LSTM(2)_DEFINITIVO.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## ANNEX 5 (2): LSTM (2)\n",
    "In this Annex the method LSTM will be implemented for the time period from 2017-10-01 to 2018-03-22. Due to the window selection, however, the values predicted that will be obtained are 167 instead of 174 (number of days between the above dates). For that reason, predictions will start in 2017-10-07. In this case, a hidden layer consisting of 20 neurons is used. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>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>8</th>\n",
       "      <td>2018-03-22</td>\n",
       "      <td>8939.44</td>\n",
       "      <td>9100.71</td>\n",
       "      <td>8564.90</td>\n",
       "      <td>8728.47</td>\n",
       "      <td>5530390000</td>\n",
       "      <td>151366000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>2018-03-21</td>\n",
       "      <td>8937.48</td>\n",
       "      <td>9177.37</td>\n",
       "      <td>8846.33</td>\n",
       "      <td>8929.28</td>\n",
       "      <td>6043130000</td>\n",
       "      <td>151316000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>2018-03-20</td>\n",
       "      <td>8619.67</td>\n",
       "      <td>9051.02</td>\n",
       "      <td>8389.89</td>\n",
       "      <td>8913.47</td>\n",
       "      <td>6361790000</td>\n",
       "      <td>145922000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>2018-03-19</td>\n",
       "      <td>8344.12</td>\n",
       "      <td>8675.87</td>\n",
       "      <td>8182.40</td>\n",
       "      <td>8630.65</td>\n",
       "      <td>6729110000</td>\n",
       "      <td>141240000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>2018-03-18</td>\n",
       "      <td>7890.52</td>\n",
       "      <td>8245.51</td>\n",
       "      <td>7397.99</td>\n",
       "      <td>8223.68</td>\n",
       "      <td>6639190000</td>\n",
       "      <td>133547000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         Date     Open     High      Low    Close      Volume    Market Cap\n",
       "8  2018-03-22  8939.44  9100.71  8564.90  8728.47  5530390000  151366000000\n",
       "9  2018-03-21  8937.48  9177.37  8846.33  8929.28  6043130000  151316000000\n",
       "10 2018-03-20  8619.67  9051.02  8389.89  8913.47  6361790000  145922000000\n",
       "11 2018-03-19  8344.12  8675.87  8182.40  8630.65  6729110000  141240000000\n",
       "12 2018-03-18  7890.52  8245.51  7397.99  8223.68  6639190000  133547000000"
      ]
     },
     "execution_count": 1,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import pandas as pd\n",
    "import time\n",
    "import seaborn as sns\n",
    "import matplotlib.pyplot as plt\n",
    "import datetime\n",
    "import numpy as np\n",
    "\n",
    "# get market info for bitcoin from 28th April 2014 to the current day\n",
    "bitcoin_market_info = pd.read_html(\"https://coinmarketcap.com/currencies/bitcoin/historical-data/?start=20130428&end=\"+time.strftime(\"%Y%m%d\"))[0]\n",
    "# convert the date string to the correct date format\n",
    "bitcoin_market_info = bitcoin_market_info.assign(Date=pd.to_datetime(bitcoin_market_info['Date']))\n",
    "# when Volume is equal to '-' convert it to 0\n",
    "bitcoin_market_info.loc[bitcoin_market_info['Volume']==\"-\",'Volume']=0\n",
    "# convert to int\n",
    "bitcoin_market_info['Volume'] = bitcoin_market_info['Volume'].astype('int64')\n",
    "# look at the first few rows\n",
    "bitcoin_market_info = bitcoin_market_info[8:] #depending on which day you want for your last prediction.\n",
    "#in this case it was 2018-03-22, as data is retrieved on 30th, we want only from row 7 onwards.\n",
    "bitcoin_market_info.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>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>8</th>\n",
       "      <td>2018-03-22</td>\n",
       "      <td>562.10</td>\n",
       "      <td>577.57</td>\n",
       "      <td>523.09</td>\n",
       "      <td>539.70</td>\n",
       "      <td>1523460000</td>\n",
       "      <td>55273800000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>2018-03-21</td>\n",
       "      <td>559.10</td>\n",
       "      <td>589.61</td>\n",
       "      <td>550.57</td>\n",
       "      <td>561.73</td>\n",
       "      <td>1781270000</td>\n",
       "      <td>54967300000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>2018-03-20</td>\n",
       "      <td>556.72</td>\n",
       "      <td>567.09</td>\n",
       "      <td>521.20</td>\n",
       "      <td>557.17</td>\n",
       "      <td>1833680000</td>\n",
       "      <td>54722100000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>2018-03-19</td>\n",
       "      <td>546.63</td>\n",
       "      <td>558.10</td>\n",
       "      <td>519.12</td>\n",
       "      <td>556.73</td>\n",
       "      <td>2046790000</td>\n",
       "      <td>53718600000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>2018-03-18</td>\n",
       "      <td>551.64</td>\n",
       "      <td>551.64</td>\n",
       "      <td>460.09</td>\n",
       "      <td>538.64</td>\n",
       "      <td>2685500000</td>\n",
       "      <td>54200100000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         Date    Open    High     Low   Close      Volume   Market Cap\n",
       "8  2018-03-22  562.10  577.57  523.09  539.70  1523460000  55273800000\n",
       "9  2018-03-21  559.10  589.61  550.57  561.73  1781270000  54967300000\n",
       "10 2018-03-20  556.72  567.09  521.20  557.17  1833680000  54722100000\n",
       "11 2018-03-19  546.63  558.10  519.12  556.73  2046790000  53718600000\n",
       "12 2018-03-18  551.64  551.64  460.09  538.64  2685500000  54200100000"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# get market info for ethereum from the start of 2016 to the current day\n",
    "eth_market_info = pd.read_html(\"https://coinmarketcap.com/currencies/ethereum/historical-data/?start=20130428&end=\"+time.strftime(\"%Y%m%d\"))[0]\n",
    "# convert the date string to the correct date format\n",
    "eth_market_info = eth_market_info.assign(Date=pd.to_datetime(eth_market_info['Date']))\n",
    "# look at the first few rows\n",
    "eth_market_info = eth_market_info[8:]\n",
    "eth_market_info.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>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>8</th>\n",
       "      <td>2018-03-22</td>\n",
       "      <td>434.44</td>\n",
       "      <td>438.40</td>\n",
       "      <td>398.59</td>\n",
       "      <td>411.36</td>\n",
       "      <td>96356400</td>\n",
       "      <td>3458220000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>2018-03-21</td>\n",
       "      <td>429.15</td>\n",
       "      <td>450.64</td>\n",
       "      <td>424.79</td>\n",
       "      <td>434.51</td>\n",
       "      <td>115173000</td>\n",
       "      <td>3415280000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>2018-03-20</td>\n",
       "      <td>412.49</td>\n",
       "      <td>442.05</td>\n",
       "      <td>399.17</td>\n",
       "      <td>428.88</td>\n",
       "      <td>109677000</td>\n",
       "      <td>3281830000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>2018-03-19</td>\n",
       "      <td>392.32</td>\n",
       "      <td>409.32</td>\n",
       "      <td>368.52</td>\n",
       "      <td>409.32</td>\n",
       "      <td>113994000</td>\n",
       "      <td>3120600000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>2018-03-18</td>\n",
       "      <td>388.19</td>\n",
       "      <td>394.15</td>\n",
       "      <td>342.51</td>\n",
       "      <td>394.15</td>\n",
       "      <td>117296000</td>\n",
       "      <td>3086950000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         Date    Open    High     Low   Close     Volume  Market Cap\n",
       "8  2018-03-22  434.44  438.40  398.59  411.36   96356400  3458220000\n",
       "9  2018-03-21  429.15  450.64  424.79  434.51  115173000  3415280000\n",
       "10 2018-03-20  412.49  442.05  399.17  428.88  109677000  3281830000\n",
       "11 2018-03-19  392.32  409.32  368.52  409.32  113994000  3120600000\n",
       "12 2018-03-18  388.19  394.15  342.51  394.15  117296000  3086950000"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#do the same for dash\n",
    "dash_market_info = pd.read_html(\"https://coinmarketcap.com/currencies/dash/historical-data/?start=20130428&end=\"+time.strftime(\"%Y%m%d\"))[0]\n",
    "# convert the date string to the correct date format\n",
    "dash_market_info = dash_market_info.assign(Date=pd.to_datetime(dash_market_info['Date']))\n",
    "#dash_market_info.loc[dash_market_info['Volume']==\"-\",'Volume']=0\n",
    "# convert to int\n",
    "dash_market_info['Volume'] = dash_market_info['Volume'].astype('int64')\n",
    "# look at the first few rows\n",
    "dash_market_info = dash_market_info[8:] \n",
    "dash_market_info.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>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>8</th>\n",
       "      <td>2018-03-22</td>\n",
       "      <td>218.56</td>\n",
       "      <td>223.68</td>\n",
       "      <td>204.90</td>\n",
       "      <td>212.71</td>\n",
       "      <td>40394500</td>\n",
       "      <td>3464170000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>2018-03-21</td>\n",
       "      <td>224.26</td>\n",
       "      <td>235.90</td>\n",
       "      <td>215.28</td>\n",
       "      <td>218.63</td>\n",
       "      <td>57121400</td>\n",
       "      <td>3553750000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>2018-03-20</td>\n",
       "      <td>217.94</td>\n",
       "      <td>233.27</td>\n",
       "      <td>206.96</td>\n",
       "      <td>223.74</td>\n",
       "      <td>60005900</td>\n",
       "      <td>3452770000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>2018-03-19</td>\n",
       "      <td>210.33</td>\n",
       "      <td>222.00</td>\n",
       "      <td>202.38</td>\n",
       "      <td>216.77</td>\n",
       "      <td>67314400</td>\n",
       "      <td>3331520000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>2018-03-18</td>\n",
       "      <td>197.43</td>\n",
       "      <td>209.63</td>\n",
       "      <td>175.08</td>\n",
       "      <td>209.63</td>\n",
       "      <td>75859900</td>\n",
       "      <td>3126390000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         Date    Open    High     Low   Close    Volume  Market Cap\n",
       "8  2018-03-22  218.56  223.68  204.90  212.71  40394500  3464170000\n",
       "9  2018-03-21  224.26  235.90  215.28  218.63  57121400  3553750000\n",
       "10 2018-03-20  217.94  233.27  206.96  223.74  60005900  3452770000\n",
       "11 2018-03-19  210.33  222.00  202.38  216.77  67314400  3331520000\n",
       "12 2018-03-18  197.43  209.63  175.08  209.63  75859900  3126390000"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "xmr_market_info = pd.read_html(\"https://coinmarketcap.com/currencies/monero/historical-data/?start=20130428&end=\"+time.strftime(\"%Y%m%d\"))[0]\n",
    "# convert the date string to the correct date format\n",
    "xmr_market_info = xmr_market_info.assign(Date=pd.to_datetime(dash_market_info['Date']))\n",
    "#dash_market_info.loc[dash_market_info['Volume']==\"-\",'Volume']=0\n",
    "# convert to int\n",
    "xmr_market_info['Volume'] = xmr_market_info['Volume'].astype('int64')\n",
    "# look at the first few rows\n",
    "xmr_market_info = xmr_market_info[8:] \n",
    "xmr_market_info.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Some python packages have been loaded and then the table in https://coinmarketcap.com/currencies/bitcoin/historical-data/ is imported. After cleaning the data that table is obtained. By replacing 'bitcoin' with 'ethereum' the corresponding table for the 'ether' token is obtained. Below, both prices are plotted. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "# getting the Bitcoin and Eth logos\n",
    "import sys\n",
    "from PIL import Image\n",
    "import io\n",
    "\n",
    "if sys.version_info[0] < 3:\n",
    "    import urllib2 as urllib\n",
    "    bt_img = urllib.urlopen(\"http://logok.org/wp-content/uploads/2016/10/Bitcoin-Logo-640x480.png\")\n",
    "    eth_img = urllib.urlopen(\"https://upload.wikimedia.org/wikipedia/commons/thumb/0/05/Ethereum_logo_2014.svg/256px-Ethereum_logo_2014.svg.png\")\n",
    "else:\n",
    "    import urllib\n",
    "    bt_img = urllib.request.urlopen(\"http://logok.org/wp-content/uploads/2016/10/Bitcoin-Logo-640x480.png\")\n",
    "    eth_img = urllib.request.urlopen(\"https://upload.wikimedia.org/wikipedia/commons/thumb/0/05/Ethereum_logo_2014.svg/256px-Ethereum_logo_2014.svg.png\")\n",
    "\n",
    "image_file = io.BytesIO(bt_img.read())\n",
    "bitcoin_im = Image.open(image_file)\n",
    "\n",
    "image_file = io.BytesIO(eth_img.read())\n",
    "eth_im = Image.open(image_file)\n",
    "width_eth_im , height_eth_im  = eth_im.size\n",
    "eth_im = eth_im.resize((int(eth_im.size[0]*0.1), int(eth_im.size[1]*0.1)), Image.ANTIALIAS)\n",
    "bitcoin_im = bitcoin_im.resize((int(bitcoin_im.size[0]*0.1), int(bitcoin_im.size[1]*0.1)), Image.ANTIALIAS)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "bitcoin_market_info.columns =[bitcoin_market_info.columns[0]]+['bt_'+i for i in bitcoin_market_info.columns[1:]]\n",
    "eth_market_info.columns =[eth_market_info.columns[0]]+['eth_'+i for i in eth_market_info.columns[1:]]\n",
    "dash_market_info.columns =[dash_market_info.columns[0]]+['dash_'+i for i in dash_market_info.columns[1:]]\n",
    "xmr_market_info.columns =[xmr_market_info.columns[0]]+['xmr_'+i for i in xmr_market_info.columns[1:]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x114b50dd8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, (ax1, ax2) = plt.subplots(2,1, gridspec_kw = {'height_ratios':[3, 1]})\n",
    "ax1.set_ylabel('Closing Price ($)',fontsize=12)\n",
    "ax2.set_ylabel('Volume ($ bn)',fontsize=12)\n",
    "ax2.set_yticks([int('%d000000000'%i) for i in range(10)])\n",
    "ax2.set_yticklabels(range(10))\n",
    "ax1.set_xticks([datetime.date(i,j,1) for i in [2014, 2016, 2018] for j in [1]])\n",
    "ax1.set_xticklabels('')\n",
    "ax2.set_xticks([datetime.date(i,j,1) for i in [2014, 2016, 2018] for j in [1]])\n",
    "ax2.set_xticklabels([datetime.date(i,j,1).strftime('%b %Y')  for i in [2014, 2016, 2018] for j in [1]])\n",
    "ax1.plot(bitcoin_market_info['Date'].astype(datetime.datetime),bitcoin_market_info['bt_Open'])\n",
    "ax2.bar(bitcoin_market_info['Date'].astype(datetime.datetime).values, bitcoin_market_info['bt_Volume'].values)\n",
    "fig.tight_layout()\n",
    "fig.figimage(bitcoin_im, 100, 120, zorder=3,alpha=.5)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x10e190828>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, (ax1, ax2) = plt.subplots(2,1, gridspec_kw = {'height_ratios':[3, 1]})\n",
    "#ax1.set_yscale('log')\n",
    "ax1.set_ylabel('Closing Price ($)',fontsize=12)\n",
    "ax2.set_ylabel('Volume ($ bn)',fontsize=12)\n",
    "ax2.set_yticks([int('%d000000000'%i) for i in range(10)])\n",
    "ax2.set_yticklabels(range(10))\n",
    "ax1.set_xticks([datetime.date(i,j,1) for i in range(2013,2019) for j in [1,7]])\n",
    "ax1.set_xticklabels('')\n",
    "ax2.set_xticks([datetime.date(i,j,1) for i in range(2013,2019) for j in [1,7]])\n",
    "ax2.set_xticklabels([datetime.date(i,j,1).strftime('%b %Y')  for i in range(2013,2019) for j in [1,7]])\n",
    "ax1.plot(eth_market_info['Date'].astype(datetime.datetime),eth_market_info['eth_Open'])\n",
    "ax2.bar(eth_market_info['Date'].astype(datetime.datetime).values, eth_market_info['eth_Volume'].values)\n",
    "fig.tight_layout()\n",
    "fig.figimage(eth_im, 300, 180, zorder=3, alpha=.6)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Date</th>\n",
       "      <th>bt_Open</th>\n",
       "      <th>bt_High</th>\n",
       "      <th>bt_Low</th>\n",
       "      <th>bt_Close</th>\n",
       "      <th>bt_Volume</th>\n",
       "      <th>bt_Market Cap</th>\n",
       "      <th>eth_Open</th>\n",
       "      <th>eth_High</th>\n",
       "      <th>eth_Low</th>\n",
       "      <th>...</th>\n",
       "      <th>xmr_Open</th>\n",
       "      <th>xmr_High</th>\n",
       "      <th>xmr_Low</th>\n",
       "      <th>xmr_Close</th>\n",
       "      <th>xmr_Volume</th>\n",
       "      <th>xmr_Market Cap</th>\n",
       "      <th>bt_day_diff</th>\n",
       "      <th>eth_day_diff</th>\n",
       "      <th>dash_day_diff</th>\n",
       "      <th>xmr_day_diff</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2018-03-22</td>\n",
       "      <td>8939.44</td>\n",
       "      <td>9100.71</td>\n",
       "      <td>8564.90</td>\n",
       "      <td>8728.47</td>\n",
       "      <td>5530390000</td>\n",
       "      <td>151366000000</td>\n",
       "      <td>562.10</td>\n",
       "      <td>577.57</td>\n",
       "      <td>523.09</td>\n",
       "      <td>...</td>\n",
       "      <td>218.56</td>\n",
       "      <td>223.68</td>\n",
       "      <td>204.90</td>\n",
       "      <td>212.71</td>\n",
       "      <td>40394500</td>\n",
       "      <td>3464170000</td>\n",
       "      <td>-0.023600</td>\n",
       "      <td>-0.039851</td>\n",
       "      <td>-0.053126</td>\n",
       "      <td>-0.026766</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2018-03-21</td>\n",
       "      <td>8937.48</td>\n",
       "      <td>9177.37</td>\n",
       "      <td>8846.33</td>\n",
       "      <td>8929.28</td>\n",
       "      <td>6043130000</td>\n",
       "      <td>151316000000</td>\n",
       "      <td>559.10</td>\n",
       "      <td>589.61</td>\n",
       "      <td>550.57</td>\n",
       "      <td>...</td>\n",
       "      <td>224.26</td>\n",
       "      <td>235.90</td>\n",
       "      <td>215.28</td>\n",
       "      <td>218.63</td>\n",
       "      <td>57121400</td>\n",
       "      <td>3553750000</td>\n",
       "      <td>-0.000917</td>\n",
       "      <td>0.004704</td>\n",
       "      <td>0.012490</td>\n",
       "      <td>-0.025105</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2018-03-20</td>\n",
       "      <td>8619.67</td>\n",
       "      <td>9051.02</td>\n",
       "      <td>8389.89</td>\n",
       "      <td>8913.47</td>\n",
       "      <td>6361790000</td>\n",
       "      <td>145922000000</td>\n",
       "      <td>556.72</td>\n",
       "      <td>567.09</td>\n",
       "      <td>521.20</td>\n",
       "      <td>...</td>\n",
       "      <td>217.94</td>\n",
       "      <td>233.27</td>\n",
       "      <td>206.96</td>\n",
       "      <td>223.74</td>\n",
       "      <td>60005900</td>\n",
       "      <td>3452770000</td>\n",
       "      <td>0.034085</td>\n",
       "      <td>0.000808</td>\n",
       "      <td>0.039734</td>\n",
       "      <td>0.026613</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2018-03-19</td>\n",
       "      <td>8344.12</td>\n",
       "      <td>8675.87</td>\n",
       "      <td>8182.40</td>\n",
       "      <td>8630.65</td>\n",
       "      <td>6729110000</td>\n",
       "      <td>141240000000</td>\n",
       "      <td>546.63</td>\n",
       "      <td>558.10</td>\n",
       "      <td>519.12</td>\n",
       "      <td>...</td>\n",
       "      <td>210.33</td>\n",
       "      <td>222.00</td>\n",
       "      <td>202.38</td>\n",
       "      <td>216.77</td>\n",
       "      <td>67314400</td>\n",
       "      <td>3331520000</td>\n",
       "      <td>0.034339</td>\n",
       "      <td>0.018477</td>\n",
       "      <td>0.043332</td>\n",
       "      <td>0.030619</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2018-03-18</td>\n",
       "      <td>7890.52</td>\n",
       "      <td>8245.51</td>\n",
       "      <td>7397.99</td>\n",
       "      <td>8223.68</td>\n",
       "      <td>6639190000</td>\n",
       "      <td>133547000000</td>\n",
       "      <td>551.64</td>\n",
       "      <td>551.64</td>\n",
       "      <td>460.09</td>\n",
       "      <td>...</td>\n",
       "      <td>197.43</td>\n",
       "      <td>209.63</td>\n",
       "      <td>175.08</td>\n",
       "      <td>209.63</td>\n",
       "      <td>75859900</td>\n",
       "      <td>3126390000</td>\n",
       "      <td>0.042223</td>\n",
       "      <td>-0.023566</td>\n",
       "      <td>0.015353</td>\n",
       "      <td>0.061794</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 29 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "        Date  bt_Open  bt_High   bt_Low  bt_Close   bt_Volume  bt_Market Cap  \\\n",
       "0 2018-03-22  8939.44  9100.71  8564.90   8728.47  5530390000   151366000000   \n",
       "1 2018-03-21  8937.48  9177.37  8846.33   8929.28  6043130000   151316000000   \n",
       "2 2018-03-20  8619.67  9051.02  8389.89   8913.47  6361790000   145922000000   \n",
       "3 2018-03-19  8344.12  8675.87  8182.40   8630.65  6729110000   141240000000   \n",
       "4 2018-03-18  7890.52  8245.51  7397.99   8223.68  6639190000   133547000000   \n",
       "\n",
       "   eth_Open  eth_High  eth_Low      ...       xmr_Open  xmr_High xmr_Low  \\\n",
       "0    562.10    577.57   523.09      ...         218.56    223.68  204.90   \n",
       "1    559.10    589.61   550.57      ...         224.26    235.90  215.28   \n",
       "2    556.72    567.09   521.20      ...         217.94    233.27  206.96   \n",
       "3    546.63    558.10   519.12      ...         210.33    222.00  202.38   \n",
       "4    551.64    551.64   460.09      ...         197.43    209.63  175.08   \n",
       "\n",
       "   xmr_Close  xmr_Volume  xmr_Market Cap  bt_day_diff  eth_day_diff  \\\n",
       "0     212.71    40394500      3464170000    -0.023600     -0.039851   \n",
       "1     218.63    57121400      3553750000    -0.000917      0.004704   \n",
       "2     223.74    60005900      3452770000     0.034085      0.000808   \n",
       "3     216.77    67314400      3331520000     0.034339      0.018477   \n",
       "4     209.63    75859900      3126390000     0.042223     -0.023566   \n",
       "\n",
       "   dash_day_diff  xmr_day_diff  \n",
       "0      -0.053126     -0.026766  \n",
       "1       0.012490     -0.025105  \n",
       "2       0.039734      0.026613  \n",
       "3       0.043332      0.030619  \n",
       "4       0.015353      0.061794  \n",
       "\n",
       "[5 rows x 29 columns]"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "market_info = pd.merge(bitcoin_market_info,eth_market_info, on=['Date'])\n",
    "market_info = pd.merge(market_info, dash_market_info, on=['Date'])\n",
    "market_info = pd.merge(market_info, xmr_market_info, on = ['Date'])\n",
    "\n",
    "market_info = market_info[market_info['Date']>='2016-01-01'] \n",
    "\n",
    "for coins in ['bt_', 'eth_', 'dash_', 'xmr_']: \n",
    "    kwargs = { coins+'day_diff': lambda x: (x[coins+'Close']-x[coins+'Open'])/x[coins+'Open']}\n",
    "    market_info = market_info.assign(**kwargs)\n",
    "    \n",
    "market_info.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x115ba0b00>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "split_date = '2017-10-01' #as with the ARIMA + GARCH (1, 1), we start in 2017-10-01; \n",
    "#however, notice that due to the window selection, the first predicted value\n",
    "#will correspond to 2017-10-07\n",
    "fig, (ax1, ax2) = plt.subplots(2,1)\n",
    "ax1.set_xticks([datetime.date(i,j,1) for i in range(2013,2019) for j in [1,7]])\n",
    "ax1.set_xticklabels('')\n",
    "ax2.set_xticks([datetime.date(i,j,1) for i in range(2013,2019) for j in [1,7]])\n",
    "ax2.set_xticklabels([datetime.date(i,j,1).strftime('%b %Y')  for i in range(2013,2019) for j in [1,7]])\n",
    "ax1.plot(market_info[market_info['Date'] < split_date]['Date'].astype(datetime.datetime),\n",
    "         market_info[market_info['Date'] < split_date]['bt_Close'], \n",
    "         color='#B08FC7', label='Training')\n",
    "ax1.plot(market_info[market_info['Date'] >= split_date]['Date'].astype(datetime.datetime),\n",
    "         market_info[market_info['Date'] >= split_date]['bt_Close'], \n",
    "         color='#8FBAC8', label='Test')\n",
    "ax2.plot(market_info[market_info['Date'] < split_date]['Date'].astype(datetime.datetime),\n",
    "         market_info[market_info['Date'] < split_date]['eth_Close'], \n",
    "         color='#B08FC7')\n",
    "ax2.plot(market_info[market_info['Date'] >= split_date]['Date'].astype(datetime.datetime),\n",
    "         market_info[market_info['Date'] >= split_date]['eth_Close'], color='#8FBAC8')\n",
    "ax1.set_xticklabels('')\n",
    "ax1.set_ylabel('Bitcoin Price ($)',fontsize=12)\n",
    "ax2.set_ylabel('Ethereum Price ($)',fontsize=12)\n",
    "plt.tight_layout()\n",
    "ax1.legend(bbox_to_anchor=(0.03, 1), loc=2, borderaxespad=0., prop={'size': 14})\n",
    "fig.figimage(bitcoin_im.resize((int(bitcoin_im.size[0]*0.65), int(bitcoin_im.size[1]*0.65)), Image.ANTIALIAS), \n",
    "             350, 170, zorder=3,alpha=.5)\n",
    "fig.figimage(eth_im.resize((int(eth_im.size[0]*0.65), int(eth_im.size[1]*0.65)), Image.ANTIALIAS), \n",
    "             360, 40, zorder=3,alpha=.5)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "First, we may want to make sure the daily change in price follows a normal distribution. We'll plot the histogram of values. This is especially important for recurrent neural networks."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x115b53d30>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, (ax1, ax2) = plt.subplots(1,2)\n",
    "ax1.hist(market_info[market_info['Date']< split_date]['bt_day_diff'].values, bins=100)\n",
    "ax2.hist(market_info[market_info['Date']< split_date]['eth_day_diff'].values, bins=100)\n",
    "ax1.set_title('Bitcoin Daily Price Changes')\n",
    "ax2.set_title('Ethereum Daily Price Changes')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Long Short Term Memory (LSTM)\n",
    "\n",
    "For an explanation of how ANN, and especially LSTM work, see the references in the sub-section Long Short-Term Memory (LSTM). The Keras (2018) package is used. First, some transformations are carried out, in order to get variables 'close_off_high' and 'volatility' for each cryptocurrency.   "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "for coins in ['bt_', 'eth_', 'dash_', 'xmr_']: \n",
    "    kwargs = { coins+'close_off_high': lambda x: 2*(x[coins+'High']- x[coins+'Close'])/(x[coins+'High']-x[coins+'Low'])-1,\n",
    "            coins+'volatility': lambda x: (x[coins+'High']- x[coins+'Low'])/(x[coins+'Open'])}\n",
    "    market_info = market_info.assign(**kwargs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Date</th>\n",
       "      <th>bt_Close</th>\n",
       "      <th>bt_Volume</th>\n",
       "      <th>bt_close_off_high</th>\n",
       "      <th>bt_volatility</th>\n",
       "      <th>eth_Close</th>\n",
       "      <th>eth_Volume</th>\n",
       "      <th>eth_close_off_high</th>\n",
       "      <th>eth_volatility</th>\n",
       "      <th>dash_Close</th>\n",
       "      <th>dash_Volume</th>\n",
       "      <th>dash_close_off_high</th>\n",
       "      <th>dash_volatility</th>\n",
       "      <th>xmr_Close</th>\n",
       "      <th>xmr_Volume</th>\n",
       "      <th>xmr_close_off_high</th>\n",
       "      <th>xmr_volatility</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>811</th>\n",
       "      <td>2016-01-01</td>\n",
       "      <td>434.33</td>\n",
       "      <td>36278900</td>\n",
       "      <td>-0.560641</td>\n",
       "      <td>0.020292</td>\n",
       "      <td>0.948024</td>\n",
       "      <td>206062</td>\n",
       "      <td>-0.418477</td>\n",
       "      <td>0.025040</td>\n",
       "      <td>3.38</td>\n",
       "      <td>98317</td>\n",
       "      <td>6.666667e-02</td>\n",
       "      <td>0.090090</td>\n",
       "      <td>0.497867</td>\n",
       "      <td>88914</td>\n",
       "      <td>0.286720</td>\n",
       "      <td>0.194582</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>810</th>\n",
       "      <td>2016-01-02</td>\n",
       "      <td>433.44</td>\n",
       "      <td>30096600</td>\n",
       "      <td>0.250597</td>\n",
       "      <td>0.009641</td>\n",
       "      <td>0.937124</td>\n",
       "      <td>255504</td>\n",
       "      <td>0.965898</td>\n",
       "      <td>0.034913</td>\n",
       "      <td>3.28</td>\n",
       "      <td>62353</td>\n",
       "      <td>5.789474e-01</td>\n",
       "      <td>0.056213</td>\n",
       "      <td>0.532668</td>\n",
       "      <td>64901</td>\n",
       "      <td>-0.040275</td>\n",
       "      <td>0.132592</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>809</th>\n",
       "      <td>2016-01-03</td>\n",
       "      <td>430.01</td>\n",
       "      <td>39633800</td>\n",
       "      <td>-0.173865</td>\n",
       "      <td>0.020827</td>\n",
       "      <td>0.971905</td>\n",
       "      <td>407632</td>\n",
       "      <td>-0.317885</td>\n",
       "      <td>0.060792</td>\n",
       "      <td>3.24</td>\n",
       "      <td>87155</td>\n",
       "      <td>-1.998401e-15</td>\n",
       "      <td>0.066869</td>\n",
       "      <td>0.500312</td>\n",
       "      <td>43237</td>\n",
       "      <td>0.619697</td>\n",
       "      <td>0.093337</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>808</th>\n",
       "      <td>2016-01-04</td>\n",
       "      <td>433.09</td>\n",
       "      <td>38477500</td>\n",
       "      <td>-0.474265</td>\n",
       "      <td>0.012649</td>\n",
       "      <td>0.954480</td>\n",
       "      <td>346245</td>\n",
       "      <td>-0.057657</td>\n",
       "      <td>0.047943</td>\n",
       "      <td>3.32</td>\n",
       "      <td>65748</td>\n",
       "      <td>-4.545455e-01</td>\n",
       "      <td>0.033846</td>\n",
       "      <td>0.517388</td>\n",
       "      <td>29381</td>\n",
       "      <td>-0.244774</td>\n",
       "      <td>0.058405</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>807</th>\n",
       "      <td>2016-01-05</td>\n",
       "      <td>431.96</td>\n",
       "      <td>34522600</td>\n",
       "      <td>-0.013333</td>\n",
       "      <td>0.010391</td>\n",
       "      <td>0.950176</td>\n",
       "      <td>219833</td>\n",
       "      <td>0.697930</td>\n",
       "      <td>0.025236</td>\n",
       "      <td>3.33</td>\n",
       "      <td>71861</td>\n",
       "      <td>4.666667e-01</td>\n",
       "      <td>0.045181</td>\n",
       "      <td>0.499272</td>\n",
       "      <td>30449</td>\n",
       "      <td>0.811877</td>\n",
       "      <td>0.092133</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          Date  bt_Close  bt_Volume  bt_close_off_high  bt_volatility  \\\n",
       "811 2016-01-01    434.33   36278900          -0.560641       0.020292   \n",
       "810 2016-01-02    433.44   30096600           0.250597       0.009641   \n",
       "809 2016-01-03    430.01   39633800          -0.173865       0.020827   \n",
       "808 2016-01-04    433.09   38477500          -0.474265       0.012649   \n",
       "807 2016-01-05    431.96   34522600          -0.013333       0.010391   \n",
       "\n",
       "     eth_Close  eth_Volume  eth_close_off_high  eth_volatility  dash_Close  \\\n",
       "811   0.948024      206062           -0.418477        0.025040        3.38   \n",
       "810   0.937124      255504            0.965898        0.034913        3.28   \n",
       "809   0.971905      407632           -0.317885        0.060792        3.24   \n",
       "808   0.954480      346245           -0.057657        0.047943        3.32   \n",
       "807   0.950176      219833            0.697930        0.025236        3.33   \n",
       "\n",
       "     dash_Volume  dash_close_off_high  dash_volatility  xmr_Close  xmr_Volume  \\\n",
       "811        98317         6.666667e-02         0.090090   0.497867       88914   \n",
       "810        62353         5.789474e-01         0.056213   0.532668       64901   \n",
       "809        87155        -1.998401e-15         0.066869   0.500312       43237   \n",
       "808        65748        -4.545455e-01         0.033846   0.517388       29381   \n",
       "807        71861         4.666667e-01         0.045181   0.499272       30449   \n",
       "\n",
       "     xmr_close_off_high  xmr_volatility  \n",
       "811            0.286720        0.194582  \n",
       "810           -0.040275        0.132592  \n",
       "809            0.619697        0.093337  \n",
       "808           -0.244774        0.058405  \n",
       "807            0.811877        0.092133  "
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model_data = market_info[['Date']+[coin+metric for coin in ['bt_', 'eth_', 'dash_', 'xmr_'] \n",
    "                                   for metric in ['Close','Volume','close_off_high','volatility']]]\n",
    "# need to reverse the data frame so that subsequent rows represent later timepoints\n",
    "model_data = model_data.sort_values(by='Date')\n",
    "model_data.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "A new data frame ('model_data') was created, removing columns Open, High and Low, and adding 'close_off_high', which represents how close the closing price is from the high price of that day. It is between 1 and -1 (1: close = high, -1: close = low). 'Volatility' is measured as the difference between high and low divided by open. \n",
    "This model uses previous data from bitcoin, ether and dash to predict next day's closing price of bitcoin and ether, separately. It has access to the previous 5 days (window = 5). After trying different windows, it was found that providing the model with too many previous data makes it remember long term relationships among the variables in the training phase that do not hold in the test phase. \n",
    "As widely varying inputs affect data negatively, the close price is normalized as in Aungiers (2017) to the first value of the window. 'Volume' column for each cryptocurrency is eliminated. After this normalization, most of the values for each window are between -1 and 1, which is desirable for deep learning models. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "# we don't need the date columns anymore\n",
    "training_set, test_set = model_data[model_data['Date']<split_date], model_data[model_data['Date']>=split_date]\n",
    "training_set = training_set.drop('Date', 1)\n",
    "test_set = test_set.drop('Date', 1)\n",
    "training_set = training_set.drop('bt_Volume', 1)\n",
    "training_set = training_set.drop('eth_Volume', 1)\n",
    "training_set = training_set.drop('dash_Volume', 1)\n",
    "test_set = test_set.drop('bt_Volume', 1)\n",
    "test_set = test_set.drop('eth_Volume', 1)\n",
    "test_set = test_set.drop('dash_Volume', 1)\n",
    "training_set = training_set.drop('dash_close_off_high', 1)\n",
    "test_set = test_set.drop('dash_close_off_high', 1)\n",
    "training_set = training_set.drop('xmr_Volume', 1)\n",
    "test_set = test_set.drop('xmr_Volume', 1)\n",
    "training_set = training_set.drop('xmr_close_off_high', 1)\n",
    "test_set = test_set.drop('xmr_close_off_high', 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['bt_Close', 'eth_Close', 'dash_Close', 'xmr_Close']\n"
     ]
    }
   ],
   "source": [
    "window_len = 5\n",
    "norm_cols = [coin+metric for coin in ['bt_', 'eth_', 'dash_', 'xmr_'] for metric in ['Close']]\n",
    "print(norm_cols)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "LSTM_training_inputs = []\n",
    "for i in range(len(training_set)-window_len):\n",
    "    temp_set = training_set[i:(i+window_len)].copy()\n",
    "    for col in norm_cols:\n",
    "        temp_set.loc[:, col] = temp_set[col]/temp_set[col].iloc[0] - 1\n",
    "    LSTM_training_inputs.append(temp_set)\n",
    "LSTM_training_outputs = (training_set['eth_Close'][window_len:].values/training_set['eth_Close'][:-window_len].values)-1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'list'>\n"
     ]
    }
   ],
   "source": [
    "LSTM_test_inputs = []\n",
    "for i in range(len(test_set)-window_len):\n",
    "    temp_set = test_set[i:(i+window_len)].copy()\n",
    "    for col in norm_cols:\n",
    "        temp_set.loc[:, col] = temp_set[col]/temp_set[col].iloc[0] - 1\n",
    "    LSTM_test_inputs.append(temp_set)\n",
    "LSTM_test_outputs = (test_set['eth_Close'][window_len:].values/test_set['eth_Close'][:-window_len].values)-1\n",
    "\n",
    "print(type(LSTM_training_inputs))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>bt_Close</th>\n",
       "      <th>bt_close_off_high</th>\n",
       "      <th>bt_volatility</th>\n",
       "      <th>eth_Close</th>\n",
       "      <th>eth_close_off_high</th>\n",
       "      <th>eth_volatility</th>\n",
       "      <th>dash_Close</th>\n",
       "      <th>dash_volatility</th>\n",
       "      <th>xmr_Close</th>\n",
       "      <th>xmr_volatility</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>811</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>-0.560641</td>\n",
       "      <td>0.020292</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>-0.418477</td>\n",
       "      <td>0.025040</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.090090</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.194582</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>810</th>\n",
       "      <td>-0.002049</td>\n",
       "      <td>0.250597</td>\n",
       "      <td>0.009641</td>\n",
       "      <td>-0.011498</td>\n",
       "      <td>0.965898</td>\n",
       "      <td>0.034913</td>\n",
       "      <td>-0.029586</td>\n",
       "      <td>0.056213</td>\n",
       "      <td>0.069900</td>\n",
       "      <td>0.132592</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>809</th>\n",
       "      <td>-0.009946</td>\n",
       "      <td>-0.173865</td>\n",
       "      <td>0.020827</td>\n",
       "      <td>0.025190</td>\n",
       "      <td>-0.317885</td>\n",
       "      <td>0.060792</td>\n",
       "      <td>-0.041420</td>\n",
       "      <td>0.066869</td>\n",
       "      <td>0.004911</td>\n",
       "      <td>0.093337</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>808</th>\n",
       "      <td>-0.002855</td>\n",
       "      <td>-0.474265</td>\n",
       "      <td>0.012649</td>\n",
       "      <td>0.006810</td>\n",
       "      <td>-0.057657</td>\n",
       "      <td>0.047943</td>\n",
       "      <td>-0.017751</td>\n",
       "      <td>0.033846</td>\n",
       "      <td>0.039209</td>\n",
       "      <td>0.058405</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>807</th>\n",
       "      <td>-0.005457</td>\n",
       "      <td>-0.013333</td>\n",
       "      <td>0.010391</td>\n",
       "      <td>0.002270</td>\n",
       "      <td>0.697930</td>\n",
       "      <td>0.025236</td>\n",
       "      <td>-0.014793</td>\n",
       "      <td>0.045181</td>\n",
       "      <td>0.002822</td>\n",
       "      <td>0.092133</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     bt_Close  bt_close_off_high  bt_volatility  eth_Close  \\\n",
       "811  0.000000          -0.560641       0.020292   0.000000   \n",
       "810 -0.002049           0.250597       0.009641  -0.011498   \n",
       "809 -0.009946          -0.173865       0.020827   0.025190   \n",
       "808 -0.002855          -0.474265       0.012649   0.006810   \n",
       "807 -0.005457          -0.013333       0.010391   0.002270   \n",
       "\n",
       "     eth_close_off_high  eth_volatility  dash_Close  dash_volatility  \\\n",
       "811           -0.418477        0.025040    0.000000         0.090090   \n",
       "810            0.965898        0.034913   -0.029586         0.056213   \n",
       "809           -0.317885        0.060792   -0.041420         0.066869   \n",
       "808           -0.057657        0.047943   -0.017751         0.033846   \n",
       "807            0.697930        0.025236   -0.014793         0.045181   \n",
       "\n",
       "     xmr_Close  xmr_volatility  \n",
       "811   0.000000        0.194582  \n",
       "810   0.069900        0.132592  \n",
       "809   0.004911        0.093337  \n",
       "808   0.039209        0.058405  \n",
       "807   0.002822        0.092133  "
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "LSTM_training_inputs[0]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "There are hundreds of similar tables like the above (remember that the model takes 5 previous days and tries to predict the 6th closing price of bitcoin and ether, separately). The next step is to build the LSTM. As the close columns are normalized so that the first value in the window is 0, changes with respect to this point will be predicted. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "# easier to work with numpy arrays rather than pandas dataframes\n",
    "# especially as we now only have numerical data\n",
    "LSTM_training_inputs = [np.array(LSTM_training_input) for LSTM_training_input in LSTM_training_inputs]\n",
    "LSTM_training_inputs = np.array(LSTM_training_inputs)\n",
    "\n",
    "LSTM_test_inputs = [np.array(LSTM_test_inputs) for LSTM_test_inputs in LSTM_test_inputs]\n",
    "LSTM_test_inputs = np.array(LSTM_test_inputs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "from keras.models import Sequential\n",
    "from keras.layers import Activation, Dense\n",
    "from keras.layers import LSTM, LeakyReLU \n",
    "from keras.layers import Dropout\n",
    "from keras.utils import plot_model\n",
    "\n",
    "np.random.seed(202) \n",
    "\n",
    "def build_model(inputs, output_size, neurons, activ_func=\"linear\",\n",
    "                dropout=0.25, loss=\"mae\", optimizer=\"adam\"):\n",
    "    \n",
    "    model = Sequential()\n",
    "\n",
    "    model.add(LSTM(neurons, input_shape=(inputs.shape[1], inputs.shape[2])))\n",
    "    model.add(Dropout(0.25))\n",
    "    model.add(Dense(units=output_size))\n",
    "    model.add(Activation(activ_func))\n",
    "    model.compile(loss=loss, optimizer=optimizer)\n",
    "    return model"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This build_model function is then used for fitting the model into the data. The LSTM layer has been shaped to fit our inputs. For the rest of the specifications in the model, see the references provided in the sub-section 'Long Short-Term Memory (LSTM)'. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/50\n",
      " - 5s - loss: 0.0965\n",
      "Epoch 2/50\n",
      " - 4s - loss: 0.0721\n",
      "Epoch 3/50\n",
      " - 4s - loss: 0.0697\n",
      "Epoch 4/50\n",
      " - 4s - loss: 0.0664\n",
      "Epoch 5/50\n",
      " - 4s - loss: 0.0628\n",
      "Epoch 6/50\n",
      " - 4s - loss: 0.0620\n",
      "Epoch 7/50\n",
      " - 4s - loss: 0.0614\n",
      "Epoch 8/50\n",
      " - 4s - loss: 0.0607\n",
      "Epoch 9/50\n",
      " - 4s - loss: 0.0575\n",
      "Epoch 10/50\n",
      " - 4s - loss: 0.0578\n",
      "Epoch 11/50\n",
      " - 4s - loss: 0.0578\n",
      "Epoch 12/50\n",
      " - 4s - loss: 0.0591\n",
      "Epoch 13/50\n",
      " - 4s - loss: 0.0575\n",
      "Epoch 14/50\n",
      " - 4s - loss: 0.0572\n",
      "Epoch 15/50\n",
      " - 4s - loss: 0.0576\n",
      "Epoch 16/50\n",
      " - 4s - loss: 0.0558\n",
      "Epoch 17/50\n",
      " - 4s - loss: 0.0544\n",
      "Epoch 18/50\n",
      " - 4s - loss: 0.0541\n",
      "Epoch 19/50\n",
      " - 4s - loss: 0.0555\n",
      "Epoch 20/50\n",
      " - 4s - loss: 0.0561\n",
      "Epoch 21/50\n",
      " - 4s - loss: 0.0568\n",
      "Epoch 22/50\n",
      " - 4s - loss: 0.0557\n",
      "Epoch 23/50\n",
      " - 4s - loss: 0.0538\n",
      "Epoch 24/50\n",
      " - 4s - loss: 0.0557\n",
      "Epoch 25/50\n",
      " - 4s - loss: 0.0558\n",
      "Epoch 26/50\n",
      " - 4s - loss: 0.0541\n",
      "Epoch 27/50\n",
      " - 4s - loss: 0.0549\n",
      "Epoch 28/50\n",
      " - 4s - loss: 0.0545\n",
      "Epoch 29/50\n",
      " - 4s - loss: 0.0547\n",
      "Epoch 30/50\n",
      " - 4s - loss: 0.0541\n",
      "Epoch 31/50\n",
      " - 4s - loss: 0.0533\n",
      "Epoch 32/50\n",
      " - 4s - loss: 0.0545\n",
      "Epoch 33/50\n",
      " - 4s - loss: 0.0542\n",
      "Epoch 34/50\n",
      " - 4s - loss: 0.0553\n",
      "Epoch 35/50\n",
      " - 4s - loss: 0.0534\n",
      "Epoch 36/50\n",
      " - 4s - loss: 0.0548\n",
      "Epoch 37/50\n",
      " - 4s - loss: 0.0537\n",
      "Epoch 38/50\n",
      " - 4s - loss: 0.0542\n",
      "Epoch 39/50\n",
      " - 4s - loss: 0.0522\n",
      "Epoch 40/50\n",
      " - 4s - loss: 0.0556\n",
      "Epoch 41/50\n",
      " - 4s - loss: 0.0547\n",
      "Epoch 42/50\n",
      " - 4s - loss: 0.0540\n",
      "Epoch 43/50\n",
      " - 4s - loss: 0.0537\n",
      "Epoch 44/50\n",
      " - 4s - loss: 0.0546\n",
      "Epoch 45/50\n",
      " - 4s - loss: 0.0552\n",
      "Epoch 46/50\n",
      " - 4s - loss: 0.0534\n",
      "Epoch 47/50\n",
      " - 4s - loss: 0.0537\n",
      "Epoch 48/50\n",
      " - 4s - loss: 0.0537\n",
      "Epoch 49/50\n",
      " - 4s - loss: 0.0549\n",
      "Epoch 50/50\n",
      " - 4s - loss: 0.0526\n"
     ]
    }
   ],
   "source": [
    "# random seed for reproducibility\n",
    "np.random.seed(202)\n",
    "# initialise model architecture\n",
    "eth_model = build_model(LSTM_training_inputs, output_size=1, neurons = 20, dropout = 0.35) \n",
    "# model output is next price normalised to 5th previous closing price\n",
    "\n",
    "# train model on data\n",
    "# note: eth_history contains information on the training error per epoch\n",
    "eth_history = eth_model.fit(LSTM_training_inputs, LSTM_training_outputs, \n",
    "                            epochs=50, batch_size=1, verbose=2, shuffle=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'twopi': '/anaconda3/bin/twopi', 'sfdp': '/anaconda3/bin/sfdp', 'fdp': '/anaconda3/bin/fdp', 'neato': '/anaconda3/bin/neato', 'dot': '/anaconda3/bin/dot', 'circo': '/anaconda3/bin/circo'}\n"
     ]
    }
   ],
   "source": [
    "import pydot_ng as pydot\n",
    "print(pydot.find_graphviz())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "plot_model(eth_model, to_file='LSTM_ethmodel_plot.png', show_shapes=True, show_layer_names=True)\n",
    "#with this, the model structure is plotted into a new file. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As it is noticeable, the error decreases throughout time, as expected. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1a290e2d30>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax1 = plt.subplots(1,1)\n",
    "\n",
    "ax1.plot(eth_history.epoch, eth_history.history['loss'])\n",
    "ax1.set_title('Training Error')\n",
    "\n",
    "if eth_model.loss == 'mae':\n",
    "    ax1.set_ylabel('Mean Absolute Error (MAE)',fontsize=12)\n",
    "# just in case you decided to change the model loss calculation\n",
    "else:\n",
    "    ax1.set_ylabel('Model Loss',fontsize=12)\n",
    "ax1.set_xlabel('# Epochs',fontsize=12)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "\"We've just built an LSTM model to predict tomorrow's Ethereum closing price. Let's see how well it performs. We start by examining its performance on the training set (data before October 2017). That number below the code represents the model's mean absolute error (mae) on the training set after the 50th training iteration (or epoch). Instead of relative changes, we can view the model output as daily closing prices.\" (Sheehan, 2017)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1a2911b160>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from mpl_toolkits.axes_grid1.inset_locator import zoomed_inset_axes\n",
    "from mpl_toolkits.axes_grid1.inset_locator import mark_inset\n",
    "\n",
    "fig, ax1 = plt.subplots(1,1)\n",
    "ax1.set_xticks([datetime.date(i,j,1) for i in range(2013,2019) for j in [1,5,9]])\n",
    "ax1.set_xticklabels([datetime.date(i,j,1).strftime('%b %Y')  for i in range(2013,2019) for j in [1,5,9]])\n",
    "ax1.plot(model_data[model_data['Date']< split_date]['Date'][window_len:].astype(datetime.datetime),\n",
    "         training_set['eth_Close'][window_len:], label='Actual')\n",
    "ax1.plot(model_data[model_data['Date']< split_date]['Date'][window_len:].astype(datetime.datetime),\n",
    "         ((np.transpose(eth_model.predict(LSTM_training_inputs))+1) * training_set['eth_Close'].values[:-window_len])[0], \n",
    "         label='Predicted')\n",
    "ax1.set_title('Training Set: Single Timepoint Prediction')\n",
    "ax1.set_ylabel('Ethereum Price ($)',fontsize=12)\n",
    "ax1.legend(bbox_to_anchor=(0.15, 1), loc=2, borderaxespad=0., prop={'size': 14})\n",
    "ax1.annotate('MAE: %.4f'%np.mean(np.abs((np.transpose(eth_model.predict(LSTM_training_inputs))+1)-\\\n",
    "            (training_set['eth_Close'].values[window_len:])/(training_set['eth_Close'].values[:-window_len]))), \n",
    "             xy=(0.75, 0.9),  xycoords='axes fraction',\n",
    "            xytext=(0.75, 0.9), textcoords='axes fraction')\n",
    "# figure inset code taken from http://akuederle.com/matplotlib-zoomed-up-inset\n",
    "axins = zoomed_inset_axes(ax1, 3.35, loc=10) # zoom-factor: 3.35, location: centre\n",
    "axins.set_xticks([datetime.date(i,j,1) for i in range(2013,2019) for j in [1,5,9]])\n",
    "axins.plot(model_data[model_data['Date']< split_date]['Date'][window_len:].astype(datetime.datetime),\n",
    "         training_set['eth_Close'][window_len:], label='Actual')\n",
    "axins.plot(model_data[model_data['Date']< split_date]['Date'][window_len:].astype(datetime.datetime),\n",
    "         ((np.transpose(eth_model.predict(LSTM_training_inputs))+1) * training_set['eth_Close'].values[:-window_len])[0], \n",
    "         label='Predicted')\n",
    "axins.set_xlim([datetime.date(2017, 3, 1), datetime.date(2017, 5, 1)])\n",
    "axins.set_ylim([10,60])\n",
    "axins.set_xticklabels('')\n",
    "mark_inset(ax1, axins, loc1=1, loc2=3, fc=\"none\", ec=\"0.5\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As can be observed, the model is very accurate in the training phase, as expected. This, however, is not very surprising, as the model has access to its error every iteration. The prediction of interest is the one in the test phase, as it will show if the prediction error is so small due to good fit or due to overfitting. For avoiding this problem, a dropout of 0.25 was set in the model. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1a291e67b8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "168\n",
      "4.73004092482\n",
      "the accuracy in percentage measured as 100 - MAPE for ether using LSTM is 95.26996\n"
     ]
    }
   ],
   "source": [
    "fig, ax1 = plt.subplots(1,1)\n",
    "#ax1.set_xticks([datetime.date(2017,i+1,1) for i in range(12)])\n",
    "#ax1.set_xticklabels([datetime.date(2017,i+1,1).strftime('%b %d %Y')  for i in range(12)])\n",
    "ax1.set_xticks([datetime.date(i,j,1) for i in range(2017,2019) for j in [1,5,9]])\n",
    "ax1.set_xticklabels([datetime.date(i,j,1).strftime('%b %Y')  for i in range(2017,2019) for j in [1,5,9]])\n",
    "ax1.plot(model_data[model_data['Date']>= split_date]['Date'][window_len:].astype(datetime.datetime),\n",
    "         test_set['eth_Close'][window_len:], label='Actual')\n",
    "ax1.plot(model_data[model_data['Date']>= split_date]['Date'][window_len:].astype(datetime.datetime),\n",
    "         ((np.transpose(eth_model.predict(LSTM_test_inputs))+1) * test_set['eth_Close'].values[:-window_len])[0], \n",
    "         label='Predicted')\n",
    "ax1.annotate('MAE: %.4f'%np.mean(np.abs((np.transpose(eth_model.predict(LSTM_test_inputs))+1)-\\\n",
    "            (test_set['eth_Close'].values[window_len:])/(test_set['eth_Close'].values[:-window_len]))), \n",
    "             xy=(0.75, 0.9),  xycoords='axes fraction',\n",
    "            xytext=(0.75, 0.9), textcoords='axes fraction')\n",
    "ax1.set_title('Test Set: Single Timepoint Prediction',fontsize=13)\n",
    "ax1.set_ylabel('Ethereum Price ($)',fontsize=12)\n",
    "ax1.legend(bbox_to_anchor=(0.1, 1), loc=2, borderaxespad=0., prop={'size': 14})\n",
    "plt.show()\n",
    "\n",
    "print(len(eth_model.predict(LSTM_test_inputs)))\n",
    "\n",
    "\n",
    "mape_eth = (100/len(eth_model.predict(LSTM_test_inputs)))*np.sum(np.abs(((np.transpose(eth_model.predict(LSTM_test_inputs)))+1)-\\\n",
    "            (test_set['eth_Close'].values[window_len:])/(test_set['eth_Close'].values[:-window_len])))\n",
    "\n",
    "print(mape_eth) \n",
    "\n",
    "print('the accuracy in percentage measured as 100 - MAPE for ether using LSTM is %.5f'%(100-mape_eth))\n",
    "\n",
    "#the accuracy in percentage measured as 100 - MAPE for ether using LSTM is 95.29775"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As it is seen, the LSTM model performs well in the test period, with some weaknesses. First, the model tends to overestimate the future ether price, which can be due to the fact that during the training period ether's price rose incredibly, therefore it expects the trend to continue. Also, in spikes, where big changes take place, the model is not able to capture these behaviors. Now, the same methodology is applied to bitcoin's prices. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/50\n",
      " - 5s - loss: 0.0513\n",
      "Epoch 2/50\n",
      " - 4s - loss: 0.0373\n",
      "Epoch 3/50\n",
      " - 4s - loss: 0.0337\n",
      "Epoch 4/50\n",
      " - 4s - loss: 0.0317\n",
      "Epoch 5/50\n",
      " - 4s - loss: 0.0294\n",
      "Epoch 6/50\n",
      " - 4s - loss: 0.0295\n",
      "Epoch 7/50\n",
      " - 4s - loss: 0.0283\n",
      "Epoch 8/50\n",
      " - 4s - loss: 0.0285\n",
      "Epoch 9/50\n",
      " - 4s - loss: 0.0269\n",
      "Epoch 10/50\n",
      " - 4s - loss: 0.0263\n",
      "Epoch 11/50\n",
      " - 4s - loss: 0.0266\n",
      "Epoch 12/50\n",
      " - 4s - loss: 0.0266\n",
      "Epoch 13/50\n",
      " - 4s - loss: 0.0257\n",
      "Epoch 14/50\n",
      " - 4s - loss: 0.0262\n",
      "Epoch 15/50\n",
      " - 4s - loss: 0.0258\n",
      "Epoch 16/50\n",
      " - 4s - loss: 0.0258\n",
      "Epoch 17/50\n",
      " - 4s - loss: 0.0247\n",
      "Epoch 18/50\n",
      " - 4s - loss: 0.0255\n",
      "Epoch 19/50\n",
      " - 4s - loss: 0.0253\n",
      "Epoch 20/50\n",
      " - 4s - loss: 0.0250\n",
      "Epoch 21/50\n",
      " - 4s - loss: 0.0245\n",
      "Epoch 22/50\n",
      " - 4s - loss: 0.0250\n",
      "Epoch 23/50\n",
      " - 4s - loss: 0.0246\n",
      "Epoch 24/50\n",
      " - 4s - loss: 0.0253\n",
      "Epoch 25/50\n",
      " - 4s - loss: 0.0242\n",
      "Epoch 26/50\n",
      " - 4s - loss: 0.0241\n",
      "Epoch 27/50\n",
      " - 4s - loss: 0.0251\n",
      "Epoch 28/50\n",
      " - 4s - loss: 0.0243\n",
      "Epoch 29/50\n",
      " - 4s - loss: 0.0246\n",
      "Epoch 30/50\n",
      " - 4s - loss: 0.0243\n",
      "Epoch 31/50\n",
      " - 4s - loss: 0.0244\n",
      "Epoch 32/50\n",
      " - 4s - loss: 0.0243\n",
      "Epoch 33/50\n",
      " - 4s - loss: 0.0243\n",
      "Epoch 34/50\n",
      " - 4s - loss: 0.0240\n",
      "Epoch 35/50\n",
      " - 4s - loss: 0.0241\n",
      "Epoch 36/50\n",
      " - 4s - loss: 0.0242\n",
      "Epoch 37/50\n",
      " - 4s - loss: 0.0247\n",
      "Epoch 38/50\n",
      " - 4s - loss: 0.0241\n",
      "Epoch 39/50\n",
      " - 4s - loss: 0.0240\n",
      "Epoch 40/50\n",
      " - 4s - loss: 0.0244\n",
      "Epoch 41/50\n",
      " - 4s - loss: 0.0240\n",
      "Epoch 42/50\n",
      " - 4s - loss: 0.0244\n",
      "Epoch 43/50\n",
      " - 4s - loss: 0.0239\n",
      "Epoch 44/50\n",
      " - 4s - loss: 0.0242\n",
      "Epoch 45/50\n",
      " - 4s - loss: 0.0242\n",
      "Epoch 46/50\n",
      " - 4s - loss: 0.0241\n",
      "Epoch 47/50\n",
      " - 4s - loss: 0.0241\n",
      "Epoch 48/50\n",
      " - 4s - loss: 0.0244\n",
      "Epoch 49/50\n",
      " - 4s - loss: 0.0242\n",
      "Epoch 50/50\n",
      " - 4s - loss: 0.0239\n"
     ]
    }
   ],
   "source": [
    "# random seed for reproducibility\n",
    "np.random.seed(202)\n",
    "# initialise model architecture\n",
    "bt_model = build_model(LSTM_training_inputs, output_size=1, neurons = 20, dropout = 0.25)\n",
    "# train model on data\n",
    "# note: btc_history contains information on the training error per epoch\n",
    "bt_history = bt_model.fit(LSTM_training_inputs, \n",
    "                            (training_set['bt_Close'][window_len:].values/training_set['bt_Close'][:-window_len].values)-1, \n",
    "                            epochs=50, batch_size=1, verbose=2, shuffle=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1a29ec6470>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from mpl_toolkits.axes_grid1.inset_locator import zoomed_inset_axes\n",
    "from mpl_toolkits.axes_grid1.inset_locator import mark_inset\n",
    "\n",
    "fig, ax1 = plt.subplots(1,1)\n",
    "ax1.set_xticks([datetime.date(i,j,1) for i in range(2013,2019) for j in [1,5,9]])\n",
    "ax1.set_xticklabels([datetime.date(i,j,1).strftime('%b %Y')  for i in range(2013,2019) for j in [1,5,9]])\n",
    "ax1.plot(model_data[model_data['Date']< split_date]['Date'][window_len:].astype(datetime.datetime),\n",
    "         training_set['bt_Close'][window_len:], label='Actual')\n",
    "ax1.plot(model_data[model_data['Date']< split_date]['Date'][window_len:].astype(datetime.datetime),\n",
    "         ((np.transpose(bt_model.predict(LSTM_training_inputs))+1) * training_set['bt_Close'].values[:-window_len])[0], \n",
    "         label='Predicted')\n",
    "ax1.set_title('Training Set: Single Timepoint Prediction')\n",
    "ax1.set_ylabel('Bitcoin Price ($)',fontsize=12)\n",
    "ax1.annotate('MAE: %.4f'%np.mean(np.abs((np.transpose(bt_model.predict(LSTM_training_inputs))+1)-\\\n",
    "            (training_set['bt_Close'].values[window_len:])/(training_set['bt_Close'].values[:-window_len]))), \n",
    "             xy=(0.75, 0.9),  xycoords='axes fraction',\n",
    "            xytext=(0.75, 0.9), textcoords='axes fraction')\n",
    "ax1.legend(bbox_to_anchor=(0.1, 1), loc=2, borderaxespad=0., prop={'size': 14})\n",
    "# figure inset code taken from http://akuederle.com/matplotlib-zoomed-up-inset\n",
    "axins = zoomed_inset_axes(ax1, 2.52, loc=10, bbox_to_anchor=(400, 307)) # zoom-factor: 2.52, location: centre\n",
    "axins.set_xticks([datetime.date(i,j,1) for i in range(2013,2019) for j in [1,5,9]])\n",
    "axins.plot(model_data[model_data['Date']< split_date]['Date'][window_len:].astype(datetime.datetime),\n",
    "         training_set['bt_Close'][window_len:], label='Actual')\n",
    "axins.plot(model_data[model_data['Date']< split_date]['Date'][window_len:].astype(datetime.datetime),\n",
    "         ((np.transpose(bt_model.predict(LSTM_training_inputs))+1) * training_set['bt_Close'].values[:-window_len])[0], \n",
    "         label='Predicted')\n",
    "axins.set_xlim([datetime.date(2017, 2, 15), datetime.date(2017, 5, 1)])\n",
    "axins.set_ylim([920, 1400])\n",
    "axins.set_xticklabels('')\n",
    "mark_inset(ax1, axins, loc1=1, loc2=3, fc=\"none\", ec=\"0.5\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1a2a3232e8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "168\n",
      "4.84961391248\n",
      "the accuracy in percentage measured as 100 - MAPE for bitcoin using LSTM is 95.15039\n"
     ]
    }
   ],
   "source": [
    "fig, ax1 = plt.subplots(1,1)\n",
    "\n",
    "ax1.set_xticks([datetime.date(i,j,1) for i in range(2017,2019) for j in [1,5,9]])\n",
    "ax1.set_xticklabels([datetime.date(i,j,1).strftime('%b %Y')  for i in range(2017,2019) for j in [1,5,9]])\n",
    "ax1.plot(model_data[model_data['Date']>= split_date]['Date'][window_len:].astype(datetime.datetime),\n",
    "         test_set['bt_Close'][window_len:], label='Actual')\n",
    "ax1.plot(model_data[model_data['Date']>= split_date]['Date'][window_len:].astype(datetime.datetime),\n",
    "         ((np.transpose(bt_model.predict(LSTM_test_inputs))+1) * test_set['bt_Close'].values[:-window_len])[0], \n",
    "         label='Predicted')\n",
    "ax1.annotate('MAE: %.4f'%np.mean(np.abs((np.transpose(bt_model.predict(LSTM_test_inputs))+1)-\\\n",
    "            (test_set['bt_Close'].values[window_len:])/(test_set['bt_Close'].values[:-window_len]))), \n",
    "             xy=(0.75, 0.9),  xycoords='axes fraction',\n",
    "            xytext=(0.75, 0.9), textcoords='axes fraction')\n",
    "ax1.set_title('Test Set: Single Timepoint Prediction',fontsize=13)\n",
    "ax1.set_ylabel('Bitcoin Price ($)',fontsize=12)\n",
    "ax1.legend(bbox_to_anchor=(0.1, 1), loc=2, borderaxespad=0., prop={'size': 11})\n",
    "plt.show()\n",
    "\n",
    "print(len(bt_model.predict(LSTM_test_inputs)))\n",
    "\n",
    "mape_bt = (100/len(bt_model.predict(LSTM_test_inputs)))*np.sum(np.abs(((np.transpose(bt_model.predict(LSTM_test_inputs)))+1)-\\\n",
    "            (test_set['bt_Close'].values[window_len:])/(test_set['bt_Close'].values[:-window_len])))\n",
    "\n",
    "print(mape_bt) \n",
    "\n",
    "print('the accuracy in percentage measured as 100 - MAPE for bitcoin using LSTM is %.5f'%(100-mape_bt))\n",
    "\n",
    "#"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "After getting the accuracy of LSTM for this time period, the next step is to check the trading strategy based in this algorithm, to calculate the sharp ratio. For that reason, the predicted prices are stored in the files below, in order to be able to work with them in R. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[  297.5556577 ]\n",
      " [  306.17867344]\n",
      " [  308.77653723]\n",
      " [  308.80544155]\n",
      " [  298.81066159]\n",
      " [  300.61905573]\n",
      " [  305.61567433]\n",
      " [  306.1067539 ]\n",
      " [  338.45396598]\n",
      " [  340.76350522]\n",
      " [  334.66166045]\n",
      " [  328.80281329]\n",
      " [  319.33842062]\n",
      " [  318.29557115]\n",
      " [  307.35152789]\n",
      " [  303.88349283]\n",
      " [  301.66865892]\n",
      " [  297.17611341]\n",
      " [  289.59591329]\n",
      " [  297.58479323]\n",
      " [  298.48196111]\n",
      " [  293.74818624]\n",
      " [  294.13120031]\n",
      " [  296.36480288]\n",
      " [  304.94881866]\n",
      " [  307.11837254]\n",
      " [  307.43021603]\n",
      " [  296.04079161]\n",
      " [  291.07846252]\n",
      " [  307.50973907]\n",
      " [  301.89713888]\n",
      " [  296.27088405]\n",
      " [  294.09348629]\n",
      " [  294.28686662]\n",
      " [  305.47288556]\n",
      " [  317.91910486]\n",
      " [  302.22850327]\n",
      " [  310.02192073]\n",
      " [  321.83361425]\n",
      " [  312.38022258]\n",
      " [  334.98477378]\n",
      " [  335.03109538]\n",
      " [  333.70755951]\n",
      " [  332.10804811]\n",
      " [  347.85806297]\n",
      " [  349.46244973]\n",
      " [  365.37752596]\n",
      " [  359.28596226]\n",
      " [  381.52247538]\n",
      " [  403.83700275]\n",
      " [  476.75816321]\n",
      " [  485.19926138]\n",
      " [  467.2596585 ]\n",
      " [  472.27730461]\n",
      " [  463.79257278]\n",
      " [  430.18691303]\n",
      " [  447.00140317]\n",
      " [  466.19893932]\n",
      " [  457.28023828]\n",
      " [  453.85501526]\n",
      " [  464.14579897]\n",
      " [  451.87689526]\n",
      " [  426.53889403]\n",
      " [  431.31903913]\n",
      " [  455.82706481]\n",
      " [  471.63207673]\n",
      " [  448.44794352]\n",
      " [  511.15869705]\n",
      " [  668.09320693]\n",
      " [  715.13570523]\n",
      " [  700.60476482]\n",
      " [  651.8003351 ]\n",
      " [  688.70986064]\n",
      " [  720.91385772]\n",
      " [  794.66647611]\n",
      " [  817.82680993]\n",
      " [  828.72572778]\n",
      " [  792.18709491]\n",
      " [  713.8583554 ]\n",
      " [  727.28983615]\n",
      " [  736.6602008 ]\n",
      " [  764.43328236]\n",
      " [  767.43333302]\n",
      " [  758.63842171]\n",
      " [  725.93302996]\n",
      " [  746.34228075]\n",
      " [  729.64382079]\n",
      " [  769.86565822]\n",
      " [  764.98589613]\n",
      " [  878.35463525]\n",
      " [  978.33312237]\n",
      " [  988.25275072]\n",
      " [  975.43335247]\n",
      " [ 1031.43934205]\n",
      " [ 1158.58638273]\n",
      " [ 1125.49768969]\n",
      " [ 1273.78414842]\n",
      " [ 1284.02515682]\n",
      " [ 1195.63500554]\n",
      " [ 1242.99179936]\n",
      " [ 1372.64039885]\n",
      " [ 1348.37267504]\n",
      " [ 1269.61733363]\n",
      " [ 1146.90469387]\n",
      " [ 1163.96696406]\n",
      " [ 1092.44437322]\n",
      " [ 1064.24738527]\n",
      " [ 1140.11457977]\n",
      " [ 1050.92285395]\n",
      " [ 1001.24829437]\n",
      " [  998.36956338]\n",
      " [ 1076.2735563 ]\n",
      " [ 1045.96016236]\n",
      " [ 1048.40462378]\n",
      " [ 1090.15027487]\n",
      " [ 1239.97123182]\n",
      " [ 1196.77948185]\n",
      " [ 1099.7910975 ]\n",
      " [ 1113.89696475]\n",
      " [ 1093.41662466]\n",
      " [ 1017.2521799 ]\n",
      " [  977.05417678]\n",
      " [  903.12871605]\n",
      " [  818.34177127]\n",
      " [  841.37641209]\n",
      " [  804.34346545]\n",
      " [  814.6689825 ]\n",
      " [  876.01269525]\n",
      " [  850.66194231]\n",
      " [  794.56797003]\n",
      " [  859.27146315]\n",
      " [  845.15581233]\n",
      " [  918.71424552]\n",
      " [  927.81465692]\n",
      " [  935.24838371]\n",
      " [  968.42736171]\n",
      " [  912.78470322]\n",
      " [  933.30512206]\n",
      " [  898.76890999]\n",
      " [  866.94588372]\n",
      " [  825.00417457]\n",
      " [  877.92366426]\n",
      " [  846.43231133]\n",
      " [  846.85508776]\n",
      " [  860.27192184]\n",
      " [  868.86935064]\n",
      " [  851.49505424]\n",
      " [  859.5651063 ]\n",
      " [  847.66267137]\n",
      " [  844.90764607]\n",
      " [  859.06953011]\n",
      " [  841.79358736]\n",
      " [  814.23057333]\n",
      " [  772.99159268]\n",
      " [  746.31407433]\n",
      " [  769.60283791]\n",
      " [  706.31656592]\n",
      " [  730.91739631]\n",
      " [  697.17342598]\n",
      " [  694.40338647]\n",
      " [  636.23618923]\n",
      " [  640.06220546]\n",
      " [  610.56609538]\n",
      " [  576.24453218]\n",
      " [  549.62070865]\n",
      " [  553.87708943]\n",
      " [  551.07471236]]\n"
     ]
    }
   ],
   "source": [
    "eth_pred_prices = ((eth_model.predict(LSTM_test_inputs)[:-1][::1]+1)*\\\n",
    "                   test_set['eth_Close'].values[:-(window_len + 1)][::1].reshape(int(np.ceil((len(LSTM_test_inputs)-1)/float(1))),1))\n",
    "\n",
    "np.savetxt(\"eth_pred_prices_lstm_2_def.csv\", eth_pred_prices, delimiter=\"\") #we save the predicted values in order to implement the rest of the backtesting for the trading strategy in R, as the language is more flexible. \n",
    "\n",
    "print(eth_pred_prices)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[  4298.85821207]\n",
      " [  4337.0090817 ]\n",
      " [  4454.58917051]\n",
      " [  4675.17977733]\n",
      " [  4789.02976215]\n",
      " [  4769.47630298]\n",
      " [  4833.05727149]\n",
      " [  5427.03311956]\n",
      " [  5678.87245831]\n",
      " [  6051.17599961]\n",
      " [  5727.19020354]\n",
      " [  5769.35954402]\n",
      " [  5575.21194475]\n",
      " [  5614.69739931]\n",
      " [  5669.99553339]\n",
      " [  5956.34772466]\n",
      " [  6006.00161015]\n",
      " [  6083.15556857]\n",
      " [  5918.31195896]\n",
      " [  5476.10460954]\n",
      " [  5681.60776677]\n",
      " [  5802.04988624]\n",
      " [  5736.38367706]\n",
      " [  5733.81929612]\n",
      " [  6100.17485485]\n",
      " [  6155.70232618]\n",
      " [  6552.95394363]\n",
      " [  6863.71457772]\n",
      " [  7060.62171307]\n",
      " [  7196.99901024]\n",
      " [  7496.49801717]\n",
      " [  7463.02423596]\n",
      " [  6992.28112373]\n",
      " [  7019.104026  ]\n",
      " [  7352.25387975]\n",
      " [  7087.5192653 ]\n",
      " [  6588.56546406]\n",
      " [  6290.85965637]\n",
      " [  5970.19489808]\n",
      " [  6377.5133997 ]\n",
      " [  6631.12049181]\n",
      " [  7336.05603418]\n",
      " [  8013.13489712]\n",
      " [  7864.51651735]\n",
      " [  7845.89397573]\n",
      " [  8118.44255421]\n",
      " [  8189.83036418]\n",
      " [  8092.85940025]\n",
      " [  8227.78484219]\n",
      " [  7903.506652  ]\n",
      " [  8203.08886913]\n",
      " [  8832.10296883]\n",
      " [  9428.16110396]\n",
      " [ 10179.26512079]\n",
      " [ 10296.30660125]\n",
      " [  9748.50960303]\n",
      " [  9953.22230232]\n",
      " [ 10885.31473955]\n",
      " [ 11082.30584602]\n",
      " [ 11337.37026067]\n",
      " [ 11847.69553986]\n",
      " [ 12068.29957991]\n",
      " [ 14001.50443878]\n",
      " [ 17186.95063477]\n",
      " [ 17016.70674834]\n",
      " [ 15640.98127789]\n",
      " [ 15325.06616861]\n",
      " [ 16743.24724762]\n",
      " [ 17348.65051842]\n",
      " [ 16548.09249122]\n",
      " [ 16571.76720643]\n",
      " [ 17670.47290306]\n",
      " [ 19393.44300745]\n",
      " [ 19156.40755212]\n",
      " [ 19157.27076244]\n",
      " [ 17749.47960162]\n",
      " [ 16597.96929479]\n",
      " [ 15564.1108449 ]\n",
      " [ 14900.36440279]\n",
      " [ 14461.64508054]\n",
      " [ 13975.25632111]\n",
      " [ 13734.97817601]\n",
      " [ 15901.48947387]\n",
      " [ 15951.64473209]\n",
      " [ 14735.95454721]\n",
      " [ 14437.79305043]\n",
      " [ 13071.45672498]\n",
      " [ 13825.84784049]\n",
      " [ 13499.20364895]\n",
      " [ 14700.39603767]\n",
      " [ 15496.45427077]\n",
      " [ 15923.48276625]\n",
      " [ 17548.98381944]\n",
      " [ 17891.54581871]\n",
      " [ 16475.94673419]\n",
      " [ 15241.08223658]\n",
      " [ 14274.12896121]\n",
      " [ 14275.55613846]\n",
      " [ 13479.24969492]\n",
      " [ 13796.96547396]\n",
      " [ 14068.82924569]\n",
      " [ 13630.24164448]\n",
      " [ 13777.20068192]\n",
      " [ 12071.78476703]\n",
      " [ 11623.30112052]\n",
      " [ 11189.93305516]\n",
      " [ 11236.16741073]\n",
      " [ 12792.78145045]\n",
      " [ 11829.92520278]\n",
      " [ 10898.25351618]\n",
      " [ 10692.97184833]\n",
      " [ 11087.43507805]\n",
      " [ 11055.71404619]\n",
      " [ 11153.73648362]\n",
      " [ 11450.12690806]\n",
      " [ 11727.43521748]\n",
      " [ 11348.82843006]\n",
      " [ 10195.55345145]\n",
      " [ 10061.73527498]\n",
      " [  9248.18033614]\n",
      " [  9034.53392057]\n",
      " [  9058.15277945]\n",
      " [  8319.15914568]\n",
      " [  7363.30001894]\n",
      " [  7766.4654163 ]\n",
      " [  7471.36437204]\n",
      " [  8071.98631146]\n",
      " [  8866.60598986]\n",
      " [  8707.01381278]\n",
      " [  8152.97797436]\n",
      " [  8807.46913262]\n",
      " [  8539.32692472]\n",
      " [  9441.57893243]\n",
      " [ 10360.07293705]\n",
      " [ 10465.40529869]\n",
      " [ 11269.43232433]\n",
      " [ 10809.029349  ]\n",
      " [ 11184.08007774]\n",
      " [ 11400.61057845]\n",
      " [ 10694.53349586]\n",
      " [  9940.16700449]\n",
      " [ 10126.86932557]\n",
      " [  9643.12787519]\n",
      " [  9523.87235165]\n",
      " [ 10257.70216227]\n",
      " [ 10687.01327566]\n",
      " [ 10477.03492992]\n",
      " [ 10862.25510615]\n",
      " [ 11153.19472563]\n",
      " [ 11537.65114346]\n",
      " [ 11669.67006873]\n",
      " [ 11669.0779227 ]\n",
      " [ 10766.92719812]\n",
      " [  9968.90365979]\n",
      " [  9343.13998032]\n",
      " [  9396.69154267]\n",
      " [  8679.29483083]\n",
      " [  9399.12459994]\n",
      " [  9113.95962914]\n",
      " [  9072.64961602]\n",
      " [  8318.78824544]\n",
      " [  8247.02688073]\n",
      " [  8124.04176334]\n",
      " [  7780.34436944]\n",
      " [  8193.82652268]\n",
      " [  8576.34585204]\n",
      " [  8883.89879653]]\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "167"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "bt_pred_prices = ((bt_model.predict(LSTM_test_inputs)[:-1][::1]+1)*\\\n",
    "                   test_set['bt_Close'].values[:-(window_len + 1)][::1].reshape(int(np.ceil((len(LSTM_test_inputs)-1)/float(1))),1))\n",
    "\n",
    "np.savetxt(\"btc_pred_prices_lstm_2_def.csv\", bt_pred_prices, delimiter=\"\")\n",
    "\n",
    "print(bt_pred_prices)\n",
    "len(bt_pred_prices)"
   ]
  }
 ],
 "metadata": {
  "anaconda-cloud": {},
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.5.5"
  },
  "widgets": {
   "state": {
    "200ff45abb3b47f3ade09c6ce91695fb": {
     "views": [
      {
       "cell_index": 21
      }
     ]
    }
   },
   "version": "1.2.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}