUjOTFB69mISLsR981BB8sqGX9HbYfrXeeN5dJpWby7\nriCkfRzgJzNHcd/89SFp/dJS+EZ2Bj8882hW5BUytG9Xrn7iE5ZtOxDI89RVU5g+Usv/ikj7E/dB\nAGD+6p2BMfZ1TR/Rl/EZPfjqpAxG9OvGffPX8df3Pm/0vft3T2Hxz85scplERNqCgoDHOccf/7Mx\nZDTOvOtPYnxGz3p5q6pr+O1bGxjSuzOr8gp59pPQYZ7XzxjBTWcdTWW1IynB1MYuIu2WgkAELN9+\ngJ+/vJo/XzyRo/p0bfPnFxFpLo0OioAJQ3ry7+9r1qyIxDYtCCMiEscUBERE4piCgIhIHFMQEBGJ\nYwoCIiJxTEFARCSOKQiIiMQxBQERkTjW7mcMm9luYGszL+8L7IlgcToyvRe19F7U0nsRKlbej6Oc\nc+mNydjug0BLmFlOY6dOxzq9F7X0XtTSexEqHt8PNQeJiMQxBQERkTgW60HgkWgXoB3Re1FL70Ut\nvReh4u79iOk+ARERObxYrwmIiMhhxGQQMLOZZrbezHLN7OZol6ctmNkWM1tlZsvNLMdL621mC8xs\no/e7l5duZvaA9/6sNLNJ0S19y5nZY2ZWYGarg9Ka/PrN7DIv/0Yzuywar6WlGngv7jCzHd7fx3Iz\nmxV07hbvvVhvZmcHpXf4/0dmNsTM3jWzz8xsjZnd6KXH5d9GWM65mPoBEoHPgWFAJ2AFMCba5WqD\n170F6Fsn7T7gZu/4ZuD/vONZwBuAAVOBxdEufwRe/ynAJGB1c18/0BvY5P3u5R33ivZri9B7cQfw\nozB5x3j/R1KAod7/ncRY+X8EDAQmecdpwAbvNcfl30a4n1isCUwGcp1zm5xzFcCzwHlRLlO0nAc8\n4R0/AZwflP6k81kE9DSzgdEoYKQ45z4A9tVJburrPxtY4Jzb55zbDywAZrZ+6SOrgfeiIecBzzrn\nyp1zm4FcfP+HYuL/kXMu3zm3zDsuAtYCg4nTv41wYjEIDAaCd4nP89JinQPeMrOlZjbbS+vvnMv3\njncC/b3jeHmPmvr6Y/19ud5r4njM3/xBHL0XZpYFTAQWo7+NgFgMAvFqunNuEnAOcJ2ZnRJ80vnq\ntHE7FCzeXz/wIDAcmADkA7+LbnHalpl1A+YCP3DOHQw+F+9/G7EYBHYAQ4IeZ3hpMc05t8P7XQC8\nhK86v8vfzOP9LvCyx8t71NTXH7Pvi3Nul3Ou2jlXAzyK7+8D4uC9MLNkfAHgaefci16y/jY8sRgE\nPgFGmtlQM+sEXAjMi3KZWpWZdTWzNP8xcBawGt/r9o9iuAx4xTueB1zqjYSYChQGVY1jSVNf/5vA\nWWbWy2suOctL6/Dq9PlcgO/vA3zvxYVmlmJmQ4GRwBJi5P+RmRnwd2Ctc+73Qaf0t+EX7Z7p1vjB\n18O/Ad/ohlujXZ42eL3D8I3eWAGs8b9moA/wNrAR+A/Q20s34C/e+7MKyI72a4jAezAHXzNHJb72\n2qua8/qBK/F1juYCV0T7dUXwvfin91pX4vugGxiU/1bvvVgPnBOU3uH/HwHT8TX1rASWez+z4vVv\nI9yPZgyLiMSxWGwOEhGRRlIQEBGJYwoCIiJxTEFARCSOKQiIiMQxBQERkTimICAiEscUBERE4tj/\nB7Q6A46VSXOJAAAAAElFTkSuQmCC\n","text/plain":["<Figure size 432x288 with 1 Axes>"]},"metadata":{"tags":[]}}]},{"cell_type":"code","metadata":{"id":"Qwu51GYWUIul","colab_type":"code","outputId":"9d7265fb-97ea-4a70-fb15-462cddfb0eb9","executionInfo":{"status":"ok","timestamp":1565632004380,"user_tz":240,"elapsed":482,"user":{"displayName":"Nupur Deshpande","photoUrl":"","userId":"13787757531456301957"}},"colab":{"base_uri":"https://localhost:8080/","height":35}},"source":["train_size = int(len(df_close) * 0.80)\n","test_size = len(df_close) - train_size\n","train, test = df_close[0:train_size,:], df_close[train_size:len(df_close),:]\n","print(len(train), len(test))"],"execution_count":0,"outputs":[{"output_type":"stream","text":["1812 453\n"],"name":"stdout"}]},{"cell_type":"markdown","metadata":{"id":"Gyaecf7d87tN","colab_type":"text"},"source":["**Normalize the values such that prices are reduced to normalized values and I can predict stock movement instead of actual prices**"]},{"cell_type":"code","metadata":{"id":"Ria-4F-4V_Xn","colab_type":"code","colab":{}},"source":["# convert an array of values into a dataset matrix\n","def create_dataset(dataset, look_back=1): #keep only similar value data if a data point is too far of the last one jump to the next value\n","\tdataX, dataY = [], []       #put this in data x\n","\tfor i in range(len(dataset)-look_back-1): #let dataY have only output values\n","\t\ta = dataset[i:(i+look_back), 0]\n","\t\tdataX.append(a)\n","\t\tdataY.append(dataset[i + look_back, 0])\n","\treturn np.array(dataX), np.array(dataY)"],"execution_count":0,"outputs":[]},{"cell_type":"code","metadata":{"id":"atRqDg15WGbV","colab_type":"code","colab":{}},"source":["# reshape into X=t and Y=t+1\n","look_back = 1\n","trainX, trainY = create_dataset(train, look_back)\n","testX, testY = create_dataset(test, look_back)"],"execution_count":0,"outputs":[]},{"cell_type":"code","metadata":{"id":"PVeSyYXXWNlT","colab_type":"code","colab":{}},"source":["trainX = np.reshape(trainX, (trainX.shape[0], 1, trainX.shape[1]))\n","testX = np.reshape(testX, (testX.shape[0], 1, testX.shape[1]))"],"execution_count":0,"outputs":[]},{"cell_type":"code","metadata":{"id":"ci3EQtyrWSgh","colab_type":"code","outputId":"624d22aa-ecde-49b4-c320-c449e50ed8d2","executionInfo":{"status":"ok","timestamp":1565632049768,"user_tz":240,"elapsed":966,"user":{"displayName":"Nupur Deshpande","photoUrl":"","userId":"13787757531456301957"}},"colab":{"base_uri":"https://localhost:8080/","height":141}},"source":["#Step 2 Build Model\n","model = Sequential()\n","\n","model.add(LSTM(\n","    input_dim=1,\n","    output_dim=50,\n","    return_sequences=True))\n","model.add(Dropout(0.2))\n","\n","model.add(LSTM(\n","    100,\n","    return_sequences=False))\n","model.add(Dropout(0.2))\n","\n","model.add(Dense(\n","    output_dim=1))\n","model.add(Activation('relu'))\n","\n","start = time.time()\n","model.compile(loss='mse', optimizer='rmsprop', metrics=['mae'])\n","print ('compilation time : ', time.time() - start)"],"execution_count":0,"outputs":[{"output_type":"stream","text":["/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:6: UserWarning: The `input_dim` and `input_length` arguments in recurrent layers are deprecated. Use `input_shape` instead.\n","  \n","/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:6: UserWarning: Update your `LSTM` call to the Keras 2 API: `LSTM(return_sequences=True, input_shape=(None, 1), units=50)`\n","  \n"],"name":"stderr"},{"output_type":"stream","text":["compilation time :  0.031675100326538086\n"],"name":"stdout"},{"output_type":"stream","text":["/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:15: UserWarning: Update your `Dense` call to the Keras 2 API: `Dense(units=1)`\n","  from ipykernel import kernelapp as app\n"],"name":"stderr"}]},{"cell_type":"code","metadata":{"id":"tzikjpoYWXKw","colab_type":"code","outputId":"bfecebd3-46d7-448e-d799-0ff42355680a","executionInfo":{"status":"ok","timestamp":1565632068696,"user_tz":240,"elapsed":16435,"user":{"displayName":"Nupur Deshpande","photoUrl":"","userId":"13787757531456301957"}},"colab":{"base_uri":"https://localhost:8080/","height":1000}},"source":["model.fit(\n","    trainX,\n","    trainY,\n","    batch_size=128,\n","    nb_epoch=80,\n","    validation_split=0.05)"],"execution_count":0,"outputs":[{"output_type":"stream","text":["/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:6: UserWarning: The `nb_epoch` argument in `fit` has been renamed `epochs`.\n","  \n"],"name":"stderr"},{"output_type":"stream","text":["Train on 1719 samples, validate on 91 samples\n","Epoch 1/80\n","1719/1719 [==============================] - 2s 932us/step - loss: 0.0118 - mean_absolute_error: 0.0760 - val_loss: 0.0522 - val_mean_absolute_error: 0.2278\n","Epoch 2/80\n","1719/1719 [==============================] - 0s 91us/step - loss: 0.0066 - mean_absolute_error: 0.0630 - val_loss: 0.0383 - val_mean_absolute_error: 0.1950\n","Epoch 3/80\n","1719/1719 [==============================] - 0s 86us/step - loss: 0.0051 - mean_absolute_error: 0.0564 - val_loss: 0.0274 - val_mean_absolute_error: 0.1649\n","Epoch 4/80\n","1719/1719 [==============================] - 0s 86us/step - loss: 0.0033 - mean_absolute_error: 0.0448 - val_loss: 0.0111 - val_mean_absolute_error: 0.1050\n","Epoch 5/80\n","1719/1719 [==============================] - 0s 86us/step - loss: 0.0016 - mean_absolute_error: 0.0317 - val_loss: 0.0045 - val_mean_absolute_error: 0.0670\n","Epoch 6/80\n","1719/1719 [==============================] - 0s 87us/step - loss: 5.8823e-04 - mean_absolute_error: 0.0186 - val_loss: 7.0137e-04 - val_mean_absolute_error: 0.0260\n","Epoch 7/80\n","1719/1719 [==============================] - 0s 92us/step - loss: 2.8856e-04 - mean_absolute_error: 0.0120 - val_loss: 3.9329e-04 - val_mean_absolute_error: 0.0192\n","Epoch 8/80\n","1719/1719 [==============================] - 0s 88us/step - loss: 2.2058e-04 - mean_absolute_error: 0.0103 - val_loss: 1.9962e-04 - val_mean_absolute_error: 0.0132\n","Epoch 9/80\n","1719/1719 [==============================] - 0s 95us/step - loss: 2.3068e-04 - mean_absolute_error: 0.0108 - val_loss: 4.4103e-05 - val_mean_absolute_error: 0.0051\n","Epoch 10/80\n","1719/1719 [==============================] - 0s 84us/step - loss: 2.1667e-04 - mean_absolute_error: 0.0103 - val_loss: 1.0165e-04 - val_mean_absolute_error: 0.0090\n","Epoch 11/80\n","1719/1719 [==============================] - 0s 88us/step - loss: 2.1358e-04 - mean_absolute_error: 0.0101 - val_loss: 1.1519e-04 - val_mean_absolute_error: 0.0097\n","Epoch 12/80\n","1719/1719 [==============================] - 0s 89us/step - loss: 1.9163e-04 - mean_absolute_error: 0.0097 - val_loss: 1.6545e-04 - val_mean_absolute_error: 0.0119\n","Epoch 13/80\n","1719/1719 [==============================] - 0s 99us/step - loss: 2.0138e-04 - mean_absolute_error: 0.0099 - val_loss: 1.3723e-04 - val_mean_absolute_error: 0.0107\n","Epoch 14/80\n","1719/1719 [==============================] - 0s 88us/step - loss: 2.1103e-04 - mean_absolute_error: 0.0104 - val_loss: 4.6632e-05 - val_mean_absolute_error: 0.0055\n","Epoch 15/80\n","1719/1719 [==============================] - 0s 91us/step - loss: 1.9416e-04 - mean_absolute_error: 0.0097 - val_loss: 6.5333e-05 - val_mean_absolute_error: 0.0068\n","Epoch 16/80\n","1719/1719 [==============================] - 0s 105us/step - loss: 1.7982e-04 - mean_absolute_error: 0.0095 - val_loss: 6.3731e-05 - val_mean_absolute_error: 0.0067\n","Epoch 17/80\n","1719/1719 [==============================] - 0s 96us/step - loss: 1.8622e-04 - mean_absolute_error: 0.0097 - val_loss: 4.5545e-04 - val_mean_absolute_error: 0.0207\n","Epoch 18/80\n","1719/1719 [==============================] - 0s 88us/step - loss: 1.6840e-04 - mean_absolute_error: 0.0090 - val_loss: 4.3901e-05 - val_mean_absolute_error: 0.0054\n","Epoch 19/80\n","1719/1719 [==============================] - 0s 89us/step - loss: 1.5566e-04 - mean_absolute_error: 0.0084 - val_loss: 5.4036e-05 - val_mean_absolute_error: 0.0060\n","Epoch 20/80\n","1719/1719 [==============================] - 0s 97us/step - loss: 1.8316e-04 - mean_absolute_error: 0.0093 - val_loss: 3.6656e-05 - val_mean_absolute_error: 0.0047\n","Epoch 21/80\n","1719/1719 [==============================] - 0s 88us/step - loss: 1.7964e-04 - mean_absolute_error: 0.0097 - val_loss: 4.4971e-05 - val_mean_absolute_error: 0.0055\n","Epoch 22/80\n","1719/1719 [==============================] - 0s 89us/step - loss: 1.8561e-04 - mean_absolute_error: 0.0092 - val_loss: 2.8935e-05 - val_mean_absolute_error: 0.0038\n","Epoch 23/80\n","1719/1719 [==============================] - 0s 92us/step - loss: 1.5764e-04 - mean_absolute_error: 0.0089 - val_loss: 4.4917e-05 - val_mean_absolute_error: 0.0055\n","Epoch 24/80\n","1719/1719 [==============================] - 0s 88us/step - loss: 1.6830e-04 - mean_absolute_error: 0.0092 - val_loss: 3.9811e-05 - val_mean_absolute_error: 0.0050\n","Epoch 25/80\n","1719/1719 [==============================] - 0s 90us/step - loss: 1.4137e-04 - mean_absolute_error: 0.0084 - val_loss: 1.7403e-04 - val_mean_absolute_error: 0.0123\n","Epoch 26/80\n","1719/1719 [==============================] - 0s 93us/step - loss: 1.7480e-04 - mean_absolute_error: 0.0093 - val_loss: 3.4332e-05 - val_mean_absolute_error: 0.0045\n","Epoch 27/80\n","1719/1719 [==============================] - 0s 89us/step - loss: 1.4636e-04 - mean_absolute_error: 0.0089 - val_loss: 1.1033e-04 - val_mean_absolute_error: 0.0095\n","Epoch 28/80\n","1719/1719 [==============================] - 0s 91us/step - loss: 1.5378e-04 - mean_absolute_error: 0.0085 - val_loss: 3.6894e-04 - val_mean_absolute_error: 0.0185\n","Epoch 29/80\n","1719/1719 [==============================] - 0s 88us/step - loss: 1.6258e-04 - mean_absolute_error: 0.0092 - val_loss: 8.0421e-05 - val_mean_absolute_error: 0.0078\n","Epoch 30/80\n","1719/1719 [==============================] - 0s 92us/step - loss: 1.6236e-04 - mean_absolute_error: 0.0093 - val_loss: 9.2499e-05 - val_mean_absolute_error: 0.0086\n","Epoch 31/80\n","1719/1719 [==============================] - 0s 90us/step - loss: 1.3921e-04 - mean_absolute_error: 0.0078 - val_loss: 8.1730e-05 - val_mean_absolute_error: 0.0079\n","Epoch 32/80\n","1719/1719 [==============================] - 0s 90us/step - loss: 1.5921e-04 - mean_absolute_error: 0.0089 - val_loss: 8.2372e-05 - val_mean_absolute_error: 0.0079\n","Epoch 33/80\n","1719/1719 [==============================] - 0s 99us/step - loss: 1.3881e-04 - mean_absolute_error: 0.0086 - val_loss: 4.2319e-05 - val_mean_absolute_error: 0.0051\n","Epoch 34/80\n","1719/1719 [==============================] - 0s 89us/step - loss: 1.7022e-04 - mean_absolute_error: 0.0092 - val_loss: 2.9845e-04 - val_mean_absolute_error: 0.0165\n","Epoch 35/80\n","1719/1719 [==============================] - 0s 87us/step - loss: 1.3844e-04 - mean_absolute_error: 0.0078 - val_loss: 1.2112e-04 - val_mean_absolute_error: 0.0100\n","Epoch 36/80\n","1719/1719 [==============================] - 0s 91us/step - loss: 1.5062e-04 - mean_absolute_error: 0.0087 - val_loss: 3.0683e-05 - val_mean_absolute_error: 0.0040\n","Epoch 37/80\n","1719/1719 [==============================] - 0s 88us/step - loss: 1.4947e-04 - mean_absolute_error: 0.0084 - val_loss: 1.3942e-04 - val_mean_absolute_error: 0.0108\n","Epoch 38/80\n","1719/1719 [==============================] - 0s 94us/step - loss: 1.5446e-04 - mean_absolute_error: 0.0087 - val_loss: 1.0166e-04 - val_mean_absolute_error: 0.0090\n","Epoch 39/80\n","1719/1719 [==============================] - 0s 101us/step - loss: 1.4610e-04 - mean_absolute_error: 0.0083 - val_loss: 8.4441e-05 - val_mean_absolute_error: 0.0081\n","Epoch 40/80\n","1719/1719 [==============================] - 0s 99us/step - loss: 1.3689e-04 - mean_absolute_error: 0.0084 - val_loss: 1.4805e-04 - val_mean_absolute_error: 0.0112\n","Epoch 41/80\n","1719/1719 [==============================] - 0s 93us/step - loss: 1.4496e-04 - mean_absolute_error: 0.0085 - val_loss: 8.2979e-05 - val_mean_absolute_error: 0.0080\n","Epoch 42/80\n","1719/1719 [==============================] - 0s 89us/step - loss: 1.5986e-04 - mean_absolute_error: 0.0089 - val_loss: 1.1679e-04 - val_mean_absolute_error: 0.0098\n","Epoch 43/80\n","1719/1719 [==============================] - 0s 92us/step - loss: 1.3162e-04 - mean_absolute_error: 0.0078 - val_loss: 9.6455e-05 - val_mean_absolute_error: 0.0088\n","Epoch 44/80\n","1719/1719 [==============================] - 0s 93us/step - loss: 1.3917e-04 - mean_absolute_error: 0.0084 - val_loss: 9.2414e-05 - val_mean_absolute_error: 0.0085\n","Epoch 45/80\n","1719/1719 [==============================] - 0s 89us/step - loss: 1.0607e-04 - mean_absolute_error: 0.0072 - val_loss: 6.4306e-05 - val_mean_absolute_error: 0.0069\n","Epoch 46/80\n","1719/1719 [==============================] - 0s 96us/step - loss: 1.5440e-04 - mean_absolute_error: 0.0086 - val_loss: 1.5145e-04 - val_mean_absolute_error: 0.0114\n","Epoch 47/80\n","1719/1719 [==============================] - 0s 90us/step - loss: 1.4677e-04 - mean_absolute_error: 0.0089 - val_loss: 5.4193e-05 - val_mean_absolute_error: 0.0062\n","Epoch 48/80\n","1719/1719 [==============================] - 0s 92us/step - loss: 1.4006e-04 - mean_absolute_error: 0.0084 - val_loss: 1.0706e-04 - val_mean_absolute_error: 0.0093\n","Epoch 49/80\n","1719/1719 [==============================] - 0s 97us/step - loss: 1.3715e-04 - mean_absolute_error: 0.0083 - val_loss: 2.8153e-05 - val_mean_absolute_error: 0.0039\n","Epoch 50/80\n","1719/1719 [==============================] - 0s 88us/step - loss: 1.4154e-04 - mean_absolute_error: 0.0084 - val_loss: 2.8350e-05 - val_mean_absolute_error: 0.0038\n","Epoch 51/80\n","1719/1719 [==============================] - 0s 89us/step - loss: 1.3085e-04 - mean_absolute_error: 0.0081 - val_loss: 9.9383e-05 - val_mean_absolute_error: 0.0089\n","Epoch 52/80\n","1719/1719 [==============================] - 0s 103us/step - loss: 1.6009e-04 - mean_absolute_error: 0.0089 - val_loss: 1.2472e-04 - val_mean_absolute_error: 0.0101\n","Epoch 53/80\n","1719/1719 [==============================] - 0s 91us/step - loss: 1.4058e-04 - mean_absolute_error: 0.0081 - val_loss: 3.4249e-05 - val_mean_absolute_error: 0.0045\n","Epoch 54/80\n","1719/1719 [==============================] - 0s 92us/step - loss: 1.1602e-04 - mean_absolute_error: 0.0074 - val_loss: 7.1764e-05 - val_mean_absolute_error: 0.0074\n","Epoch 55/80\n","1719/1719 [==============================] - 0s 95us/step - loss: 1.3542e-04 - mean_absolute_error: 0.0084 - val_loss: 9.2665e-05 - val_mean_absolute_error: 0.0086\n","Epoch 56/80\n","1719/1719 [==============================] - 0s 95us/step - loss: 1.3735e-04 - mean_absolute_error: 0.0085 - val_loss: 3.5881e-05 - val_mean_absolute_error: 0.0046\n","Epoch 57/80\n","1719/1719 [==============================] - 0s 88us/step - loss: 1.3500e-04 - mean_absolute_error: 0.0083 - val_loss: 5.8830e-05 - val_mean_absolute_error: 0.0065\n","Epoch 58/80\n","1719/1719 [==============================] - 0s 93us/step - loss: 1.2315e-04 - mean_absolute_error: 0.0080 - val_loss: 2.8182e-05 - val_mean_absolute_error: 0.0038\n","Epoch 59/80\n","1719/1719 [==============================] - 0s 98us/step - loss: 1.3483e-04 - mean_absolute_error: 0.0083 - val_loss: 1.3298e-04 - val_mean_absolute_error: 0.0105\n","Epoch 60/80\n","1719/1719 [==============================] - 0s 86us/step - loss: 1.3363e-04 - mean_absolute_error: 0.0084 - val_loss: 9.4996e-05 - val_mean_absolute_error: 0.0087\n","Epoch 61/80\n","1719/1719 [==============================] - 0s 87us/step - loss: 1.0705e-04 - mean_absolute_error: 0.0072 - val_loss: 1.8190e-04 - val_mean_absolute_error: 0.0125\n","Epoch 62/80\n","1719/1719 [==============================] - 0s 95us/step - loss: 1.5339e-04 - mean_absolute_error: 0.0092 - val_loss: 1.9177e-04 - val_mean_absolute_error: 0.0129\n","Epoch 63/80\n","1719/1719 [==============================] - 0s 93us/step - loss: 1.4165e-04 - mean_absolute_error: 0.0085 - val_loss: 9.9931e-05 - val_mean_absolute_error: 0.0090\n","Epoch 64/80\n","1719/1719 [==============================] - 0s 94us/step - loss: 1.2340e-04 - mean_absolute_error: 0.0080 - val_loss: 1.2305e-04 - val_mean_absolute_error: 0.0101\n","Epoch 65/80\n","1719/1719 [==============================] - 0s 122us/step - loss: 1.2870e-04 - mean_absolute_error: 0.0082 - val_loss: 2.0341e-04 - val_mean_absolute_error: 0.0134\n","Epoch 66/80\n","1719/1719 [==============================] - 0s 91us/step - loss: 1.3016e-04 - mean_absolute_error: 0.0084 - val_loss: 4.2077e-05 - val_mean_absolute_error: 0.0052\n","Epoch 67/80\n","1719/1719 [==============================] - 0s 87us/step - loss: 1.2615e-04 - mean_absolute_error: 0.0082 - val_loss: 4.1921e-05 - val_mean_absolute_error: 0.0052\n","Epoch 68/80\n","1719/1719 [==============================] - 0s 85us/step - loss: 1.1662e-04 - mean_absolute_error: 0.0076 - val_loss: 6.1197e-04 - val_mean_absolute_error: 0.0242\n","Epoch 69/80\n","1719/1719 [==============================] - 0s 89us/step - loss: 1.5171e-04 - mean_absolute_error: 0.0089 - val_loss: 5.4299e-05 - val_mean_absolute_error: 0.0062\n","Epoch 70/80\n","1719/1719 [==============================] - 0s 87us/step - loss: 1.3168e-04 - mean_absolute_error: 0.0083 - val_loss: 1.6580e-04 - val_mean_absolute_error: 0.0120\n","Epoch 71/80\n","1719/1719 [==============================] - 0s 101us/step - loss: 1.2382e-04 - mean_absolute_error: 0.0081 - val_loss: 6.4290e-05 - val_mean_absolute_error: 0.0069\n","Epoch 72/80\n","1719/1719 [==============================] - 0s 96us/step - loss: 1.1309e-04 - mean_absolute_error: 0.0076 - val_loss: 3.5509e-05 - val_mean_absolute_error: 0.0046\n","Epoch 73/80\n","1719/1719 [==============================] - 0s 88us/step - loss: 1.4478e-04 - mean_absolute_error: 0.0087 - val_loss: 1.0949e-04 - val_mean_absolute_error: 0.0094\n","Epoch 74/80\n","1719/1719 [==============================] - 0s 92us/step - loss: 1.3794e-04 - mean_absolute_error: 0.0085 - val_loss: 1.5246e-04 - val_mean_absolute_error: 0.0114\n","Epoch 75/80\n","1719/1719 [==============================] - 0s 94us/step - loss: 1.2279e-04 - mean_absolute_error: 0.0080 - val_loss: 1.0124e-04 - val_mean_absolute_error: 0.0090\n","Epoch 76/80\n","1719/1719 [==============================] - 0s 92us/step - loss: 1.3769e-04 - mean_absolute_error: 0.0084 - val_loss: 3.0729e-05 - val_mean_absolute_error: 0.0041\n","Epoch 77/80\n","1719/1719 [==============================] - 0s 101us/step - loss: 1.0772e-04 - mean_absolute_error: 0.0072 - val_loss: 7.3066e-05 - val_mean_absolute_error: 0.0075\n","Epoch 78/80\n","1719/1719 [==============================] - 0s 104us/step - loss: 1.4476e-04 - mean_absolute_error: 0.0091 - val_loss: 5.3765e-05 - val_mean_absolute_error: 0.0062\n","Epoch 79/80\n","1719/1719 [==============================] - 0s 91us/step - loss: 1.1429e-04 - mean_absolute_error: 0.0078 - val_loss: 1.5640e-04 - val_mean_absolute_error: 0.0116\n","Epoch 80/80\n","1719/1719 [==============================] - 0s 91us/step - loss: 1.3166e-04 - mean_absolute_error: 0.0080 - val_loss: 3.4405e-05 - val_mean_absolute_error: 0.0045\n"],"name":"stdout"},{"output_type":"execute_result","data":{"text/plain":["<keras.callbacks.History at 0x7ff300116240>"]},"metadata":{"tags":[]},"execution_count":24}]},{"cell_type":"markdown","metadata":{"id":"PQazcuL1c9QZ","colab_type":"text"},"source":["**is it overtrained? lets run again with 40 epochs**"]},{"cell_type":"code","metadata":{"id":"FvK2f7nXW3rH","colab_type":"code","outputId":"d36535c1-c5a0-4157-c39f-6c48bb3b3575","executionInfo":{"status":"ok","timestamp":1565632084219,"user_tz":240,"elapsed":6632,"user":{"displayName":"Nupur Deshpande","photoUrl":"","userId":"13787757531456301957"}},"colab":{"base_uri":"https://localhost:8080/","height":1000}},"source":["model.fit(\n","    trainX,\n","    trainY,\n","    batch_size=128,\n","    nb_epoch=40,\n","    validation_split=0.05)"],"execution_count":0,"outputs":[{"output_type":"stream","text":["Train on 1719 samples, validate on 91 samples\n","Epoch 1/40\n","1719/1719 [==============================] - 0s 102us/step - loss: 1.4395e-04 - mean_absolute_error: 0.0088 - val_loss: 7.2878e-05 - val_mean_absolute_error: 0.0075\n","Epoch 2/40\n"," 128/1719 [=>............................] - ETA: 0s - loss: 9.1589e-05 - mean_absolute_error: 0.0062"],"name":"stdout"},{"output_type":"stream","text":["/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:6: UserWarning: The `nb_epoch` argument in `fit` has been renamed `epochs`.\n","  \n"],"name":"stderr"},{"output_type":"stream","text":["1719/1719 [==============================] - 0s 95us/step - loss: 1.1378e-04 - mean_absolute_error: 0.0077 - val_loss: 7.1421e-05 - val_mean_absolute_error: 0.0073\n","Epoch 3/40\n","1719/1719 [==============================] - 0s 89us/step - loss: 1.3333e-04 - mean_absolute_error: 0.0084 - val_loss: 3.9828e-05 - val_mean_absolute_error: 0.0050\n","Epoch 4/40\n","1719/1719 [==============================] - 0s 93us/step - loss: 1.2639e-04 - mean_absolute_error: 0.0082 - val_loss: 3.1434e-05 - val_mean_absolute_error: 0.0042\n","Epoch 5/40\n","1719/1719 [==============================] - 0s 93us/step - loss: 1.2105e-04 - mean_absolute_error: 0.0080 - val_loss: 6.5291e-05 - val_mean_absolute_error: 0.0070\n","Epoch 6/40\n","1719/1719 [==============================] - 0s 88us/step - loss: 1.3959e-04 - mean_absolute_error: 0.0085 - val_loss: 4.1084e-04 - val_mean_absolute_error: 0.0196\n","Epoch 7/40\n","1719/1719 [==============================] - 0s 87us/step - loss: 1.2085e-04 - mean_absolute_error: 0.0080 - val_loss: 3.6565e-04 - val_mean_absolute_error: 0.0184\n","Epoch 8/40\n","1719/1719 [==============================] - 0s 93us/step - loss: 1.2532e-04 - mean_absolute_error: 0.0081 - val_loss: 9.1453e-05 - val_mean_absolute_error: 0.0085\n","Epoch 9/40\n","1719/1719 [==============================] - 0s 90us/step - loss: 1.3458e-04 - mean_absolute_error: 0.0084 - val_loss: 3.2829e-05 - val_mean_absolute_error: 0.0044\n","Epoch 10/40\n","1719/1719 [==============================] - 0s 93us/step - loss: 1.1974e-04 - mean_absolute_error: 0.0079 - val_loss: 2.6812e-05 - val_mean_absolute_error: 0.0037\n","Epoch 11/40\n","1719/1719 [==============================] - 0s 94us/step - loss: 1.1889e-04 - mean_absolute_error: 0.0079 - val_loss: 8.9790e-05 - val_mean_absolute_error: 0.0085\n","Epoch 12/40\n","1719/1719 [==============================] - 0s 87us/step - loss: 1.2687e-04 - mean_absolute_error: 0.0083 - val_loss: 1.0581e-04 - val_mean_absolute_error: 0.0092\n","Epoch 13/40\n","1719/1719 [==============================] - 0s 88us/step - loss: 1.0807e-04 - mean_absolute_error: 0.0075 - val_loss: 9.7315e-05 - val_mean_absolute_error: 0.0088\n","Epoch 14/40\n","1719/1719 [==============================] - 0s 91us/step - loss: 1.1440e-04 - mean_absolute_error: 0.0075 - val_loss: 1.2471e-04 - val_mean_absolute_error: 0.0101\n","Epoch 15/40\n","1719/1719 [==============================] - 0s 92us/step - loss: 1.2797e-04 - mean_absolute_error: 0.0084 - val_loss: 3.1272e-04 - val_mean_absolute_error: 0.0169\n","Epoch 16/40\n","1719/1719 [==============================] - 0s 95us/step - loss: 1.2103e-04 - mean_absolute_error: 0.0080 - val_loss: 1.1063e-04 - val_mean_absolute_error: 0.0095\n","Epoch 17/40\n","1719/1719 [==============================] - 0s 107us/step - loss: 1.2381e-04 - mean_absolute_error: 0.0082 - val_loss: 5.4325e-05 - val_mean_absolute_error: 0.0062\n","Epoch 18/40\n","1719/1719 [==============================] - 0s 89us/step - loss: 1.0797e-04 - mean_absolute_error: 0.0073 - val_loss: 6.9013e-05 - val_mean_absolute_error: 0.0071\n","Epoch 19/40\n","1719/1719 [==============================] - 0s 109us/step - loss: 1.2457e-04 - mean_absolute_error: 0.0082 - val_loss: 1.2270e-04 - val_mean_absolute_error: 0.0100\n","Epoch 20/40\n","1719/1719 [==============================] - 0s 89us/step - loss: 1.2081e-04 - mean_absolute_error: 0.0082 - val_loss: 2.5328e-04 - val_mean_absolute_error: 0.0151\n","Epoch 21/40\n","1719/1719 [==============================] - 0s 92us/step - loss: 1.2580e-04 - mean_absolute_error: 0.0083 - val_loss: 4.9837e-05 - val_mean_absolute_error: 0.0059\n","Epoch 22/40\n","1719/1719 [==============================] - 0s 91us/step - loss: 1.2509e-04 - mean_absolute_error: 0.0082 - val_loss: 4.6585e-05 - val_mean_absolute_error: 0.0056\n","Epoch 23/40\n","1719/1719 [==============================] - 0s 91us/step - loss: 1.1245e-04 - mean_absolute_error: 0.0076 - val_loss: 2.7249e-05 - val_mean_absolute_error: 0.0038\n","Epoch 24/40\n","1719/1719 [==============================] - 0s 88us/step - loss: 1.2061e-04 - mean_absolute_error: 0.0080 - val_loss: 2.4126e-04 - val_mean_absolute_error: 0.0146\n","Epoch 25/40\n","1719/1719 [==============================] - 0s 87us/step - loss: 1.1572e-04 - mean_absolute_error: 0.0077 - val_loss: 1.2319e-04 - val_mean_absolute_error: 0.0100\n","Epoch 26/40\n","1719/1719 [==============================] - 0s 88us/step - loss: 1.2772e-04 - mean_absolute_error: 0.0085 - val_loss: 2.6375e-05 - val_mean_absolute_error: 0.0037\n","Epoch 27/40\n","1719/1719 [==============================] - 0s 85us/step - loss: 1.1028e-04 - mean_absolute_error: 0.0079 - val_loss: 1.6489e-04 - val_mean_absolute_error: 0.0119\n","Epoch 28/40\n","1719/1719 [==============================] - 0s 88us/step - loss: 1.1402e-04 - mean_absolute_error: 0.0076 - val_loss: 4.7650e-05 - val_mean_absolute_error: 0.0056\n","Epoch 29/40\n","1719/1719 [==============================] - 0s 85us/step - loss: 1.0562e-04 - mean_absolute_error: 0.0070 - val_loss: 2.3996e-04 - val_mean_absolute_error: 0.0147\n","Epoch 30/40\n","1719/1719 [==============================] - 0s 96us/step - loss: 1.1833e-04 - mean_absolute_error: 0.0080 - val_loss: 4.2833e-05 - val_mean_absolute_error: 0.0052\n","Epoch 31/40\n","1719/1719 [==============================] - 0s 93us/step - loss: 1.1623e-04 - mean_absolute_error: 0.0075 - val_loss: 4.6854e-04 - val_mean_absolute_error: 0.0210\n","Epoch 32/40\n","1719/1719 [==============================] - 0s 92us/step - loss: 1.2763e-04 - mean_absolute_error: 0.0082 - val_loss: 4.7272e-05 - val_mean_absolute_error: 0.0056\n","Epoch 33/40\n","1719/1719 [==============================] - 0s 87us/step - loss: 1.0776e-04 - mean_absolute_error: 0.0073 - val_loss: 1.5516e-04 - val_mean_absolute_error: 0.0115\n","Epoch 34/40\n","1719/1719 [==============================] - 0s 94us/step - loss: 1.0795e-04 - mean_absolute_error: 0.0074 - val_loss: 3.5491e-05 - val_mean_absolute_error: 0.0046\n","Epoch 35/40\n","1719/1719 [==============================] - 0s 89us/step - loss: 1.2660e-04 - mean_absolute_error: 0.0081 - val_loss: 2.8242e-04 - val_mean_absolute_error: 0.0160\n","Epoch 36/40\n","1719/1719 [==============================] - 0s 89us/step - loss: 1.1183e-04 - mean_absolute_error: 0.0074 - val_loss: 9.7360e-05 - val_mean_absolute_error: 0.0087\n","Epoch 37/40\n","1719/1719 [==============================] - 0s 106us/step - loss: 1.1645e-04 - mean_absolute_error: 0.0077 - val_loss: 5.6632e-05 - val_mean_absolute_error: 0.0063\n","Epoch 38/40\n","1719/1719 [==============================] - 0s 89us/step - loss: 1.1465e-04 - mean_absolute_error: 0.0080 - val_loss: 2.0228e-04 - val_mean_absolute_error: 0.0134\n","Epoch 39/40\n","1719/1719 [==============================] - 0s 87us/step - loss: 1.0808e-04 - mean_absolute_error: 0.0075 - val_loss: 2.2450e-04 - val_mean_absolute_error: 0.0141\n","Epoch 40/40\n","1719/1719 [==============================] - 0s 95us/step - loss: 1.1486e-04 - mean_absolute_error: 0.0081 - val_loss: 5.5542e-04 - val_mean_absolute_error: 0.0230\n"],"name":"stdout"},{"output_type":"execute_result","data":{"text/plain":["<keras.callbacks.History at 0x7ff31164e208>"]},"metadata":{"tags":[]},"execution_count":25}]},{"cell_type":"markdown","metadata":{"id":"PNNUkkpIdCfK","colab_type":"text"},"source":["**Final Predictions**"]},{"cell_type":"code","metadata":{"id":"27AkkDUDXLkP","colab_type":"code","outputId":"4601f90c-80d1-40ad-e2c4-ee01ddd6c567","executionInfo":{"status":"ok","timestamp":1565632090995,"user_tz":240,"elapsed":1095,"user":{"displayName":"Nupur Deshpande","photoUrl":"","userId":"13787757531456301957"}},"colab":{"base_uri":"https://localhost:8080/","height":283}},"source":["import matplotlib.pyplot as plt\n","predicted_stateful = model.predict(testX)\n","plt.plot(testY, color = 'red',label = 'Actual Closing Prices')\n","plt.plot(predicted_stateful.flatten(), color = 'blue', label  = 'Predicted closing Prices')\n","plt.xlabel('Days')\n","plt.ylabel('Normalised Prices')\n","plt.legend(loc='best')\n","plt.show()"],"execution_count":0,"outputs":[{"output_type":"display_data","data":{"image/png":"iVBORw0KGgoAAAANSUhEUgAAAYUAAAEKCAYAAAD9xUlFAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAAAASUVORK5CYII=\n","text/plain":["<Figure size 432x288 with 1 Axes>"]},"metadata":{"tags":[]}}]},{"cell_type":"markdown","metadata":{"id":"nc-mIZspkToH","colab_type":"text"},"source":["**So LSTM  works great to predict Closing Prices when we normalize the prices and predict the stock Price movement.**"]},{"cell_type":"code","metadata":{"id":"6Im_DUAVIr3K","colab_type":"code","outputId":"528dd187-3a05-42d0-d097-484700ac6a9d","executionInfo":{"status":"ok","timestamp":1565632147664,"user_tz":240,"elapsed":368,"user":{"displayName":"Nupur Deshpande","photoUrl":"","userId":"13787757531456301957"}},"colab":{"base_uri":"https://localhost:8080/","height":52}},"source":["mae = model.evaluate(trainX,\n","    trainY,\n","    batch_size=128,verbose=1)"],"execution_count":0,"outputs":[{"output_type":"stream","text":["1810/1810 [==============================] - 0s 28us/step\n"],"name":"stdout"},{"output_type":"execute_result","data":{"text/plain":["[0.00017531904821587696, 0.01169782763974772]"]},"metadata":{"tags":[]},"execution_count":27}]},{"cell_type":"markdown","metadata":{"id":"SbLZwCUodL0R","colab_type":"text"},"source":["**Seems like using LSTM to predict stock price movement gives much better prediction resullts!!**\n","\n","**Will work more on this in the next notebook..**"]},{"cell_type":"code","metadata":{"id":"FvtbDsAr8uES","colab_type":"code","colab":{}},"source":[""],"execution_count":0,"outputs":[]}]}