Skip to content

Instantly share code, notes, and snippets.

@analyticsindiamagazine

analyticsindiamagazine/Bitcoin_Price_Predicting_RNN.ipynb

Created August 28, 2019 12:14
Show Gist options

  • Select an option

    Learn more about clone URLs
  • Save analyticsindiamagazine/8aec3310d9346480b4d22532292ae53e to your computer and use it in GitHub Desktop.

Select an option

Learn more about clone URLs
Save analyticsindiamagazine/8aec3310d9346480b4d22532292ae53e to your computer and use it in GitHub Desktop.
Download ZIP
Display the source blob
Display the rendered blob
Raw
Bitcoin_Price_Predicting_RNN.ipynb
{
"nbformat": 4,
"nbformat_minor": 0,
"metadata": {
"kernelspec": {
"name": "python3",
"display_name": "Python 3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.6.8"
},
"colab": {
"name": "Final_Article.ipynb",
"version": "0.3.2",
"provenance": [],
"collapsed_sections": []
},
"accelerator": "GPU"
},
"cells": [
{
"cell_type": "markdown",
"metadata": {
"id": "q6pm25mX4s3k",
"colab_type": "text"
},
"source": [
"# Time Series Forecasting: Predicting Bitcoin Price\n",
"\n",
"\n",
"The cryptocurrency market has seen its rise and fall in the past few years. With a variety of coins being exchanged for real money, it is important to know the trend in the coin price. In this article, we will build a fairly simple LSTM network to predict or forecast the prices of Bitcoin.\n",
"\n",
"## Obtaining Bitcoin Data\n",
"\n",
"There are plenty of open sources available on the internet to extract historical data of Bitcoin prices. The one that I have used below is from Coinmarketcap.\n",
"\n",
"You can view and download the dataset [here](https://coinmarketcap.com/currencies/bitcoin/historical-data/).\n"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "-378lqZkUD36",
"colab_type": "text"
},
"source": [
"## Step 1. Loading and Understanding The Data"
]
},
{
"cell_type": "code",
"metadata": {
"id": "yZICbvdieWsJ",
"colab_type": "code",
"colab": {}
},
"source": [
"#Importing the libraries\n",
"import numpy as np\n",
"import matplotlib.pyplot as plt\n",
"import pandas as pd"
],
"execution_count": 0,
"outputs": []
},
{
"cell_type": "code",
"metadata": {
"id": "C_pr4CzLfA3Y",
"colab_type": "code",
"outputId": "a03415d1-e966-4f7b-9875-70a155f7f141",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 359
}
},
"source": [
"#Importing the data\n",
"complete_data = pd.read_excel('BitCoin_prices.xlsx')\n",
"\n",
"#Displaying the first 10 rows\n",
"complete_data.head(10)\n"
],
"execution_count": 3,
"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>Close</th>\n",
" <th>Volume</th>\n",
" <th>Market Cap</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>0</th>\n",
" <td>Aug 20, 2019</td>\n",
" <td>10916.35</td>\n",
" <td>10947.04</td>\n",
" <td>10618.96</td>\n",
" <td>10763.23</td>\n",
" <td>15053082175</td>\n",
" <td>192530283565</td>\n",
" </tr>\n",
" <tr>\n",
" <th>1</th>\n",
" <td>Aug 19, 2019</td>\n",
" <td>10350.28</td>\n",
" <td>10916.05</td>\n",
" <td>10313.20</td>\n",
" <td>10916.05</td>\n",
" <td>16038264603</td>\n",
" <td>195243306008</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2</th>\n",
" <td>Aug 18, 2019</td>\n",
" <td>10233.01</td>\n",
" <td>10487.07</td>\n",
" <td>10119.09</td>\n",
" <td>10345.81</td>\n",
" <td>12999813869</td>\n",
" <td>185022920955</td>\n",
" </tr>\n",
" <tr>\n",
" <th>3</th>\n",
" <td>Aug 17, 2019</td>\n",
" <td>10358.72</td>\n",
" <td>10452.62</td>\n",
" <td>10086.70</td>\n",
" <td>10231.74</td>\n",
" <td>13778035685</td>\n",
" <td>182966857173</td>\n",
" </tr>\n",
" <tr>\n",
" <th>4</th>\n",
" <td>Aug 16, 2019</td>\n",
" <td>10319.42</td>\n",
" <td>10524.35</td>\n",
" <td>9855.48</td>\n",
" <td>10374.34</td>\n",
" <td>20228207096</td>\n",
" <td>185500055339</td>\n",
" </tr>\n",
" <tr>\n",
" <th>5</th>\n",
" <td>Aug 15, 2019</td>\n",
" <td>10038.42</td>\n",
" <td>10437.41</td>\n",
" <td>9675.32</td>\n",
" <td>10311.55</td>\n",
" <td>22899115082</td>\n",
" <td>184357666577</td>\n",
" </tr>\n",
" <tr>\n",
" <th>6</th>\n",
" <td>Aug 14, 2019</td>\n",
" <td>10889.49</td>\n",
" <td>10889.56</td>\n",
" <td>10028.14</td>\n",
" <td>10051.70</td>\n",
" <td>19990838300</td>\n",
" <td>179692803424</td>\n",
" </tr>\n",
" <tr>\n",
" <th>7</th>\n",
" <td>Aug 13, 2019</td>\n",
" <td>11385.05</td>\n",
" <td>11420.05</td>\n",
" <td>10830.33</td>\n",
" <td>10895.83</td>\n",
" <td>16681503537</td>\n",
" <td>194762696644</td>\n",
" </tr>\n",
" <tr>\n",
" <th>8</th>\n",
" <td>Aug 12, 2019</td>\n",
" <td>11528.19</td>\n",
" <td>11528.19</td>\n",
" <td>11320.95</td>\n",
" <td>11382.62</td>\n",
" <td>13647198229</td>\n",
" <td>203441494985</td>\n",
" </tr>\n",
" <tr>\n",
" <th>9</th>\n",
" <td>Aug 11, 2019</td>\n",
" <td>11349.74</td>\n",
" <td>11523.58</td>\n",
" <td>11248.29</td>\n",
" <td>11523.58</td>\n",
" <td>15774371518</td>\n",
" <td>205941632235</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>"
],
"text/plain": [
" Date Open High ... Close Volume Market Cap\n",
"0 Aug 20, 2019 10916.35 10947.04 ... 10763.23 15053082175 192530283565\n",
"1 Aug 19, 2019 10350.28 10916.05 ... 10916.05 16038264603 195243306008\n",
"2 Aug 18, 2019 10233.01 10487.07 ... 10345.81 12999813869 185022920955\n",
"3 Aug 17, 2019 10358.72 10452.62 ... 10231.74 13778035685 182966857173\n",
"4 Aug 16, 2019 10319.42 10524.35 ... 10374.34 20228207096 185500055339\n",
"5 Aug 15, 2019 10038.42 10437.41 ... 10311.55 22899115082 184357666577\n",
"6 Aug 14, 2019 10889.49 10889.56 ... 10051.70 19990838300 179692803424\n",
"7 Aug 13, 2019 11385.05 11420.05 ... 10895.83 16681503537 194762696644\n",
"8 Aug 12, 2019 11528.19 11528.19 ... 11382.62 13647198229 203441494985\n",
"9 Aug 11, 2019 11349.74 11523.58 ... 11523.58 15774371518 205941632235\n",
"\n",
"[10 rows x 7 columns]"
]
},
"metadata": {
"tags": []
},
"execution_count": 3
}
]
},
{
"cell_type": "code",
"metadata": {
"id": "3MdpXWYbSsLQ",
"colab_type": "code",
"outputId": "e8ef9c97-e2b5-4362-989a-c18b5c5e4a5e",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 238
}
},
"source": [
"#Printing the dataset info\n",
"print(complete_data.info())"
],
"execution_count": 4,
"outputs": [
{
"output_type": "stream",
"text": [
"<class 'pandas.core.frame.DataFrame'>\n",
"RangeIndex: 2306 entries, 0 to 2305\n",
"Data columns (total 7 columns):\n",
"Date 2306 non-null object\n",
"Open 2306 non-null float64\n",
"High 2306 non-null float64\n",
"Low 2306 non-null float64\n",
"Close 2306 non-null float64\n",
"Volume 2306 non-null object\n",
"Market Cap 2306 non-null int64\n",
"dtypes: float64(4), int64(1), object(2)\n",
"memory usage: 126.2+ KB\n",
"None\n"
],
"name": "stdout"
}
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "_8DMG_mfSWG8",
"colab_type": "text"
},
"source": [
"**Key Observations:** \n",
"\n",
"* The dataset has 7 features/columns.\n",
"* The Date column is of type object and has to be converted to datetime.\n",
"* The data is arranged in the decreasing order of date which has to be reversed\n",
"\n",
"\n",
"**Note**\n",
"\n",
"In this project, we will try to predict the closing price of bitcons and will only focus on Date, Open and Close columns in the dataset.\n",
"\n",
"\n"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "V7z7s5s2USZ-",
"colab_type": "text"
},
"source": [
"## Step 2. Creating Training and Test Data"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "rLBuGZLAUoxB",
"colab_type": "text"
},
"source": [
"\n",
"**Important Note:**\n",
"\n",
"The dataset consists of observations in the decreasing order of date. We need to preserve this serial nature of the data and hence we cannot split the data randomly. So we will perform a more straight forward approach to splitting the dataset.\n"
]
},
{
"cell_type": "code",
"metadata": {
"id": "-rXD6xqpprP9",
"colab_type": "code",
"outputId": "c891d2ae-4fc8-4d51-8cd5-5439ee758f8d",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 51
}
},
"source": [
"#Setting The training set ratio\n",
"training_ratio = 80\n",
"\n",
"#Calculating the test set ratio\n",
"test_ratio = 100-training_ratio\n",
"\n",
"#Rounding the training set length to avoid fractions\n",
"training_len = round(len(complete_data)*(training_ratio/100))\n",
"\n",
"#Setting the Test set length\n",
"test_len = round(len(complete_data)-training_len)\n",
"\n",
"#Splitting the data based on the calculated lengths\n",
"dataset_train = complete_data.tail(training_len)\n",
"dataset_test = complete_data.head(test_len)\n",
"\n",
"#Printing the shapes of training and test sets\n",
"\n",
"print(\"Shape Of Training Set :\", dataset_train.shape)\n",
"print(\"Shape Of Test Set :\", dataset_test.shape)\n",
"\n"
],
"execution_count": 5,
"outputs": [
{
"output_type": "stream",
"text": [
"Shape Of Training Set : (1845, 7)\n",
"Shape Of Test Set : (461, 7)\n"
],
"name": "stdout"
}
]
},
{
"cell_type": "code",
"metadata": {
"id": "cLA5Gdh2rLE_",
"colab_type": "code",
"outputId": "5edafcf5-1a93-4cba-ba11-ffe95dbefb7b",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 391
}
},
"source": [
"#Printing the training and test sets\n",
"print(dataset_test.tail(10))\n",
"print(dataset_train.head(10))"
],
"execution_count": 6,
"outputs": [
{
"output_type": "stream",
"text": [
" Date Open High Low Close Volume Market Cap\n",
"451 May 26, 2018 7486.48 7595.16 7349.12 7355.88 4051539968 125469061263\n",
"452 May 25, 2018 7592.30 7659.14 7392.65 7480.14 4867829760 127573690458\n",
"453 May 24, 2018 7561.12 7738.60 7331.14 7587.34 6049220096 129385391552\n",
"454 May 23, 2018 8037.08 8054.66 7507.88 7557.82 6491120128 128868479514\n",
"455 May 22, 2018 8419.87 8423.25 8004.58 8041.78 5137010176 137104106176\n",
"456 May 21, 2018 8522.33 8557.52 8365.12 8418.99 5154990080 143518943480\n",
"457 May 20, 2018 8246.99 8562.41 8205.24 8513.25 5191059968 145109512565\n",
"458 May 19, 2018 8255.73 8372.06 8183.35 8247.18 4712399872 140559162894\n",
"459 May 18, 2018 8091.83 8274.12 7974.82 8250.97 5764190208 140607667610\n",
"460 May 17, 2018 8370.05 8445.54 8054.12 8094.32 5862530048 137923772714\n",
" Date Open High Low Close Volume Market Cap\n",
"461 May 16, 2018 8504.41 8508.43 8175.49 8368.83 6760220160 142587497878\n",
"462 May 15, 2018 8705.19 8836.19 8456.45 8510.38 6705710080 144979744412\n",
"463 May 14, 2018 8713.10 8881.12 8367.97 8716.79 7364149760 148480275422\n",
"464 May 13, 2018 8515.49 8773.55 8395.12 8723.94 5866379776 148587777457\n",
"465 May 12, 2018 8441.44 8664.86 8223.50 8504.89 6821380096 144841040789\n",
"466 May 11, 2018 9052.96 9052.96 8394.46 8441.49 8488520192 143743802092\n",
"467 May 10, 2018 9325.96 9396.04 9040.52 9043.94 6906699776 153988453198\n",
"468 May 09, 2018 9223.73 9374.76 9031.62 9325.18 7226890240 158758858205\n",
"469 May 08, 2018 9380.87 9462.75 9127.77 9234.82 7415869952 157202142973\n",
"470 May 07, 2018 9645.67 9665.85 9231.53 9373.01 7394019840 159538115686\n"
],
"name": "stdout"
}
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "NDLKjGo7Vw-n",
"colab_type": "text"
},
"source": [
"Note:\n",
"\n",
"The test set data should be the continuation of the training set data."
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "0NemsfDIsIL2",
"colab_type": "text"
},
"source": [
"## Step 3. Preprocessing The Training Set"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "voH3ptGAXdmz",
"colab_type": "text"
},
"source": [
"###3.1 Preparing The Data"
]
},
{
"cell_type": "code",
"metadata": {
"id": "mJkWdNRqrmyj",
"colab_type": "code",
"colab": {}
},
"source": [
"#Setting the date colum to date format\n",
"dataset_train['Date'] = pd.to_datetime(dataset_train['Date'])"
],
"execution_count": 0,
"outputs": []
},
{
"cell_type": "code",
"metadata": {
"id": "WdTNe5WynTvF",
"colab_type": "code",
"colab": {}
},
"source": [
"#Setting Date column as Index\n",
"dataset_train.set_index('Date', inplace = True)"
],
"execution_count": 0,
"outputs": []
},
{
"cell_type": "code",
"metadata": {
"id": "UXM48sqVnTzs",
"colab_type": "code",
"colab": {}
},
"source": [
"#Sorting the dataset based on increasing Dates\n",
"dataset_train.sort_values(by='Date',ascending=True, inplace = True)"
],
"execution_count": 0,
"outputs": []
},
{
"cell_type": "code",
"metadata": {
"id": "TBmryV5KnTxQ",
"colab_type": "code",
"outputId": "54bce995-f459-4112-97e1-6eebb9397aa3",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 221
}
},
"source": [
"print(dataset_train.head(10))"
],
"execution_count": 10,
"outputs": [
{
"output_type": "stream",
"text": [
" Open High Low Close Volume Market Cap\n",
"Date \n",
"2013-04-28 135.30 135.98 132.10 134.21 - 1488566728\n",
"2013-04-29 134.44 147.49 134.00 144.54 - 1603768865\n",
"2013-04-30 144.00 146.93 134.05 139.00 - 1542813125\n",
"2013-05-01 139.00 139.89 107.72 116.99 - 1298954594\n",
"2013-05-02 116.38 125.60 92.28 105.21 - 1168517495\n",
"2013-05-03 106.25 108.13 79.10 97.75 - 1085995169\n",
"2013-05-04 98.10 115.00 92.50 112.50 - 1250316563\n",
"2013-05-05 112.90 118.80 107.14 115.91 - 1288693176\n",
"2013-05-06 115.98 124.66 106.64 112.30 - 1249023060\n",
"2013-05-07 112.25 113.44 97.70 111.50 - 1240593600\n"
],
"name": "stdout"
}
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "rWcbDwN5Wm3k",
"colab_type": "text"
},
"source": [
"The training set has now been reorganized by arranging the data in increasing order of dates."
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "QZNkaUWXXmCw",
"colab_type": "text"
},
"source": [
"### 3.2 Scaling and Sequencing "
]
},
{
"cell_type": "code",
"metadata": {
"id": "Gj4GoLU0wEbv",
"colab_type": "code",
"colab": {}
},
"source": [
"#A method to preprocess the data in to sequences and to return x and y \n",
"\n",
"#Initializing the MinMaxScaler object\n",
"from sklearn.preprocessing import MinMaxScaler\n",
"sc = MinMaxScaler(feature_range = (0,1))\n",
"\n",
"def bit_pre_process(raw_data , seq_len, column = 1):\n",
" \n",
" #Select the feature/column \n",
" data = raw_data.iloc[:, column].values\n",
" data = data.reshape(-1, 1)\n",
" \n",
" #Feature Scaling\n",
" data = sc.fit_transform(data)\n",
" \n",
" #Making sequences\n",
" \n",
" X = []\n",
" y = []\n",
"\n",
" for i in range(seq_len, len(data)):\n",
" X.append(data[i-seq_len:i, 0])\n",
" y.append(data[i, 0])\n",
" X, y = np.array(X), np.array(y)\n",
"\n",
" # Reshaping\n",
" X = np.reshape(X, (X.shape[0], X.shape[1], 1))\n",
" \n",
" return X, y"
],
"execution_count": 0,
"outputs": []
},
{
"cell_type": "code",
"metadata": {
"id": "S98-4oBL3Bge",
"colab_type": "code",
"colab": {}
},
"source": [
"\n",
"#Setting the sequence length (Try different values)\n",
"sequence_length = 60\n",
"\n",
"#Choosing the idex of the Close column\n",
"comumn_index= 3\n",
"\n",
"#Preprocessing the training set\n",
"X_train, y_train = bit_pre_process(dataset_train , sequence_length, comumn_index)"
],
"execution_count": 0,
"outputs": []
},
{
"cell_type": "markdown",
"metadata": {
"id": "1I8IdGglY78j",
"colab_type": "text"
},
"source": [
"Sequence length is the number of previous observation to consider in order to predict for a given day. For example, in this case to predict the closing price of Bitcoin at a given date, the algorithm will consider the closing price of Bitcoins for the past 60 days.\n",
"\n"
]
},
{
"cell_type": "code",
"metadata": {
"id": "O96Cqx49eWsf",
"colab_type": "code",
"outputId": "d0ac5f8b-9efa-46f3-a6fa-a1005d34d627",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 34
}
},
"source": [
"X_train.shape"
],
"execution_count": 13,
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"(1785, 60, 1)"
]
},
"metadata": {
"tags": []
},
"execution_count": 13
}
]
},
{
"cell_type": "code",
"metadata": {
"id": "qloDj6rQeWsh",
"colab_type": "code",
"outputId": "7aafec29-2ddb-4ef1-e99f-c23717a954fe",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 34
}
},
"source": [
"y_train.shape"
],
"execution_count": 14,
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"(1785,)"
]
},
"metadata": {
"tags": []
},
"execution_count": 14
}
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "OxzZEFG5aSQJ",
"colab_type": "text"
},
"source": [
"The independent variable set(x_train) will now consist of 1785 observations (Original length of training set - sequence length (1845 - 60)). Each observation in x_train is a sequence of 60 closing prices i,e. the first row in x_train will be an array of the first 60 observations from the training set and the corresponding y_train will be the closing price of the 61st observation.\n",
"\n",
"\n",
"For a sequence length of 60, x_train and y_train will look like whats shown below:\n",
"\n",
"\n",
"![alt text](https://www.analyticsindiamag.com/wp-content/uploads/2019/08/true_x_train_y_train.png)"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "g--bVpuU3jZp",
"colab_type": "text"
},
"source": [
"## Step 4. Building a Recurrent Neural Network"
]
},
{
"cell_type": "code",
"metadata": {
"id": "MdgKyMuleWsl",
"colab_type": "code",
"colab": {}
},
"source": [
"#Importing the Keras libraries and packages\n",
"from keras.models import Sequential\n",
"from keras.layers import Dense\n",
"from keras.layers import LSTM, CuDNNLSTM\n",
"from keras.layers import Dropout\n",
"\n",
"#Initialising the RNN\n",
"regressor = Sequential()\n",
"\n",
"#Adding the first LSTM layer and some Dropout regularisation\n",
"regressor.add(CuDNNLSTM(units = 50, return_sequences = True, input_shape = (X_train.shape[1], 1)))\n",
"regressor.add(Dropout(0.2))\n",
"\n",
"#Adding a second LSTM layer and some Dropout regularisation\n",
"regressor.add(CuDNNLSTM(units = 50, return_sequences = True))\n",
"regressor.add(Dropout(0.2))\n",
"\n",
"#Adding a third LSTM layer and some Dropout regularisation\n",
"regressor.add(CuDNNLSTM(units = 50, return_sequences = True))\n",
"regressor.add(Dropout(0.2))\n",
"\n",
"#Adding a fourth LSTM layer and some Dropout regularisation\n",
"regressor.add(CuDNNLSTM(units = 50))\n",
"regressor.add(Dropout(0.2))\n",
"\n",
"#Adding the output layer\n",
"regressor.add(Dense(units = 1))\n"
],
"execution_count": 0,
"outputs": []
},
{
"cell_type": "markdown",
"metadata": {
"id": "AlhRxJsF2FWA",
"colab_type": "text"
},
"source": [
"**Important Note:**\n",
"\n",
"CuDNNLSTM requires a GPU to execute. If there is no GPU, use LSTM instead."
]
},
{
"cell_type": "code",
"metadata": {
"id": "1gMOj97Y0bd0",
"colab_type": "code",
"colab": {}
},
"source": [
"#Compiling the RNN\n",
"regressor.compile(optimizer = 'adam', loss = 'mean_squared_error', metrics = ['mse', 'mae'])\n"
],
"execution_count": 0,
"outputs": []
},
{
"cell_type": "code",
"metadata": {
"id": "LcsjQspS0dDP",
"colab_type": "code",
"outputId": "b1da7c69-81bb-4388-eb92-5a65b8c63b1e",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 1000
}
},
"source": [
"#Fitting the RNN to the Training set and training the RNN\n",
"regressor.fit(X_train, y_train, epochs = 50, batch_size = 50)"
],
"execution_count": 19,
"outputs": [
{
"output_type": "stream",
"text": [
"Epoch 1/50\n",
"1785/1785 [==============================] - 3s 2ms/step - loss: 0.0061 - mean_squared_error: 0.0061 - mean_absolute_error: 0.0424\n",
"Epoch 2/50\n",
"1785/1785 [==============================] - 2s 1ms/step - loss: 0.0024 - mean_squared_error: 0.0024 - mean_absolute_error: 0.0224\n",
"Epoch 3/50\n",
"1785/1785 [==============================] - 2s 1ms/step - loss: 0.0020 - mean_squared_error: 0.0020 - mean_absolute_error: 0.0223\n",
"Epoch 4/50\n",
"1785/1785 [==============================] - 2s 1ms/step - loss: 0.0020 - mean_squared_error: 0.0020 - mean_absolute_error: 0.0212\n",
"Epoch 5/50\n",
"1785/1785 [==============================] - 2s 1ms/step - loss: 0.0025 - mean_squared_error: 0.0025 - mean_absolute_error: 0.0284\n",
"Epoch 6/50\n",
"1785/1785 [==============================] - 2s 1ms/step - loss: 0.0015 - mean_squared_error: 0.0015 - mean_absolute_error: 0.0181\n",
"Epoch 7/50\n",
"1785/1785 [==============================] - 2s 1ms/step - loss: 0.0015 - mean_squared_error: 0.0015 - mean_absolute_error: 0.0180\n",
"Epoch 8/50\n",
"1785/1785 [==============================] - 2s 1ms/step - loss: 0.0016 - mean_squared_error: 0.0016 - mean_absolute_error: 0.0189\n",
"Epoch 9/50\n",
"1785/1785 [==============================] - 2s 1ms/step - loss: 0.0017 - mean_squared_error: 0.0017 - mean_absolute_error: 0.0203\n",
"Epoch 10/50\n",
"1785/1785 [==============================] - 2s 1ms/step - loss: 0.0014 - mean_squared_error: 0.0014 - mean_absolute_error: 0.0180\n",
"Epoch 11/50\n",
"1785/1785 [==============================] - 2s 1ms/step - loss: 0.0015 - mean_squared_error: 0.0015 - mean_absolute_error: 0.0172\n",
"Epoch 12/50\n",
"1785/1785 [==============================] - 2s 1ms/step - loss: 0.0014 - mean_squared_error: 0.0014 - mean_absolute_error: 0.0162\n",
"Epoch 13/50\n",
"1785/1785 [==============================] - 2s 1ms/step - loss: 0.0011 - mean_squared_error: 0.0011 - mean_absolute_error: 0.0150\n",
"Epoch 14/50\n",
"1785/1785 [==============================] - 2s 1ms/step - loss: 0.0014 - mean_squared_error: 0.0014 - mean_absolute_error: 0.0187\n",
"Epoch 15/50\n",
"1785/1785 [==============================] - 2s 1ms/step - loss: 0.0012 - mean_squared_error: 0.0012 - mean_absolute_error: 0.0168\n",
"Epoch 16/50\n",
"1785/1785 [==============================] - 2s 1ms/step - loss: 0.0010 - mean_squared_error: 0.0010 - mean_absolute_error: 0.0152\n",
"Epoch 17/50\n",
"1785/1785 [==============================] - 2s 1ms/step - loss: 0.0012 - mean_squared_error: 0.0012 - mean_absolute_error: 0.0144\n",
"Epoch 18/50\n",
"1785/1785 [==============================] - 2s 1ms/step - loss: 0.0014 - mean_squared_error: 0.0014 - mean_absolute_error: 0.0183\n",
"Epoch 19/50\n",
"1785/1785 [==============================] - 2s 1ms/step - loss: 0.0011 - mean_squared_error: 0.0011 - mean_absolute_error: 0.0182\n",
"Epoch 20/50\n",
"1785/1785 [==============================] - 2s 1ms/step - loss: 0.0010 - mean_squared_error: 0.0010 - mean_absolute_error: 0.0159\n",
"Epoch 21/50\n",
"1785/1785 [==============================] - 2s 1ms/step - loss: 0.0010 - mean_squared_error: 0.0010 - mean_absolute_error: 0.0162\n",
"Epoch 22/50\n",
"1785/1785 [==============================] - 2s 1ms/step - loss: 0.0011 - mean_squared_error: 0.0011 - mean_absolute_error: 0.0159\n",
"Epoch 23/50\n",
"1785/1785 [==============================] - 2s 1ms/step - loss: 0.0010 - mean_squared_error: 0.0010 - mean_absolute_error: 0.0146\n",
"Epoch 24/50\n",
"1785/1785 [==============================] - 2s 1ms/step - loss: 0.0011 - mean_squared_error: 0.0011 - mean_absolute_error: 0.0167\n",
"Epoch 25/50\n",
"1785/1785 [==============================] - 2s 1ms/step - loss: 9.9813e-04 - mean_squared_error: 9.9813e-04 - mean_absolute_error: 0.0154\n",
"Epoch 26/50\n",
"1785/1785 [==============================] - 2s 1ms/step - loss: 0.0011 - mean_squared_error: 0.0011 - mean_absolute_error: 0.0158\n",
"Epoch 27/50\n",
"1785/1785 [==============================] - 2s 1ms/step - loss: 0.0010 - mean_squared_error: 0.0010 - mean_absolute_error: 0.0156\n",
"Epoch 28/50\n",
"1785/1785 [==============================] - 2s 1ms/step - loss: 0.0011 - mean_squared_error: 0.0011 - mean_absolute_error: 0.0153\n",
"Epoch 29/50\n",
"1785/1785 [==============================] - 2s 1ms/step - loss: 9.3647e-04 - mean_squared_error: 9.3647e-04 - mean_absolute_error: 0.0150\n",
"Epoch 30/50\n",
"1785/1785 [==============================] - 2s 1ms/step - loss: 0.0010 - mean_squared_error: 0.0010 - mean_absolute_error: 0.0161\n",
"Epoch 31/50\n",
"1785/1785 [==============================] - 2s 1ms/step - loss: 0.0012 - mean_squared_error: 0.0012 - mean_absolute_error: 0.0170\n",
"Epoch 32/50\n",
"1785/1785 [==============================] - 2s 1ms/step - loss: 0.0010 - mean_squared_error: 0.0010 - mean_absolute_error: 0.0168\n",
"Epoch 33/50\n",
"1785/1785 [==============================] - 2s 1ms/step - loss: 9.4160e-04 - mean_squared_error: 9.4160e-04 - mean_absolute_error: 0.0190\n",
"Epoch 34/50\n",
"1785/1785 [==============================] - 2s 1ms/step - loss: 0.0010 - mean_squared_error: 0.0010 - mean_absolute_error: 0.0150\n",
"Epoch 35/50\n",
"1785/1785 [==============================] - 2s 1ms/step - loss: 0.0010 - mean_squared_error: 0.0010 - mean_absolute_error: 0.0150 \n",
"Epoch 36/50\n",
"1785/1785 [==============================] - 2s 1ms/step - loss: 9.3799e-04 - mean_squared_error: 9.3799e-04 - mean_absolute_error: 0.0143\n",
"Epoch 37/50\n",
"1785/1785 [==============================] - 2s 1ms/step - loss: 8.2048e-04 - mean_squared_error: 8.2048e-04 - mean_absolute_error: 0.0137\n",
"Epoch 38/50\n",
"1785/1785 [==============================] - 2s 1ms/step - loss: 9.8252e-04 - mean_squared_error: 9.8252e-04 - mean_absolute_error: 0.0159\n",
"Epoch 39/50\n",
"1785/1785 [==============================] - 2s 1ms/step - loss: 7.9782e-04 - mean_squared_error: 7.9782e-04 - mean_absolute_error: 0.0132\n",
"Epoch 40/50\n",
"1785/1785 [==============================] - 2s 1ms/step - loss: 9.5424e-04 - mean_squared_error: 9.5424e-04 - mean_absolute_error: 0.0147\n",
"Epoch 41/50\n",
"1785/1785 [==============================] - 2s 1ms/step - loss: 7.0718e-04 - mean_squared_error: 7.0718e-04 - mean_absolute_error: 0.0132\n",
"Epoch 42/50\n",
"1785/1785 [==============================] - 2s 1ms/step - loss: 6.8666e-04 - mean_squared_error: 6.8666e-04 - mean_absolute_error: 0.0128\n",
"Epoch 43/50\n",
"1785/1785 [==============================] - 2s 1ms/step - loss: 8.9949e-04 - mean_squared_error: 8.9949e-04 - mean_absolute_error: 0.0137\n",
"Epoch 44/50\n",
"1785/1785 [==============================] - 2s 1ms/step - loss: 6.5313e-04 - mean_squared_error: 6.5313e-04 - mean_absolute_error: 0.0137\n",
"Epoch 45/50\n",
"1785/1785 [==============================] - 2s 1ms/step - loss: 8.9759e-04 - mean_squared_error: 8.9759e-04 - mean_absolute_error: 0.0154\n",
"Epoch 46/50\n",
"1785/1785 [==============================] - 2s 1ms/step - loss: 8.0164e-04 - mean_squared_error: 8.0164e-04 - mean_absolute_error: 0.0135\n",
"Epoch 47/50\n",
"1785/1785 [==============================] - 2s 1ms/step - loss: 7.0854e-04 - mean_squared_error: 7.0854e-04 - mean_absolute_error: 0.0127\n",
"Epoch 48/50\n",
"1785/1785 [==============================] - 2s 1ms/step - loss: 7.1875e-04 - mean_squared_error: 7.1875e-04 - mean_absolute_error: 0.0137\n",
"Epoch 49/50\n",
"1785/1785 [==============================] - 2s 1ms/step - loss: 7.8526e-04 - mean_squared_error: 7.8526e-04 - mean_absolute_error: 0.0132\n",
"Epoch 50/50\n",
"1785/1785 [==============================] - 2s 1ms/step - loss: 7.0160e-04 - mean_squared_error: 7.0160e-04 - mean_absolute_error: 0.0123\n"
],
"name": "stdout"
},
{
"output_type": "execute_result",
"data": {
"text/plain": [
"<keras.callbacks.History at 0x7f9df0190908>"
]
},
"metadata": {
"tags": []
},
"execution_count": 19
}
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "HHS1Tb7-0lqe",
"colab_type": "text"
},
"source": [
"## Step 5. Predicting For The Test Set"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "k6GJXRg-3rlY",
"colab_type": "text"
},
"source": [
"### 5.1 Preparing Test Data"
]
},
{
"cell_type": "code",
"metadata": {
"id": "jZF6zlZB4FO8",
"colab_type": "code",
"outputId": "0e90acfb-8e28-43fe-bd8e-05b484117ac1",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 204
}
},
"source": [
"#Displaying the dataset\n",
"dataset_test.head()"
],
"execution_count": 20,
"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>Close</th>\n",
" <th>Volume</th>\n",
" <th>Market Cap</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>0</th>\n",
" <td>Aug 20, 2019</td>\n",
" <td>10916.35</td>\n",
" <td>10947.04</td>\n",
" <td>10618.96</td>\n",
" <td>10763.23</td>\n",
" <td>15053082175</td>\n",
" <td>192530283565</td>\n",
" </tr>\n",
" <tr>\n",
" <th>1</th>\n",
" <td>Aug 19, 2019</td>\n",
" <td>10350.28</td>\n",
" <td>10916.05</td>\n",
" <td>10313.20</td>\n",
" <td>10916.05</td>\n",
" <td>16038264603</td>\n",
" <td>195243306008</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2</th>\n",
" <td>Aug 18, 2019</td>\n",
" <td>10233.01</td>\n",
" <td>10487.07</td>\n",
" <td>10119.09</td>\n",
" <td>10345.81</td>\n",
" <td>12999813869</td>\n",
" <td>185022920955</td>\n",
" </tr>\n",
" <tr>\n",
" <th>3</th>\n",
" <td>Aug 17, 2019</td>\n",
" <td>10358.72</td>\n",
" <td>10452.62</td>\n",
" <td>10086.70</td>\n",
" <td>10231.74</td>\n",
" <td>13778035685</td>\n",
" <td>182966857173</td>\n",
" </tr>\n",
" <tr>\n",
" <th>4</th>\n",
" <td>Aug 16, 2019</td>\n",
" <td>10319.42</td>\n",
" <td>10524.35</td>\n",
" <td>9855.48</td>\n",
" <td>10374.34</td>\n",
" <td>20228207096</td>\n",
" <td>185500055339</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>"
],
"text/plain": [
" Date Open High ... Close Volume Market Cap\n",
"0 Aug 20, 2019 10916.35 10947.04 ... 10763.23 15053082175 192530283565\n",
"1 Aug 19, 2019 10350.28 10916.05 ... 10916.05 16038264603 195243306008\n",
"2 Aug 18, 2019 10233.01 10487.07 ... 10345.81 12999813869 185022920955\n",
"3 Aug 17, 2019 10358.72 10452.62 ... 10231.74 13778035685 182966857173\n",
"4 Aug 16, 2019 10319.42 10524.35 ... 10374.34 20228207096 185500055339\n",
"\n",
"[5 rows x 7 columns]"
]
},
"metadata": {
"tags": []
},
"execution_count": 20
}
]
},
{
"cell_type": "code",
"metadata": {
"colab_type": "code",
"id": "W0yV-yvHufeC",
"colab": {}
},
"source": [
"#Converting the Date colum values from object to datetime format\n",
"dataset_test['Date'] = pd.to_datetime(dataset_test['Date'])"
],
"execution_count": 0,
"outputs": []
},
{
"cell_type": "code",
"metadata": {
"colab_type": "code",
"id": "YGOpdYdHufeH",
"colab": {}
},
"source": [
"#Setting the Date column as Index to the dataset\n",
"dataset_test.set_index('Date', inplace = True)"
],
"execution_count": 0,
"outputs": []
},
{
"cell_type": "code",
"metadata": {
"colab_type": "code",
"id": "ijeQvGrDufeP",
"colab": {}
},
"source": [
"#Sorting the data based on Increasing value of date\n",
"dataset_test.sort_values(by='Date',ascending=True, inplace = True)"
],
"execution_count": 0,
"outputs": []
},
{
"cell_type": "code",
"metadata": {
"colab_type": "code",
"outputId": "66449867-78e3-45b3-caa4-3f19cdac9b1f",
"id": "JUtSe1THufeK",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 235
}
},
"source": [
"#Displaying the prepared dataset\n",
"dataset_test.head()"
],
"execution_count": 25,
"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>Open</th>\n",
" <th>High</th>\n",
" <th>Low</th>\n",
" <th>Close</th>\n",
" <th>Volume</th>\n",
" <th>Market Cap</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>2018-05-17</th>\n",
" <td>8370.05</td>\n",
" <td>8445.54</td>\n",
" <td>8054.12</td>\n",
" <td>8094.32</td>\n",
" <td>5862530048</td>\n",
" <td>137923772714</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2018-05-18</th>\n",
" <td>8091.83</td>\n",
" <td>8274.12</td>\n",
" <td>7974.82</td>\n",
" <td>8250.97</td>\n",
" <td>5764190208</td>\n",
" <td>140607667610</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2018-05-19</th>\n",
" <td>8255.73</td>\n",
" <td>8372.06</td>\n",
" <td>8183.35</td>\n",
" <td>8247.18</td>\n",
" <td>4712399872</td>\n",
" <td>140559162894</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2018-05-20</th>\n",
" <td>8246.99</td>\n",
" <td>8562.41</td>\n",
" <td>8205.24</td>\n",
" <td>8513.25</td>\n",
" <td>5191059968</td>\n",
" <td>145109512565</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2018-05-21</th>\n",
" <td>8522.33</td>\n",
" <td>8557.52</td>\n",
" <td>8365.12</td>\n",
" <td>8418.99</td>\n",
" <td>5154990080</td>\n",
" <td>143518943480</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>"
],
"text/plain": [
" Open High Low Close Volume Market Cap\n",
"Date \n",
"2018-05-17 8370.05 8445.54 8054.12 8094.32 5862530048 137923772714\n",
"2018-05-18 8091.83 8274.12 7974.82 8250.97 5764190208 140607667610\n",
"2018-05-19 8255.73 8372.06 8183.35 8247.18 4712399872 140559162894\n",
"2018-05-20 8246.99 8562.41 8205.24 8513.25 5191059968 145109512565\n",
"2018-05-21 8522.33 8557.52 8365.12 8418.99 5154990080 143518943480"
]
},
"metadata": {
"tags": []
},
"execution_count": 25
}
]
},
{
"cell_type": "code",
"metadata": {
"id": "5qJoK5cLtHqb",
"colab_type": "code",
"outputId": "3a49b86b-1040-4949-989a-e28a44080407",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 34
}
},
"source": [
"dataset_test.shape"
],
"execution_count": 24,
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"(461, 6)"
]
},
"metadata": {
"tags": []
},
"execution_count": 24
}
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "NDNocNP2rqyk",
"colab_type": "text"
},
"source": [
"Now we have arranged the test data in increasing order of dates. However the test data is still not prepared for forecasting. In order to predict the the closing price of the first day in the test set we will require the closing prices of the previous 60 days. So we will attach the closing prices of the previous 60 days from the training set to the test set."
]
},
{
"cell_type": "code",
"metadata": {
"id": "OWaDf6ee75I4",
"colab_type": "code",
"colab": {}
},
"source": [
"#Adding the previous 60 days cosing price to the test data\n",
"test_set = pd.concat((dataset_train.tail(sequence_length), dataset_test), axis = 0)"
],
"execution_count": 0,
"outputs": []
},
{
"cell_type": "code",
"metadata": {
"id": "Z5lQy0UNtvGj",
"colab_type": "code",
"outputId": "69a69caa-3ace-4110-ba8d-084bd9a05cc8",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 390
}
},
"source": [
"test_set.head(10)"
],
"execution_count": 27,
"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>Open</th>\n",
" <th>High</th>\n",
" <th>Low</th>\n",
" <th>Close</th>\n",
" <th>Volume</th>\n",
" <th>Market Cap</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>2018-03-18</th>\n",
" <td>7890.52</td>\n",
" <td>8245.51</td>\n",
" <td>7397.99</td>\n",
" <td>8223.68</td>\n",
" <td>6639190016</td>\n",
" <td>139201713268</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2018-03-19</th>\n",
" <td>8344.12</td>\n",
" <td>8675.87</td>\n",
" <td>8182.40</td>\n",
" <td>8630.65</td>\n",
" <td>6729110016</td>\n",
" <td>146107514353</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2018-03-20</th>\n",
" <td>8619.67</td>\n",
" <td>9051.02</td>\n",
" <td>8389.89</td>\n",
" <td>8913.47</td>\n",
" <td>6361789952</td>\n",
" <td>150909503835</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2018-03-21</th>\n",
" <td>8937.48</td>\n",
" <td>9177.37</td>\n",
" <td>8846.33</td>\n",
" <td>8929.28</td>\n",
" <td>6043129856</td>\n",
" <td>151193917440</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2018-03-22</th>\n",
" <td>8939.44</td>\n",
" <td>9100.71</td>\n",
" <td>8564.90</td>\n",
" <td>8728.47</td>\n",
" <td>5530390016</td>\n",
" <td>147809220250</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2018-03-23</th>\n",
" <td>8736.25</td>\n",
" <td>8879.62</td>\n",
" <td>8360.62</td>\n",
" <td>8879.62</td>\n",
" <td>5954120192</td>\n",
" <td>150383574951</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2018-03-24</th>\n",
" <td>8901.95</td>\n",
" <td>8996.18</td>\n",
" <td>8665.70</td>\n",
" <td>8668.12</td>\n",
" <td>5664600064</td>\n",
" <td>146818882936</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2018-03-25</th>\n",
" <td>8612.81</td>\n",
" <td>8682.01</td>\n",
" <td>8449.10</td>\n",
" <td>8495.78</td>\n",
" <td>4569880064</td>\n",
" <td>143914265310</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2018-03-26</th>\n",
" <td>8498.47</td>\n",
" <td>8530.08</td>\n",
" <td>7921.43</td>\n",
" <td>8209.40</td>\n",
" <td>5921039872</td>\n",
" <td>139078211968</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2018-03-27</th>\n",
" <td>8200.00</td>\n",
" <td>8232.78</td>\n",
" <td>7797.28</td>\n",
" <td>7833.04</td>\n",
" <td>5378250240</td>\n",
" <td>132717053150</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>"
],
"text/plain": [
" Open High Low Close Volume Market Cap\n",
"Date \n",
"2018-03-18 7890.52 8245.51 7397.99 8223.68 6639190016 139201713268\n",
"2018-03-19 8344.12 8675.87 8182.40 8630.65 6729110016 146107514353\n",
"2018-03-20 8619.67 9051.02 8389.89 8913.47 6361789952 150909503835\n",
"2018-03-21 8937.48 9177.37 8846.33 8929.28 6043129856 151193917440\n",
"2018-03-22 8939.44 9100.71 8564.90 8728.47 5530390016 147809220250\n",
"2018-03-23 8736.25 8879.62 8360.62 8879.62 5954120192 150383574951\n",
"2018-03-24 8901.95 8996.18 8665.70 8668.12 5664600064 146818882936\n",
"2018-03-25 8612.81 8682.01 8449.10 8495.78 4569880064 143914265310\n",
"2018-03-26 8498.47 8530.08 7921.43 8209.40 5921039872 139078211968\n",
"2018-03-27 8200.00 8232.78 7797.28 7833.04 5378250240 132717053150"
]
},
"metadata": {
"tags": []
},
"execution_count": 27
}
]
},
{
"cell_type": "code",
"metadata": {
"id": "xBQdpSGw8l4i",
"colab_type": "code",
"outputId": "1af5574a-eaac-435b-a072-969249cd6960",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 34
}
},
"source": [
"test_set.shape"
],
"execution_count": 28,
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"(521, 6)"
]
},
"metadata": {
"tags": []
},
"execution_count": 28
}
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "QjQSuzf0wxFv",
"colab_type": "text"
},
"source": [
"Now test set has 521 rows (60 + 461) and is ready to be preprocessed."
]
},
{
"cell_type": "code",
"metadata": {
"id": "kHufio7c8q1h",
"colab_type": "code",
"colab": {}
},
"source": [
"#Preprocessing the test data\n",
"x_test, y_true = bit_pre_process(test_set , sequence_length, comumn_index)"
],
"execution_count": 0,
"outputs": []
},
{
"cell_type": "code",
"metadata": {
"id": "6X1eIS6t_X8d",
"colab_type": "code",
"outputId": "f42a8876-4d95-42a8-f972-8208127d28ae",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 34
}
},
"source": [
"x_test.shape"
],
"execution_count": 30,
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"(461, 60, 1)"
]
},
"metadata": {
"tags": []
},
"execution_count": 30
}
]
},
{
"cell_type": "code",
"metadata": {
"id": "loznRTDUU96i",
"colab_type": "code",
"outputId": "4798daf2-333b-4174-d817-edaff193b076",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 34
}
},
"source": [
"y_true.shape"
],
"execution_count": 31,
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"(461,)"
]
},
"metadata": {
"tags": []
},
"execution_count": 31
}
]
},
{
"cell_type": "code",
"metadata": {
"id": "q522kNAKVcoC",
"colab_type": "code",
"colab": {}
},
"source": [
"#Predicting the closing price of the test data\n",
"predicted_stock_price = regressor.predict(x_test)"
],
"execution_count": 0,
"outputs": []
},
{
"cell_type": "code",
"metadata": {
"id": "ZU4fwVzY1ez2",
"colab_type": "code",
"colab": {}
},
"source": [
"#Inversing the scaled values \n",
"predicted_stock_price = sc.inverse_transform(predicted_stock_price)"
],
"execution_count": 0,
"outputs": []
},
{
"cell_type": "code",
"metadata": {
"id": "c672JItLVqf5",
"colab_type": "code",
"colab": {}
},
"source": [
"#Inversing the scaled actual prices from test data\n",
"real_stock_price = sc.inverse_transform(y_true.reshape(-1, 1))"
],
"execution_count": 0,
"outputs": []
},
{
"cell_type": "markdown",
"metadata": {
"id": "rLPRoe5tt8Xf",
"colab_type": "text"
},
"source": [
"## Step 6. Visualizing The Prediction"
]
},
{
"cell_type": "code",
"metadata": {
"id": "JlCyB8w4WDdJ",
"colab_type": "code",
"colab": {}
},
"source": [
"\n",
"# A Method to plot predicted pirce vs Actual price\n",
"def plot_predictions(real_price, predicted_price, title, x_label, y_label):\n",
" plt.plot(real_price, color = 'green', label = 'Real Stock Price')\n",
" plt.plot(predicted_price, color = 'red', label = 'Predicted Stock Price')\n",
" plt.title(title)\n",
" plt.xlabel('Time')\n",
" plt.ylabel('Google Stock Price')\n",
" plt.legend()\n",
" plt.show()"
],
"execution_count": 0,
"outputs": []
},
{
"cell_type": "code",
"metadata": {
"id": "vqRXlx85aL2m",
"colab_type": "code",
"outputId": "053682bb-f0dd-48fe-f65a-538e909d16d4",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 295
}
},
"source": [
"#Plotting real_stock_pric vs predicted_stock_price\n",
"plot_predictions(real_stock_price, predicted_stock_price, \"Bitcoin Closing Price Prediction\", \"Time\", \"Closing Price\")"
],
"execution_count": 36,
"outputs": [
{
"output_type": "display_data",
"data": {
"image/png": "iVBORw0KGgoAAAANSUhEUgAAAZUAAAEWCAYAAACufwpNAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\ny2hQUUop5TIaVJRSSrmMBhWllFIu8/8B7iA0ZuzEBGgAAAAASUVORK5CYII=\n",
"text/plain": [
"<Figure size 432x288 with 1 Axes>"
]
},
"metadata": {
"tags": []
}
}
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "QkRH8hk2uZah",
"colab_type": "text"
},
"source": [
"## Step 7. Comparing the Predicted and Actual prices"
]
},
{
"cell_type": "code",
"metadata": {
"id": "_okB5Ew714KD",
"colab_type": "code",
"colab": {}
},
"source": [
"dataset_test['Predicted_Close'] = predicted_stock_price"
],
"execution_count": 0,
"outputs": []
},
{
"cell_type": "code",
"metadata": {
"id": "OrIHqzf62KZn",
"colab_type": "code",
"colab": {}
},
"source": [
"compare = dataset_test[['Open','Close','Predicted_Close' ]]"
],
"execution_count": 0,
"outputs": []
},
{
"cell_type": "code",
"metadata": {
"id": "sq-lb2YZ2kmU",
"colab_type": "code",
"outputId": "2994380f-3ad2-43f1-fc9e-b0a6c92c861d",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 700
}
},
"source": [
"compare.head(20)"
],
"execution_count": 39,
"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>Open</th>\n",
" <th>Close</th>\n",
" <th>Predicted_Close</th>\n",
" </tr>\n",
" <tr>\n",
" <th>Date</th>\n",
" <th></th>\n",
" <th></th>\n",
" <th></th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>2018-05-17</th>\n",
" <td>8370.05</td>\n",
" <td>8094.32</td>\n",
" <td>8908.131836</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2018-05-18</th>\n",
" <td>8091.83</td>\n",
" <td>8250.97</td>\n",
" <td>8823.910156</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2018-05-19</th>\n",
" <td>8255.73</td>\n",
" <td>8247.18</td>\n",
" <td>8739.717773</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2018-05-20</th>\n",
" <td>8246.99</td>\n",
" <td>8513.25</td>\n",
" <td>8670.551758</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2018-05-21</th>\n",
" <td>8522.33</td>\n",
" <td>8418.99</td>\n",
" <td>8650.153320</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2018-05-22</th>\n",
" <td>8419.87</td>\n",
" <td>8041.78</td>\n",
" <td>8668.397461</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2018-05-23</th>\n",
" <td>8037.08</td>\n",
" <td>7557.82</td>\n",
" <td>8668.617188</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2018-05-24</th>\n",
" <td>7561.12</td>\n",
" <td>7587.34</td>\n",
" <td>8588.743164</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2018-05-25</th>\n",
" <td>7592.30</td>\n",
" <td>7480.14</td>\n",
" <td>8445.044922</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2018-05-26</th>\n",
" <td>7486.48</td>\n",
" <td>7355.88</td>\n",
" <td>8267.884766</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2018-05-27</th>\n",
" <td>7362.08</td>\n",
" <td>7368.22</td>\n",
" <td>8082.372070</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2018-05-28</th>\n",
" <td>7371.31</td>\n",
" <td>7135.99</td>\n",
" <td>7915.746094</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2018-05-29</th>\n",
" <td>7129.46</td>\n",
" <td>7472.59</td>\n",
" <td>7765.708984</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2018-05-30</th>\n",
" <td>7469.73</td>\n",
" <td>7406.52</td>\n",
" <td>7666.460938</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2018-05-31</th>\n",
" <td>7406.15</td>\n",
" <td>7494.17</td>\n",
" <td>7620.408691</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2018-06-01</th>\n",
" <td>7500.70</td>\n",
" <td>7541.45</td>\n",
" <td>7624.316895</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2018-06-02</th>\n",
" <td>7536.72</td>\n",
" <td>7643.45</td>\n",
" <td>7666.868164</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2018-06-03</th>\n",
" <td>7632.09</td>\n",
" <td>7720.25</td>\n",
" <td>7738.747559</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2018-06-04</th>\n",
" <td>7722.53</td>\n",
" <td>7514.47</td>\n",
" <td>7829.735840</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2018-06-05</th>\n",
" <td>7500.90</td>\n",
" <td>7633.76</td>\n",
" <td>7903.632812</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>"
],
"text/plain": [
" Open Close Predicted_Close\n",
"Date \n",
"2018-05-17 8370.05 8094.32 8908.131836\n",
"2018-05-18 8091.83 8250.97 8823.910156\n",
"2018-05-19 8255.73 8247.18 8739.717773\n",
"2018-05-20 8246.99 8513.25 8670.551758\n",
"2018-05-21 8522.33 8418.99 8650.153320\n",
"2018-05-22 8419.87 8041.78 8668.397461\n",
"2018-05-23 8037.08 7557.82 8668.617188\n",
"2018-05-24 7561.12 7587.34 8588.743164\n",
"2018-05-25 7592.30 7480.14 8445.044922\n",
"2018-05-26 7486.48 7355.88 8267.884766\n",
"2018-05-27 7362.08 7368.22 8082.372070\n",
"2018-05-28 7371.31 7135.99 7915.746094\n",
"2018-05-29 7129.46 7472.59 7765.708984\n",
"2018-05-30 7469.73 7406.52 7666.460938\n",
"2018-05-31 7406.15 7494.17 7620.408691\n",
"2018-06-01 7500.70 7541.45 7624.316895\n",
"2018-06-02 7536.72 7643.45 7666.868164\n",
"2018-06-03 7632.09 7720.25 7738.747559\n",
"2018-06-04 7722.53 7514.47 7829.735840\n",
"2018-06-05 7500.90 7633.76 7903.632812"
]
},
"metadata": {
"tags": []
},
"execution_count": 39
}
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "uc74FsC92HWH",
"colab_type": "text"
},
"source": [
"## Step 8. Further Analysis - Buy or Sell"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "Zk7H8132ypyn",
"colab_type": "text"
},
"source": [
"In this step we will simply analyse the predicted and real closing prices and based on the data we will decide whether it is good to buy or sell the Bitcoins.\n",
"\n",
"Condition :\n",
"\n",
"If the closing price is greater than the opening price we will decide to sell the coins, otherwise will will buy more coins.\n",
"\n"
]
},
{
"cell_type": "code",
"metadata": {
"id": "qN8ZGbTG29ko",
"colab_type": "code",
"colab": {}
},
"source": [
"#A function to decide buy or sell (1 for sell and 0 for buy)\n",
"\n",
"def buy_sell(op, cl):\n",
" \n",
" #If closing price is greater than the opening price, its a buy\n",
" if cl > op: \n",
" return 1\n",
" #If closing price is lesser than the opening price, its a sell \n",
" else:\n",
" return 0\n",
" "
],
"execution_count": 0,
"outputs": []
},
{
"cell_type": "code",
"metadata": {
"id": "-hPnwtJS3QJ2",
"colab_type": "code",
"colab": {}
},
"source": [
"#Mapping the buy_sell method to the actual closing price in the test data and saving it as a column\n",
"compare[\"Buy-(Actual)\"] = list(map(buy_sell, compare['Open'], compare['Close']))"
],
"execution_count": 0,
"outputs": []
},
{
"cell_type": "code",
"metadata": {
"id": "tki3AZup4adv",
"colab_type": "code",
"colab": {}
},
"source": [
"#Mapping the buy_sell method to the predicted closing price in the test data and saving it as a column\n",
"compare[\"Buy-(Predicted)\"] = list(map(buy_sell, compare['Open'], compare['Predicted_Close']))"
],
"execution_count": 0,
"outputs": []
},
{
"cell_type": "code",
"metadata": {
"id": "2kefDiR54UFA",
"colab_type": "code",
"outputId": "bcf773df-61b6-472f-ab0a-03998fd7b91d",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 235
}
},
"source": [
"#Checkin the new dataframe\n",
"compare.tail()"
],
"execution_count": 50,
"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>Open</th>\n",
" <th>Close</th>\n",
" <th>Predicted_Close</th>\n",
" <th>Buy-(Actual)</th>\n",
" <th>Buy-(Predicted)</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",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>2019-08-16</th>\n",
" <td>10319.42</td>\n",
" <td>10374.34</td>\n",
" <td>10936.112305</td>\n",
" <td>1</td>\n",
" <td>1</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2019-08-17</th>\n",
" <td>10358.72</td>\n",
" <td>10231.74</td>\n",
" <td>10685.478516</td>\n",
" <td>0</td>\n",
" <td>1</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2019-08-18</th>\n",
" <td>10233.01</td>\n",
" <td>10345.81</td>\n",
" <td>10514.767578</td>\n",
" <td>1</td>\n",
" <td>1</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2019-08-19</th>\n",
" <td>10350.28</td>\n",
" <td>10916.05</td>\n",
" <td>10435.475586</td>\n",
" <td>1</td>\n",
" <td>1</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2019-08-20</th>\n",
" <td>10916.35</td>\n",
" <td>10763.23</td>\n",
" <td>10501.110352</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>"
],
"text/plain": [
" Open Close Predicted_Close Buy-(Actual) Buy-(Predicted)\n",
"Date \n",
"2019-08-16 10319.42 10374.34 10936.112305 1 1\n",
"2019-08-17 10358.72 10231.74 10685.478516 0 1\n",
"2019-08-18 10233.01 10345.81 10514.767578 1 1\n",
"2019-08-19 10350.28 10916.05 10435.475586 1 1\n",
"2019-08-20 10916.35 10763.23 10501.110352 0 0"
]
},
"metadata": {
"tags": []
},
"execution_count": 50
}
]
},
{
"cell_type": "code",
"metadata": {
"id": "UZq7SSoN4V9y",
"colab_type": "code",
"colab": {}
},
"source": [
"#Calculating Buy or Sell accuracy\n",
"from sklearn.metrics import confusion_matrix\n",
"cm = confusion_matrix(compare['Buy-(Actual)'], compare['Buy-(Predicted)'])"
],
"execution_count": 0,
"outputs": []
},
{
"cell_type": "code",
"metadata": {
"id": "TbAAWlqN40VV",
"colab_type": "code",
"outputId": "0c5d05b8-a693-44c8-bc74-0d9dbe45ac7a",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 34
}
},
"source": [
"Accuracy = cm.diagonal().sum()/cm.sum()\n",
"print(\"Buy or Sell Accuracy : \", Accuracy)"
],
"execution_count": 45,
"outputs": [
{
"output_type": "stream",
"text": [
"Buy or Sell Accuracy : 0.5249457700650759\n"
],
"name": "stdout"
}
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "-EyXeLPH0GPU",
"colab_type": "text"
},
"source": [
"### Best Approach\n",
"\n",
"The Buy or Sell can be predicted using the RNN. In the above approach we used regression to predict the closing prices and then mapped the Buy or Sell method. Instead we can use the Buy/Sell as the target variable and use the RNN for categorical prediction."
]
}
]
}
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment