Results_LSTM.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "#modules to load saved models\n",
    "from joblib import dump, load\n",
    "from tensorflow.keras.models import load_model\n",
    "\n",
    "#data preprocessing modules\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "from sklearn.preprocessing import StandardScaler, MinMaxScaler, RobustScaler\n",
    "from sklearn.pipeline import Pipeline\n",
    "from sklearn.model_selection import train_test_split\n",
    "from keras.utils import to_categorical\n",
    "\n",
    "#regression metrics\n",
    "from sklearn.metrics import mean_absolute_error\n",
    "from sklearn.metrics import r2_score\n",
    "from sklearn.metrics import mean_squared_error\n",
    "from commons import mean_absolute_percentage_error #keep commons.py in notebooks folder\n",
    "\n",
    "#classification metrics\n",
    "from sklearn.metrics import accuracy_score, recall_score, precision_score, roc_auc_score, f1_score\n",
    "\n",
    "#modules for plots if required\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "from keras.utils import plot_model\n",
    "from IPython.display import SVG\n",
    "from keras.utils import model_to_dot "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Regression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "#file path\n",
    "data=pd.read_csv('datasets/reg_interval1.csv',sep=',')\n",
    "\n",
    "# 'datasets/reg_interval1.csv'\n",
    "# 'datasets/reg_interval2.csv'\n",
    "# 'datasets/reg_interval3.csv'\n",
    "# 'datasets/reg_seven.csv'\n",
    "# 'datasets/reg_thirty.csv'\n",
    "# 'datasets/reg_ninety.csv'"
   ]
  },
  {
   "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>difficulty</th>\n",
       "      <th>fee_to_rewardUSD</th>\n",
       "      <th>hashrate3std</th>\n",
       "      <th>mediantransactionvalue90smaUSD</th>\n",
       "      <th>mediantransactionvalueUSD</th>\n",
       "      <th>mining_profitability7rsi</th>\n",
       "      <th>price30wmaUSD</th>\n",
       "      <th>price3wmaUSD</th>\n",
       "      <th>price7wmaUSD</th>\n",
       "      <th>price90momUSD</th>\n",
       "      <th>sentinusd30varUSD</th>\n",
       "      <th>transactionfeesUSD</th>\n",
       "      <th>priceUSD</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>4.123083e+08</td>\n",
       "      <td>1.079</td>\n",
       "      <td>1.613338e+14</td>\n",
       "      <td>64.903</td>\n",
       "      <td>231.380</td>\n",
       "      <td>51.766</td>\n",
       "      <td>181.591</td>\n",
       "      <td>220.730</td>\n",
       "      <td>210.518</td>\n",
       "      <td>134.013</td>\n",
       "      <td>6.583452e+15</td>\n",
       "      <td>0.2130</td>\n",
       "      <td>231.403</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>4.397166e+10</td>\n",
       "      <td>0.487</td>\n",
       "      <td>3.917670e+16</td>\n",
       "      <td>99.888</td>\n",
       "      <td>109.132</td>\n",
       "      <td>33.611</td>\n",
       "      <td>293.321</td>\n",
       "      <td>254.923</td>\n",
       "      <td>267.518</td>\n",
       "      <td>-154.347</td>\n",
       "      <td>3.962830e+15</td>\n",
       "      <td>0.0363</td>\n",
       "      <td>239.441</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>4.397166e+10</td>\n",
       "      <td>0.396</td>\n",
       "      <td>5.429477e+16</td>\n",
       "      <td>99.292</td>\n",
       "      <td>88.082</td>\n",
       "      <td>41.201</td>\n",
       "      <td>247.544</td>\n",
       "      <td>222.808</td>\n",
       "      <td>215.902</td>\n",
       "      <td>-123.405</td>\n",
       "      <td>6.909133e+15</td>\n",
       "      <td>0.0333</td>\n",
       "      <td>232.311</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     difficulty  fee_to_rewardUSD  hashrate3std  \\\n",
       "0  4.123083e+08             1.079  1.613338e+14   \n",
       "1  4.397166e+10             0.487  3.917670e+16   \n",
       "2  4.397166e+10             0.396  5.429477e+16   \n",
       "\n",
       "   mediantransactionvalue90smaUSD  mediantransactionvalueUSD  \\\n",
       "0                          64.903                    231.380   \n",
       "1                          99.888                    109.132   \n",
       "2                          99.292                     88.082   \n",
       "\n",
       "   mining_profitability7rsi  price30wmaUSD  price3wmaUSD  price7wmaUSD  \\\n",
       "0                    51.766        181.591       220.730       210.518   \n",
       "1                    33.611        293.321       254.923       267.518   \n",
       "2                    41.201        247.544       222.808       215.902   \n",
       "\n",
       "   price90momUSD  sentinusd30varUSD  transactionfeesUSD  priceUSD  \n",
       "0        134.013       6.583452e+15              0.2130   231.403  \n",
       "1       -154.347       3.962830e+15              0.0363   239.441  \n",
       "2       -123.405       6.909133e+15              0.0333   232.311  "
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#visualize the data\n",
    "data.head(3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "#create train-test splits for ANN\n",
    "length=data.shape[1]-1\n",
    "X=data.iloc[:,:length]\n",
    "y=data.iloc[:,length:]\n",
    "X_train, X_test, y_train, y_test = train_test_split(X,y,test_size=0.2, shuffle=True, random_state=7)\n",
    "y_train=np.ravel(y_train)\n",
    "y_test=np.ravel(y_test)\n",
    "estimators=[]\n",
    "estimators.append(['robust',RobustScaler()])\n",
    "estimators.append(['mixmax',MinMaxScaler()])\n",
    "scaling=Pipeline(estimators)\n",
    "X_train=scaling.fit_transform(X_train)\n",
    "X_test=scaling.transform(X_test)\n",
    "X_train=np.reshape(X_train,(X_train.shape[0],1,X_train.shape[1]))\n",
    "X_test=np.reshape(X_test,(X_test.shape[0],1,X_test.shape[1]))\n",
    "y_train=np.reshape(y_train, (y_train.shape[0],1,1))\n",
    "y_test=np.reshape(y_test,(y_test.shape[0],1,1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:From /home/mudassir/btcpaper/lib/python3.6/site-packages/tensorflow/python/ops/resource_variable_ops.py:435: colocate_with (from tensorflow.python.framework.ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Colocations handled automatically by placer.\n"
     ]
    }
   ],
   "source": [
    "#load saved model\n",
    "LSTM=load_model('trained_models/LSTM_reg_interval1.hdf5',compile=False)\n",
    "\n",
    "# 'trained_models/LSTM_reg_interval1.hdf5'\n",
    "# 'trained_models/LSTM_reg_interval2.hdf5'\n",
    "# 'trained_models/LSTM_reg_interval3.hdf5'\n",
    "# 'trained_models/LSTM_reg_seven.hdf5'\n",
    "# 'trained_models/LSTM_reg_thirty.hdf5'\n",
    "# 'trained_models/LSTM_reg_ninety.hdf5'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'name': 'sequential_1',\n",
       " 'layers': [{'class_name': 'Bidirectional',\n",
       "   'config': {'name': 'bidirectional_1',\n",
       "    'trainable': True,\n",
       "    'batch_input_shape': (None, 1, 12),\n",
       "    'dtype': 'float32',\n",
       "    'layer': {'class_name': 'LSTM',\n",
       "     'config': {'name': 'lstm_1',\n",
       "      'trainable': True,\n",
       "      'dtype': 'float32',\n",
       "      'return_sequences': True,\n",
       "      'return_state': False,\n",
       "      'go_backwards': False,\n",
       "      'stateful': False,\n",
       "      'unroll': False,\n",
       "      'time_major': False,\n",
       "      'zero_output_for_mask': True,\n",
       "      'units': 300,\n",
       "      'activation': 'relu',\n",
       "      'recurrent_activation': 'sigmoid',\n",
       "      'use_bias': True,\n",
       "      'kernel_initializer': {'class_name': 'VarianceScaling',\n",
       "       'config': {'scale': 1.0,\n",
       "        'mode': 'fan_avg',\n",
       "        'distribution': 'uniform',\n",
       "        'seed': None,\n",
       "        'dtype': 'float32'}},\n",
       "      'recurrent_initializer': {'class_name': 'Orthogonal',\n",
       "       'config': {'gain': 1.0, 'seed': None, 'dtype': 'float32'}},\n",
       "      'bias_initializer': {'class_name': 'Zeros',\n",
       "       'config': {'dtype': 'float32'}},\n",
       "      'unit_forget_bias': True,\n",
       "      'kernel_regularizer': None,\n",
       "      'recurrent_regularizer': None,\n",
       "      'bias_regularizer': None,\n",
       "      'activity_regularizer': None,\n",
       "      'kernel_constraint': None,\n",
       "      'recurrent_constraint': None,\n",
       "      'bias_constraint': None,\n",
       "      'dropout': 0.0,\n",
       "      'recurrent_dropout': 0.0,\n",
       "      'implementation': 2}},\n",
       "    'merge_mode': 'concat'}},\n",
       "  {'class_name': 'Bidirectional',\n",
       "   'config': {'name': 'bidirectional_2',\n",
       "    'trainable': True,\n",
       "    'dtype': 'float32',\n",
       "    'layer': {'class_name': 'LSTM',\n",
       "     'config': {'name': 'lstm_2',\n",
       "      'trainable': True,\n",
       "      'dtype': 'float32',\n",
       "      'return_sequences': True,\n",
       "      'return_state': False,\n",
       "      'go_backwards': False,\n",
       "      'stateful': False,\n",
       "      'unroll': False,\n",
       "      'time_major': False,\n",
       "      'zero_output_for_mask': True,\n",
       "      'units': 300,\n",
       "      'activation': 'relu',\n",
       "      'recurrent_activation': 'sigmoid',\n",
       "      'use_bias': True,\n",
       "      'kernel_initializer': {'class_name': 'VarianceScaling',\n",
       "       'config': {'scale': 1.0,\n",
       "        'mode': 'fan_avg',\n",
       "        'distribution': 'uniform',\n",
       "        'seed': None,\n",
       "        'dtype': 'float32'}},\n",
       "      'recurrent_initializer': {'class_name': 'Orthogonal',\n",
       "       'config': {'gain': 1.0, 'seed': None, 'dtype': 'float32'}},\n",
       "      'bias_initializer': {'class_name': 'Zeros',\n",
       "       'config': {'dtype': 'float32'}},\n",
       "      'unit_forget_bias': True,\n",
       "      'kernel_regularizer': None,\n",
       "      'recurrent_regularizer': None,\n",
       "      'bias_regularizer': None,\n",
       "      'activity_regularizer': None,\n",
       "      'kernel_constraint': None,\n",
       "      'recurrent_constraint': None,\n",
       "      'bias_constraint': None,\n",
       "      'dropout': 0.0,\n",
       "      'recurrent_dropout': 0.0,\n",
       "      'implementation': 2}},\n",
       "    'merge_mode': 'concat'}},\n",
       "  {'class_name': 'Dense',\n",
       "   'config': {'name': 'dense_1',\n",
       "    'trainable': True,\n",
       "    'dtype': 'float32',\n",
       "    'units': 1,\n",
       "    'activation': 'linear',\n",
       "    'use_bias': True,\n",
       "    'kernel_initializer': {'class_name': 'VarianceScaling',\n",
       "     'config': {'scale': 1.0,\n",
       "      'mode': 'fan_avg',\n",
       "      'distribution': 'uniform',\n",
       "      'seed': None,\n",
       "      'dtype': 'float32'}},\n",
       "    'bias_initializer': {'class_name': 'Zeros',\n",
       "     'config': {'dtype': 'float32'}},\n",
       "    'kernel_regularizer': None,\n",
       "    'bias_regularizer': None,\n",
       "    'activity_regularizer': None,\n",
       "    'kernel_constraint': None,\n",
       "    'bias_constraint': None}}]}"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#show model configuration\n",
    "LSTM.get_config()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABHcAAABoCAYAAAB7YBxGAAAABmJLR0QA/wD/AP+gvaeTAAAgAElEQVR4nO3deVxU9f4/8NewqkgoauKCCyphYZsPI5fMTMWHudxKoFyILDG719RI1MT79XrFpZsLV2mxzDCXAOtetzIzK8stWywtt5uhoiiIsgiyv39/+JsTwwxwwBnOnJnX8/HgUXzmLO+Z9/u858zHMweDiAiIiIiIiIiIiEiPUl20joCIiIiIiIiIiOqPkztERERERERERDrGyR0iIiIiIiIiIh3j5A4RERERERERkY65VR04cOAAli1bpkUsRGQlqampWodQL+w/RM7h5ZdfRu/evbUOo17CwsK0DoGI7ATPt4hIK5b6j9mVO+fPn8fmzZsbJCAisq709HRdH7/sP0SOb/PmzTh//rzWYdTb5s2bkZ6ernUYRKQhnm8RkVZq6j9mV+4Y6XUmmsiZpaSkICIiQuswbhn7D5HjMhgMWodwy6ZPn47w8HCtwyAijfB8i4i0UlP/4T13iIiIiIiIiIh0jJM7REREREREREQ6xskdIiIiIiIiIiId4+QOEREREREREZGOcXKHiIiIiIiIiEjHOLlDRES1ys3N1ToEu5CZmYnU1FQsXLhQ61CIqBbsWzexbxEROQdO7tjAt99+i9mzZ8NgMMBgMOCZZ57B1q1btQ4LX331FcLDw5W4XnjhBezfv1/rsIjICqZOnYqWLVvCYDDA3d0dI0aMwNChQ9GrVy8MHToUmzdvtrjegw8+iNjYWIuPFRcXY+HChejTpw9atGiher2GoMX+T5w4gfnz5yM8PBwffPBBg+6byBGxb9ke+xaRfdu9ezeGDRumfD4bOHAgBg4ciF69emHUqFFYs2YNSkpKtA6TdIKTOzbQr18/LFq0CB07dgQAvPXWWxg5cqQmsaSnpyv/P2DAACQlJQEAOnbsiLfeegt9+vTRJC4isq6EhAQcO3YMABAQEIBt27Zh586d2LdvH7p3746wsDAsXbrUbL3OnTujUaNGFrfp6emJl19+GSdPnkR5ebnq9aytch/TYv9GQUFBFl9DIqof9i3bY98ism+DBg3Cu+++C+Bmj9izZw/27NmD7777DhMnTsSiRYsQHByM3377TeNISQ/ctA7AkTVu3Njkvw0tLS0NkZGR2Lt3r93ERES24+fnBwBwcflz3t7DwwP/+te/8Pbbb+ONN95ATEyMyTqbNm2qcZuNGjXC7bffjqtXr9ZpPWux1Mcacv9VeXp6arJfIkfFvmV77FtE9q1t27YATI9Vg8GA4cOHo2fPnujZsydGjhyJY8eONfgEMekLr9xxUBcuXMDw4cORlZWldShEpDE3Nzd4e3sjLy9P61DqhH2MyHmxbxERAW3atME///lP/P7777wKj2rFyZ0GtHXrVkyaNAn+/v7IyclBVFQUWrZsiR49euCHH34AABw8eBCvvPIKOnfujMuXL2P06NFo0aIFevTogY8//hgA8M4778DFxQUGgwEAkJ+fj2XLlpmMvf/++/j1119x6dIlTJ48uV7xnj59GmFhYZg1axYiIyPRv39/HD16FACwYcMGeHl5wWAwYMmSJcqlzxs3boSnp6fy9a+ioiK89tpreP7559GrVy8MHjwYx44dQ0VFBb7++mtMnz4dnTt3xsWLFzFgwAB07NgROTk59X+RicjM5s2bkZmZiQkTJihjFRUVSE1NRVRUFB5++GFl/MaNG4iJicGkSZMwd+5cvPrqqygoKKh2vdqO5ep6gFFBQQEWLFiA8ePHY+rUqRgwYAASEhIAWO5j1cWdl5eHmTNnYvbs2YiJiUFoaChiYmKUfqKm/wI19z0iajjsW+xbRHTT6NGj4erqil27diljNfUptb3j+++/x4MPPoi//e1v+Pvf/w53d3eld9bWB8lOSRXJycliYZjqISgoyOS1TE9Pl6ZNmwoAiY+Pl7Nnz8r69esFgISEhEh5ebls375dGjduLABkypQpsnfvXtm4caN4e3sLANm3b5+IiHTp0sUsT1XHAEhQUJBZXNWNV9WtWzfp0qWLiIiUlpZKs2bNJDg4WHk8Li5OAMivv/6qjJ07d04ef/xx5feJEyfKiRMnlN+HDBkirVu3litXrsj+/fulSZMmAkAWLVoku3fvlueff16uX79ea2xkmd6PX73Hbw8AiI+Pj0RFRcm4ceOkT58+0rx5c1m9erVUVFSYLHvu3DmTflBWViYhISEyceJEZZnff/9d3NzcTPJSeb3i4uIaj+XqekBeXp6UlpbKgAEDZPz48Upsa9euFQCybds25flU7VdV487Pz5fAwECZN2+eskxmZqYEBgZKQECA5OTk1Np/jWrre9XFROoBkOTkZK3DqDe9x2+P2LduYt/SD72fr+g9fkdU2zHapk0badGihfJ7TX1Kbe8IDAwUX19f5feIiAjJzMysdfukrRqO3xRO7thQ1ckdEZE77rjDbKx169bi6emp/B4YGCgApKCgQBlbsWKFAJCnnnqq2m1XHbvVyZ1ly5bJpk2bRESkoqJCunTpIu7u7srj2dnZ4u3tbXJCtWjRItm+fbuIiBw6dEgAWPwxLmN8Pa5evVprPFQ7vR+/eo/fHgCQrl27ytmzZ+X48eOya9cumTx5sjRq1EhiYmKkvLzcbHljP1i1apUAkOPHj5ssY+xJ1a0nYvlYrq0HLFu2TADIyZMnlXXKyspk7dq1cu3aNYv7sbT/OXPmCADJyMgwWWbdunUCQGJjY01irKxq/62t79UUE6mj98kRvcdvj9i3/sS+pQ96P1/Re/yOqLZj1N/fX9q2bSsidfuMVVnV3tGqVSsBIAkJCVJRUSHHjh2TvLw8Vdsn7dQ0ucMbKjcw49emKmvevDkuX76s/G68qWCTJk2UsZEjR2LatGk4ffq07YP8/6ZPn46CggK88cYbuHr1KoqLi1FaWqo87uvriylTpuD111/HvHnz0LZtW3zxxReYMWMGAODw4cMIDg6u8dJg4+vRvHlz2z4ZIifi5uaGDh06ALj5l1IGDx6MO++8E1OmTEGrVq0wc+ZMi+sZL/ft1KmTyXjlG51Wx9KxXFsPWL16NQCgffv2ypirqyuioqJq3V9l+/btAwB4e3ubjPfv3x8AsH//fpMYK6vaf2vre0RkG+xbN7FvEVFVpaWluHz5MgYNGgSgbp+xKqvaO9588008++yzmDp1Kj744AOsWrUK3t7eqrZP9on33NEJ413U/f39bb6vrKwslJWV4fDhw+jRowcCAgIQFxeHpk2bmi378ssvw8PDAytWrMAPP/yABx54AK6urgCA7OxsnDlzBoWFhWbrVVRU2Px5ENGfwsLCAABbtmypdpkLFy4AuHnsWkNtPcB4gnGrk9bGD3FpaWkm461btwYA+Pj4qN6Wmr5HRA2DfUsd9i0ix7Znzx6UlJTg0UcfBWC9z1hPPvkkjhw5gtDQUHz//fd46KGHkJSUxM9wOsbJHZ0wnrQYZ2yNs7ElJSUAABFBbm6uyToGgwFlZWV13teLL74IV1dXREZGorS0FEOHDgVg+WBu0aIFJk+ejLfeegv//ve/TW58GBQUhMLCQixZssRknePHj2PVqlV1jouI6s/4gaRNmzbVLhMUFAQA2LFjh1X2WVsPuOeeewAA8fHxEBHl8bNnz+LTTz8FoK6PGf+lu2rc58+fB/Bn31RDTd8joobBvqUO+xaR4yopKcGrr76K++67Dy+99BIA633G+r//+z8EBARg586d2LRpE0pLSxEXF8fPcHpWh+9wUR116NDB7N45nTp1Mnt927VrJwCktLRURP68d05ZWZmyTFJSkvTs2VNZ5vHHHxcAMnfuXDl9+rQsX75cfH19BYDs3LlTysvLpWvXruLl5SXnzp1TtnPx4kUBIO3atTO7SWFubq5ER0fLuHHjRETEx8dHDAaD7Nq1SzZs2CC33367AJBDhw7J+fPnlfUuXboknp6eMmDAAJPtFRUVSUBAgACQCRMmyIYNGyQuLk6GDBmi3IzL+HrwJsrWoffjV+/xa+3GjRsCQDp06GAyfvnyZenTp494eHjId999p4zn5+cLAOU73EeOHBE3Nzdp0aKF7Ny5UwoLC2XPnj1y2223CQD5448/LK4nYvlYrq0HnDlzRry8vASADBw4UBITE2Xu3LkyadIkpT9Z6mNV919YWCjBwcHSvn17k/tXTJ06Vfr27av0TTX9t7a+V1hYKACkU6dO9UsS6f6eNXqP396wb7Fv6ZHez1f0Hr+jqe4Y/fHHH6V///7SuXNn+e2335TxunzGqqxq72jSpIlyr7DS0lLx8fGRkJAQVdsn7fCGyg3sm2++kVmzZik3nho7dqxs2bJFEhMTlbEFCxZIbm6ucqNkADJr1iy5ceOGMrnz+uuvy5UrVyQzM1MWL15scvJx6tQpCQkJES8vLxkyZIicOnVKHnroIRk/frx8+OGHUlxcLLNnz5Y2bdrIRx99JCIie/bskVGjRin7CwoKkkceeUQeeeQRueOOO8TT01MASFJSkoiIJCYmio+PjzzwwANy8OBBSUhIkObNm8uoUaMkOzvb5DkPHz5cPvjgA7PXIi0tTUaOHCm+vr7i5+cn0dHRkpWVJQUFBTJ//nwllujoaPnpp59smBXnoPfjV+/xa+mjjz6S0aNHK8dUSEiIDB06VPr06SPdu3eXp59+Wo4dO6YsX1BQILNnz1aWX7ZsmeTl5cnevXulb9++4u3tLQEBAbJ48WLp37+/vPDCC/LFF19Ifn6+yXoLFiyQGTNmVHssV9cDjI4ePSqhoaHSvHlzadeunUybNk1yc3OVx6v2serizs/Pl9jYWBkyZIjExMRIbGyszJ8/X4qLi0VEVPffmvre999/Ly+99JKyzooVK5STIlJP75Mjeo/fnrBvsW/pld7PV/QevyP59ttv5bnnnlOO0QEDBkhoaKiMHDlSnnzySUlMTLT4j+A19Sm1vQOA3H///bJ48WIZO3asDB8+XJkQr60PknZqmtwxiFS6phRASkoKIiIiUGWYGlD37t1x4sQJ3eSgsLAQ99xzD3755Rc0btxY63Ccmt6PX73HT0S1MxgMSE5ORnh4uNah1Ive4yeiW6f38xW9x0/kzGo4flN5zx26ZYmJiZgyZQondoiIiIiIiIg0wD+FbocKCgqU/3p5eWkcjWWHDh1CdHQ0CgsLUV5ejhMnTmgdEhEREREREZFT4pU7dqSgoABz5sxR/lLCSy+9hIMHD2oclWVeXl7Iy8uDi4sLNm7cCA8PD61DIiIiIiIiInJKvHLHjnh5eSE+Ph7x8fFah1Kr4OBg/PHHH1qHQUREREREROT0eOUOEREREREREZGOcXKHiIiIiIiIiEjHOLlDRERERERERKRjnNwhIiIiIiIiItIxTu4QEREREREREelYtX8ty2AwNGQcREQK9h8ismcRERGIiIjQOgwiolvC8y0ix1Lt5E5ycnJDxkE2sHz5cgDA9OnTNY6EGsqBAwewYsUKrcO4Zew/jslYn8yvc3OESZFp06ahd+/eWodBDSAiIoL5JjM83yJ7wc97zqem/lPt5E54eLjNAqKGkZqaCoC5dDaOcLLBmnVcK1asYH6dnCNM7vTu3Zt17CQiIiKYb7KI51tkD/h5zzlV1394zx0iIiIiIiIiIh3j5A4RERERERERkY5xcoeIiIiIiIiISMc4uUNEREREREREpGOc3CEiIiIiIiIi0jFO7hARERERERER6ZiuJncefPBBxMbGah0GETk59iIi0gP2KiLSG/YtovrT1eRO586d0ahRI832n56ertm+qXq2zAtzXj9FRUV48803ceXKFa1DsQn2IlKDvcm+JSUl4dSpU1qHYVPsVVRf7F/6sXr1apw9e1brMKyGfYssYU9SR1eTO5s2bcL8+fM12XdaWhrGjBmjyb6perbMC3NefyUlJXjxxRfh5+eHoUOHYv369cjPz9c6LKthL6LasDfZv+XLl+OOO+7APffcg6VLlzrUyZ0RexXVB/uXvsybNw+dO3dGSEgIEhMTkZmZqXVIt4R9i6piT1JPV5M7Wrlw4QKGDx+OrKwsrUOhSmyZF+bcOsrLy/H555/jmWeeQcuWLREeHo4tW7aguLhY69B0iXWpD+xN+lBeXg4AOHr0KGbNmoUOHTqgX79+WL16Na5evapxdPrGOtUv9i/9qaiogIjg8OHDmDp1Ktq0aYNBgwZh3bp1yMvL0zo83WB92if2pLrRxeRORUUFUlNTERUVhYcffhgAsHXrVkyaNAn+/v7IyclBVFQUWrZsiR49euCHH34AABw8eBCvvPIKOnfujMuXL2P06NFo0aIFevTogY8//hgA8M4778DFxQUGgwEAkJ+fj2XLlpmMvf/++/j1119x6dIlTJ48WYnryy+/hL+/P/bu3duQL4dDyMvLw8yZMzF79mzExMQgNDQUMTExyMnJAVD/vDDn9qeiogIVFRUoKSnBf//7Xzz++ONo0aIFxo8fj23btqGsrEzrEFVjL3J87E3OR0RQVlYGEcH+/fvx4osvolWrVhg4cCDWrVuny6sO2aucE/uXcxMRlJeXo6KiAl999RUmTJiAli1bYtiwYVi3bh0KCgq0DrFG7FuOhz1JA1JFcnKyWBjW3Llz5wSABAUFiYhIenq6NG3aVABIfHy8nD17VtavXy8AJCQkRMrLy2X79u3SuHFjASBTpkyRvXv3ysaNG8Xb21sAyL59+0REpEuXLmbPuepY5X0bbdmyRZo0aSLbtm2z8bOvn9GjR8vo0aO1DsNMfn6+BAYGyrx585SxzMxMCQwMlICAAMnJyRGRuueFObef4zc3N1cA1Pjj4eEhAMTHx0eio6Plm2++kQ8//NAu4q8Je1H92Ut9Voe9qWEAkOTkZK3DkODg4Bp7lKurq7i4uIi7u7sMGzZMUlJSpLi42G7irw17lXXoJd/sXw3Lnt7PWrduXWMvc3NzE4PBIE2aNJFx48bJ1q1bZePGjXYTf2XsW3XHz3vOl/Ma+k+KW62zP3bC39/f5Pd27dqhXbt2OHnyJF599VUAwNixYxETE4MjR47AxcUFjz32GPz9/XHq1CksXrwYTZo0AQBkZmZi2rRpWLlyJfr06QN3d3ez/Vkaq2rkyJHIy8uDq6urFZ6h81i8eDFOnTqFSZMmKWOtWrVCXFwcIiMjsXDhQixZsqTOeWHO/xQeHq7p/ktLS2tdpqSkBACQm5uLtWvXYvXq1WjZsiUA4Pjx4+jevbtNY6wv9iLHxd7UcBISErB582ZNY7h27VqNjxu/tlVRUYFdu3bh008/RdOmTQEAv/32GyoqKuDiYr8XQLNXORf2L21ofb4FoNavuhuvkC4sLMSHH36I9evXK73swIEDePDBB5UrGbTGvuU42JO0Yb9nJSpYakTNmzc3aXLGEy9j0oGbCQOA06dP33IMeky61vbt2wcA8Pb2Nhnv378/AGD//v23tH3mnBoae5FjYG8iR8de5bjYv8hRsW/pE3uSNnRz5Y41tW3bFoD57DA1DOPBmJaWhrvuuksZb926NQDAx8fH6vt0tpynpKRouv+8vLxa8+jh4YGSkhL4+PggIiIC48ePx4ULF/DUU0/Z7VU71uZsdWnv2JsaztSpUzX/F+8ePXrgwoUL1T7u6uoKEYGrqysGDx6MqKgojBo1Cp6enrjzzjvt+qoda3PWOtUT9i9taH2+BQB+fn41Pu7m5oby8nI0btwYTzzxBMLDw3H9+nWMGTMGvXv3bqAoGx7rU1vsSdpwnjOTSrKzswEAgwYNAvDnjLDxayIigtzcXJN1DAaDxRu/Gi/bJvWMM7Y7duwwGT9//jwA6+SlKubcPri7u8NgMMDLywvh4eHYunUrrly5grfffhv9+vWzm8uCGwrr0r6wN5HBYICrqytcXFzQv39/rF27FtnZ2dixYwfCwsLg4eGhdYiaYJ3aP/YvqszV1RWurq5wd3fH4MGD8f777yMzMxMffPABRowYodurEuqC9akt9iRt6GZy5/r16wBg8if9ioqKzJYz/lWLqkmqnKDdu3ejZ8+eyncAg4KCAAALFizA//73PyQkJCiX+n322WeoqKhAly5dkJGRoRQkcLNYmzVrhp07d1rjKTqN2NhYBAcHY+XKlbh06ZIynpiYiL59++Jvf/sbgPrnxYg5tw8uLi5wcXGBh4cH/vKXv+A///kPsrOzlRMMNzd9XUDIXuS42Juck8FggJubGwwGA/r06YM33ngDWVlZ2LNnDyIjI80uKdcL9irnwv5FlSenBwwYgPfeew9XrlzBJ598gsjISHh5eWkdYq3YtxwHe5I2dDG5U1hYiIULFwIALl68iOXLl2PJkiVIS0sDAMTHxyMvLw8JCQnKJdZz5841aQYrVqxAdnY2srKykJGRga+//lr5ULlkyRKEhIRg2bJl+Otf/4rHHnsMd911F8aPH4+cnByUlZUhLCwMt912Gw4fPqxs09PTE7fddhs8PT0b6JVwDI0bN8aBAwcwZswYPPPMM3jllVcwc+ZMtGjRAnv27LnlvBgx59ozfp0hKSkJV65cQUpKivK1Bj1iL3Js7E3Oxfgv1z169MDixYtx7tw5fPvtt4iOjoavr6/G0d0a9irnw/7lvIx/2rlXr15ISEhARkYGdu/ejcjISNx2221ah6ca+5ZjYU/ShkFEpPJASkoKIiIiUGVYt7p3744TJ044zPOpi7CwMABAamqqxpE0LGfOub0cv0VFRVi7di3CwsKUv4Clhr3EbwvOXJdGjpxfNVgDNxkMBiQnJ2t+z52kpCT07t0bgYGBdVrPXuK3FdapKUfPt1qsC1P29H62evVqhIaGomPHjqrXsaf4rcGZ65Of95wv5zUcv6n6+j4EEelCo0aNMHnyZK3DICKq1jPPPKN1CEREtyw6OlrrEIjITujia1m3oqCgwOS/5PiYc7JHrEtiDZAesE7JEtYF2TPWp/Nhzi1z2MmdgoICzJkzR7k50ksvvYSDBw9qHBXZEnNO9oh1SawB0gPWKVnCuiB7xvp0Psx5zRz2a1leXl6Ij49HfHy81qFQA2HOyR6xLok1QHrAOiVLWBdkz1ifzoc5r5nDXrlDREREREREROQMOLlDRERERERERKRjnNwhIiIiIiIiItIxTu4QEREREREREelYtTdUTklJacg4yAbS09MBMJfO5MCBA1qHYBWsWcdkrE/ml/TOUXotqcN8U1WOUhN8P9Y/ft5zPjX1H4OISOWBlJQURERE2DwoIrKdKoe1brD/EDmH5ORkhIeHax1GvRgMBq1DICI7wfMtItKKhf6TWu2VO3ptVlQzYzNnfh2To7xZsz4dE/sPAY4xOaLnySm6dWFhYQCA1NRUjSMhrfB8ixyBwWDg+5kO1dR/eM8dIiIiIiIiIiId4+QOEREREREREZGOcXKHiIiIiIiIiEjHOLlDRERERERERKRjnNwhIiIiIiIiItIxTu4QEREREREREelYtX8KnYiIiPTpjz/+wLZt21BcXIzHH38cXbt21TokojphDRPAOiD9Yw1TQ7L6lTunT5/G0qVLkZKSgnvvvRcGgwHBwcG4ceOGyXJffPEFhg4dCoPBgF69eiElJcXaoVjNxYsXsXbtWkRERKBPnz4239aaNWtw3333wdvbG/feey/Wrl2rPFZeXo5Zs2bhwoULtxRHfTG/t74te86v3rE+b31b9lyfzG/t8vPzMWXKFAwePBh33303ZsyYYXIi+d577yE8PBxxcXGYOHEiNm3aZLaNmpbRugb0jjVcO2eoYdZB7ZyhDvSMNVw7R69h1kD9t2XT3EkVycnJYmFYla+++krGjBkjJSUlIiKSm5srAASAREdHmy2flpYmAOTkyZP12l9DOnfunACQoKAgm25r1qxZMm7cOElMTJSpU6dK48aNBYCsXLlSWebq1avyxBNPyJkzZ+q8b+bXMubXPrA+LWN9Mr9qZGZmyv333y+BgYGSlZVl9vj8+fOlU6dOcu3aNRERuXbtmnTq1EkSEhLqtMyt1ICICABJTk6u17r2oL7xs4Zrp5caHj16tIwePbpe67IOaqeHOuD5Fmu4JnqoYRG+n1nSUOfcNuo/KVab3Pntt9+kQ4cOkp2dbTIOQPr372+xeEpLSwWAUhj2zlqJrm5b58+fl7Fjx5qMffbZZwJAunbtajL+888/S3BwsFy/fr1O+2V+q+fM+bUXrM/qOXN9Mr/qDBs2TFxdXeXgwYNmj507d07c3d1l0aJFJuPx8fHSpEkTuXLliqpljOpbAyLOObnDGlZHLzVc38kd1oE6eqgDnm+xhmuihxoW4ftZdWx9zm1kg/6TYpWvZYkIxo0bh2effRa+vr5mjycnJ6NNmzaYOHEi/vjjD2Xcze3mLX/c3d2tEYbunT17FkuXLjUZGzJkCFq1aoXMzEyT8bvvvhtdunTBjBkzbB4X82sd9ppfvWN9Woe91ifzq8727dvxySefIDQ0FCEhIWaPr1+/HqWlpXj00UdNxgcOHIjCwkKsWbNG1TJG7FHqsYbVcfQaZh2o4+h1oGesYXUcuYZZA9Zli9xZZXJn69at+PHHHzF06FCLj/v5+SElJQWFhYWIiIhAaWlptdvKy8vDzJkzMXv2bMTExCA0NBQxMTHIyclR9jVp0iT4+/sjJycHUVFRaNmyJXr06IEffvhB2U5RURFee+01PP/88+jVqxcGDx6MY8eOWePp2kzfvn3RunVrs/GSkhI89NBDZuOhoaF45513cObMGZvGxfxah73mV+9Yn9Zhr/XJ/KqTlJQEAOjQoQMefvhheHt7o2fPntixYwcA4NtvvwUAtG/f3mQ9f39/AMDPP/+sapnK2KPUYQ2r4+g1zDpQx9HrQM9Yw+o4cg2zBqzP6rmrw2U+1Xr66afFYDBIaWmp2WOVt7V8+XIBIK+88orFx/Pz8yUwMFDmzZunjGVmZkpgYKAEBARITk6OpKenS9OmTQWAxMfHy9mzZ2X9+vUCQEJCQpT1Jk6cKFkrAnUAAAh6SURBVCdOnFB+HzJkiLRu3Vry8vLq9NyqPpeGuESrsn379knjxo3lxx9/NHvsp59+EgBml+zVhPmtnrPm156wPqvnrPXJ/KrTqVMnASBLly6VjIwMOXjwoPj7+4vBYJDvvvtO7r33XgEgN27cMFmvsLBQAEjv3r1VLVNZfWpAxPm+lsUaVkdPNVyfr2WxDtTRSx3wfMsUa/hPeqlhEb6fVachz7mt3H+sc8+dTp06SbNmzSw+VnVb4eHhYjAYZMeOHWaPz5kzRwBIRkaGyTrr1q0TABIbGysiInfccYfZdlu3bi2enp4iInLo0CEBYPFn+/btdXpuVZ9LQ364Kisrk4cfflg2bdpk8fGLFy8KABk2bJjq/TK/1XPW/NoT1mf1nLU+mV91GjVqJG3atDEZM54EjRs3TvkefFFRkckyN27cEADSs2dPVctUVp8aEHG+yR3WsDp6quH6TO6wDtTRSx3wfMsUa/hPeqlhEb6fVachz7mt3H+sc8+dS5cuoXnz5qqWXbNmDYKCghAVFYWLFy+aPLZv3z4AgLe3t8l4//79AQD79+8HABgMBrPtNm/eHMXFxQCAw4cPIzg4GCJi9vPYY4/V7clp6B//+AceffRRPPXUUxYfb9asGQDg8uXLNo2D+bUNe8mv3rE+bcNe6pP5VcfPz8/su+yPPPIIAODkyZMICgoCAOVyZ6Nr164BANq2batqmcrYo9RhDavj6DXMOlDH0etAz1jD6jhyDbMGrM/aubPK5I6rqyvKy8tVLdu0aVN8/PHHuHHjBsaNG2cajMvNcNLS0kzGjfeB8PHxUbWP7OxsnDlzBoWFhWaPVVRUqNqG1rZv3w4vLy/MnTu32mUsFbwtML/WZ0/51TvWp/XZU30yv+p069bN7MbXLVu2BAD4+vrirrvuAgCzE6yMjAwAQL9+/VQtUxl7lDqsYXUcvYZZB+o4eh3oGWtYHUeuYdaA9Vk7d1aZ3GnTpo3ZzCLw54ta9cUNCgrCe++9hy+//NJk3DhbZ7zhlNH58+cBAIMGDVIVT1BQEAoLC7FkyRKT8ePHj2PVqlWqtqGlzz//HOnp6Zg5c6bJ+IEDB0x+N87e+vn52TQe5te67C2/esf6tC57q0/mV50xY8agqKgIR44cUcauXLkCAHjggQcwfvx4NGvWzOx12bNnDzw8PDBmzBhVy1TGHqUOa1gdR69h1oE6jl4HesYaVseRa5g1YH1Wz10dvsNVreeee04MBoPk5+ebjGdkZAgAuXjxosX1pk+fbrKvwsJCCQ4Olvbt25t8B2/q1KnSt29f5eZNxhtVVdauXTsBIKWlpVJUVCQBAQECQCZMmCAbNmyQuLg4GTJkiHJzpddff13uvPPOau8nUZXxBlbdunUze8ya29q9e7cMHDhQVq1apfysXLlSpk+fLnFxcSbL/vLLLwLY/gZwzK9j59eesD4tc+b6ZH7VbausrEyCg4NlzJgxytiqVavEz89Prl27JiIiS5YskW7duimvZV5ennTr1k3mz5+vrKNmGaP61ICI891zhzXseDVcn3vusA4cqw54vvUn1rApvdSwCN/PLGmoc24jK/cf69xQ+euvvxYAsmvXLmXs448/lmHDhgkAGT58uHzzzTdm65WWlkq/fv1MxvLz8yU2NlaGDBkiMTExEhsbK/Pnz5fi4mIREUlMTFRulLRgwQLJzc2VFStWKGOzZs2SGzduSFpamowcOVJ8fX3Fz89PoqOjJSsrS9nPiy++KC4uLtKuXbtan9+XX34p0dHRAkDc3d3ltddekyNHjlh9W/v375cmTZpYvCmUwWCQ33//3WRbb775pri6upqN14T5Nefs+bUnrE9zzl6fzK/6bV27dk0mTJggkZGREhcXJ+PGjZP09HSTZdasWSPjx4+XOXPmSFhYmKxevdpsO2qWEalfDYg43+QOa9jxarg+kzusA8eqA55v3cQatkwPNSzC97OqGuqcuzIr9x/rTO6IiAwbNkymTZtW5/W0dPLkSZM/pWYv21JrxIgRMnHixDqtw/xqvy21GjK/9oL1qf221GL/qZne86tGfWpAxPkmd0RYw45Ww/WZ3BFhHThSHfB8Sz9Yw9Xj+5m22xKxev+xzl/LAoC1a9fik08+0c2d5gsLC7Fy5Uq8++67drUttQ4dOoRTp05h6dKlDbI/5tex86t3rE/Hrk/mt2HzqwZ7VN2whlnDAOuAdaB/rGHWMGvAejVgk9zVYSaoVkePHpWnn35aCgoK6rV+Qzp69KjyXTx72pYaFy9elBEjRsj58+frvC7zq+221NAqv/aA9antttRg/6mdnvOrxq3UgIhzXrkjwhq2J7daw/W9ckeEdWBPeL7F92OttmUtWr0fswZunY36j/Wu3AGA4OBgxMfHIzEx0ZqbtYng4GB4e3vb3bZqU1ZWhnXr1mHDhg1o3759g+zTiPm1PS3zq3esT9tj/1FHr/lVgz2q/ljD9kHrGmYd2Aet60DPWMP2gedk6thjDdgydwYRkcoDKSkpiIiIQJVhchDMr2PTe371Hj/VjPklADAYDEhOTkZ4eLjWodSL3uOnWxcWFgYASE1N1TgS0ore38/0Hj9ZB9/P9KmG4zfVqlfuEBERERERERFRw+LkDhERERERERGRjnFyh4iIiIiIiIhIxzi5Q0RERERERESkY27VPWC8WRw5lvT0dADMr6My5lfvWJ+Oif2HHMXy5ct5M10ndvDgQQDsZc6M51vkKPh+pj819R+zv5Z14MABLFu2zOZBEZHt6LVJs/8QOYeXX34ZvXv31jqMeuGHISIy4vkWEWnFQv9JNZvcISIiIiIiIiIi3eCfQiciIiIiIiIi0jNO7hARERERERER6Rgnd4iIiIiIiIiIdIyTO0REREREREREOvb/AAjObMUQY+iuAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "plot_model(LSTM,show_shapes=True,show_layer_names=False,rankdir='LR')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "#make predictions\n",
    "y_pred=LSTM.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_test=y_test.ravel() #remove [] from y_test\n",
    "y_pred=y_pred.ravel() #remove [] from y_pred"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "scrolled": true
   },
   "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>y_test</th>\n",
       "      <th>y_pred</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>238.622</td>\n",
       "      <td>237.669479</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>239.916</td>\n",
       "      <td>235.678116</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>241.816</td>\n",
       "      <td>236.916702</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>230.940</td>\n",
       "      <td>233.600815</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>237.873</td>\n",
       "      <td>239.443283</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>237.015</td>\n",
       "      <td>235.624252</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>240.003</td>\n",
       "      <td>236.275787</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>235.295</td>\n",
       "      <td>235.335770</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>242.991</td>\n",
       "      <td>239.320053</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>230.117</td>\n",
       "      <td>230.458710</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>235.993</td>\n",
       "      <td>235.768173</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>232.400</td>\n",
       "      <td>241.623764</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>231.588</td>\n",
       "      <td>232.910995</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>241.370</td>\n",
       "      <td>240.779312</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>231.602</td>\n",
       "      <td>234.663315</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>233.409</td>\n",
       "      <td>235.239609</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>238.794</td>\n",
       "      <td>239.228333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>242.205</td>\n",
       "      <td>240.489044</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>243.136</td>\n",
       "      <td>240.306076</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>232.979</td>\n",
       "      <td>234.456802</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>228.792</td>\n",
       "      <td>230.755020</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>237.832</td>\n",
       "      <td>237.969345</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     y_test      y_pred\n",
       "0   238.622  237.669479\n",
       "1   239.916  235.678116\n",
       "2   241.816  236.916702\n",
       "3   230.940  233.600815\n",
       "4   237.873  239.443283\n",
       "5   237.015  235.624252\n",
       "6   240.003  236.275787\n",
       "7   235.295  235.335770\n",
       "8   242.991  239.320053\n",
       "9   230.117  230.458710\n",
       "10  235.993  235.768173\n",
       "11  232.400  241.623764\n",
       "12  231.588  232.910995\n",
       "13  241.370  240.779312\n",
       "14  231.602  234.663315\n",
       "15  233.409  235.239609\n",
       "16  238.794  239.228333\n",
       "17  242.205  240.489044\n",
       "18  243.136  240.306076\n",
       "19  232.979  234.456802\n",
       "20  228.792  230.755020\n",
       "21  237.832  237.969345"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#show predictions in tabular format\n",
    "combine=zip(y_test,y_pred)\n",
    "pd.DataFrame(combine,columns=['y_test','y_pred'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "#calculate metrics\n",
    "mae=mean_absolute_error(y_test,y_pred)\n",
    "mape=mean_absolute_percentage_error(y_test, y_pred)\n",
    "rmse=np.sqrt(mean_squared_error(y_test,y_pred))\n",
    "r2=abs(r2_score(y_test, y_pred))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>MAE</td>\n",
       "      <td>2.195671</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>MAPE</td>\n",
       "      <td>0.927766</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>RMSE</td>\n",
       "      <td>3.007450</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>R^2</td>\n",
       "      <td>0.534922</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      0         1\n",
       "0   MAE  2.195671\n",
       "1  MAPE  0.927766\n",
       "2  RMSE  3.007450\n",
       "3   R^2  0.534922"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#show metrics in tabular format\n",
    "metrics=[mae,mape,rmse,r2]\n",
    "metrics_labels=['MAE','MAPE','RMSE','R^2']\n",
    "pd.DataFrame(zip( metrics_labels,metrics))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Classification"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "#file path\n",
    "data=pd.read_csv('datasets/cls_interval1.csv',sep=',')\n",
    "\n",
    "# 'datasets/cls_interval1.csv'\n",
    "# 'datasets/cls_interval2.csv'\n",
    "# 'datasets/cls_interval3.csv'\n",
    "# 'datasets/cls_seven.csv'\n",
    "# 'datasets/cls_thirty.csv'\n",
    "# 'datasets/cls_ninety.csv'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>activeaddresses30trx</th>\n",
       "      <th>activeaddresses7ema</th>\n",
       "      <th>activeaddresses90std</th>\n",
       "      <th>confirmationtime</th>\n",
       "      <th>confirmationtime3wma</th>\n",
       "      <th>confirmationtime7sma</th>\n",
       "      <th>confirmationtime7wma</th>\n",
       "      <th>confirmationtime90ema</th>\n",
       "      <th>difficulty</th>\n",
       "      <th>fee_to_reward7trxUSD</th>\n",
       "      <th>...</th>\n",
       "      <th>sentbyaddress7var</th>\n",
       "      <th>sentinusd90varUSD</th>\n",
       "      <th>sentinusd90wmaUSD</th>\n",
       "      <th>sentinusdUSD</th>\n",
       "      <th>size30rsi</th>\n",
       "      <th>size90mom</th>\n",
       "      <th>size90sma</th>\n",
       "      <th>transactionfeesUSD</th>\n",
       "      <th>transactions3std</th>\n",
       "      <th>category</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.538</td>\n",
       "      <td>74692.0</td>\n",
       "      <td>18183.0</td>\n",
       "      <td>7.273</td>\n",
       "      <td>7.864</td>\n",
       "      <td>8.684</td>\n",
       "      <td>8.282</td>\n",
       "      <td>8.930</td>\n",
       "      <td>6695826.0</td>\n",
       "      <td>1.014</td>\n",
       "      <td>...</td>\n",
       "      <td>30129504.0</td>\n",
       "      <td>1.972499e+15</td>\n",
       "      <td>73371893</td>\n",
       "      <td>136779480.0</td>\n",
       "      <td>48.671</td>\n",
       "      <td>50223.0</td>\n",
       "      <td>154528.0</td>\n",
       "      <td>0.0777</td>\n",
       "      <td>12463.0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.569</td>\n",
       "      <td>78988.0</td>\n",
       "      <td>19348.0</td>\n",
       "      <td>7.956</td>\n",
       "      <td>7.822</td>\n",
       "      <td>8.459</td>\n",
       "      <td>8.099</td>\n",
       "      <td>8.909</td>\n",
       "      <td>6695826.0</td>\n",
       "      <td>0.730</td>\n",
       "      <td>...</td>\n",
       "      <td>45880948.0</td>\n",
       "      <td>2.512207e+15</td>\n",
       "      <td>78292153</td>\n",
       "      <td>278422775.0</td>\n",
       "      <td>52.003</td>\n",
       "      <td>61367.0</td>\n",
       "      <td>155210.0</td>\n",
       "      <td>0.1050</td>\n",
       "      <td>15762.0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.611</td>\n",
       "      <td>86026.0</td>\n",
       "      <td>21903.0</td>\n",
       "      <td>8.229</td>\n",
       "      <td>7.978</td>\n",
       "      <td>8.064</td>\n",
       "      <td>8.042</td>\n",
       "      <td>8.894</td>\n",
       "      <td>6695826.0</td>\n",
       "      <td>1.028</td>\n",
       "      <td>...</td>\n",
       "      <td>78761044.0</td>\n",
       "      <td>3.061579e+15</td>\n",
       "      <td>83274140</td>\n",
       "      <td>284039909.0</td>\n",
       "      <td>52.771</td>\n",
       "      <td>76185.0</td>\n",
       "      <td>156056.0</td>\n",
       "      <td>0.1330</td>\n",
       "      <td>10252.0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>3 rows × 42 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   activeaddresses30trx  activeaddresses7ema  activeaddresses90std  \\\n",
       "0                 0.538              74692.0               18183.0   \n",
       "1                 0.569              78988.0               19348.0   \n",
       "2                 0.611              86026.0               21903.0   \n",
       "\n",
       "   confirmationtime  confirmationtime3wma  confirmationtime7sma  \\\n",
       "0             7.273                 7.864                 8.684   \n",
       "1             7.956                 7.822                 8.459   \n",
       "2             8.229                 7.978                 8.064   \n",
       "\n",
       "   confirmationtime7wma  confirmationtime90ema  difficulty  \\\n",
       "0                 8.282                  8.930   6695826.0   \n",
       "1                 8.099                  8.909   6695826.0   \n",
       "2                 8.042                  8.894   6695826.0   \n",
       "\n",
       "   fee_to_reward7trxUSD  ...  sentbyaddress7var  sentinusd90varUSD  \\\n",
       "0                 1.014  ...         30129504.0       1.972499e+15   \n",
       "1                 0.730  ...         45880948.0       2.512207e+15   \n",
       "2                 1.028  ...         78761044.0       3.061579e+15   \n",
       "\n",
       "   sentinusd90wmaUSD  sentinusdUSD  size30rsi  size90mom  size90sma  \\\n",
       "0           73371893   136779480.0     48.671    50223.0   154528.0   \n",
       "1           78292153   278422775.0     52.003    61367.0   155210.0   \n",
       "2           83274140   284039909.0     52.771    76185.0   156056.0   \n",
       "\n",
       "   transactionfeesUSD  transactions3std  category  \n",
       "0              0.0777           12463.0         1  \n",
       "1              0.1050           15762.0         1  \n",
       "2              0.1330           10252.0         1  \n",
       "\n",
       "[3 rows x 42 columns]"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#visualize the data\n",
    "data.head(3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "#create train-test splits for SVM\n",
    "length=data.shape[1]-1\n",
    "X=data.iloc[:,:length]\n",
    "y=data.iloc[:,length:]\n",
    "X_train, X_test, y_train, y_test = train_test_split(X,y,test_size=0.2, shuffle=False, random_state=7)\n",
    "y_train=np.ravel(y_train)\n",
    "y_test=np.ravel(y_test)\n",
    "estimators=[]\n",
    "estimators.append(['robust',RobustScaler()])\n",
    "estimators.append(['mixmax',MinMaxScaler()])\n",
    "scaling=Pipeline(estimators)\n",
    "X_train=scaling.fit_transform(X_train)\n",
    "X_test=scaling.transform(X_test)\n",
    "X_train=np.reshape(X_train,(X_train.shape[0],1,X_train.shape[1]))\n",
    "X_test=np.reshape(X_test,(X_test.shape[0],1,X_test.shape[1]))\n",
    "unique1, id1 = np.unique(y_train, return_inverse=True)\n",
    "y_train=to_categorical(id1,num_classes=2)\n",
    "y_train=np.reshape(y_train, (y_train.shape[0],1,y_train.shape[1]))\n",
    "unique2, id2 = np.unique(y_test, return_inverse=True)\n",
    "y_test = to_categorical(id2, num_classes=2)\n",
    "y_test=np.reshape(y_test,(y_test.shape[0],1,y_test.shape[1]))\n",
    "y_test=y_test.argmax(-1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "#load saved model\n",
    "LSTM=load_model('trained_models/LSTM_cls_interval1.hdf5',compile=False)\n",
    "\n",
    "# 'trained_models/LSTM_cls_interval1.hdf5'\n",
    "# 'trained_models/LSTM_cls_interval2.hdf5'\n",
    "# 'trained_models/LSTM_cls_interval3.hdf5'\n",
    "# 'trained_models/LSTM_cls_seven.hdf5'\n",
    "# 'trained_models/LSTM_cls_thirty.hdf5'\n",
    "# 'trained_models/LSTM_cls_ninety.hdf5'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'name': 'sequential_1',\n",
       " 'layers': [{'class_name': 'Bidirectional',\n",
       "   'config': {'name': 'bidirectional_1',\n",
       "    'trainable': True,\n",
       "    'batch_input_shape': (None, 1, 41),\n",
       "    'dtype': 'float32',\n",
       "    'layer': {'class_name': 'LSTM',\n",
       "     'config': {'name': 'lstm_1',\n",
       "      'trainable': True,\n",
       "      'dtype': 'float32',\n",
       "      'return_sequences': True,\n",
       "      'return_state': False,\n",
       "      'go_backwards': False,\n",
       "      'stateful': False,\n",
       "      'unroll': False,\n",
       "      'time_major': False,\n",
       "      'zero_output_for_mask': True,\n",
       "      'units': 300,\n",
       "      'activation': 'relu',\n",
       "      'recurrent_activation': 'sigmoid',\n",
       "      'use_bias': True,\n",
       "      'kernel_initializer': {'class_name': 'VarianceScaling',\n",
       "       'config': {'scale': 1.0,\n",
       "        'mode': 'fan_avg',\n",
       "        'distribution': 'uniform',\n",
       "        'seed': None,\n",
       "        'dtype': 'float32'}},\n",
       "      'recurrent_initializer': {'class_name': 'Orthogonal',\n",
       "       'config': {'gain': 1.0, 'seed': None, 'dtype': 'float32'}},\n",
       "      'bias_initializer': {'class_name': 'Zeros',\n",
       "       'config': {'dtype': 'float32'}},\n",
       "      'unit_forget_bias': True,\n",
       "      'kernel_regularizer': None,\n",
       "      'recurrent_regularizer': None,\n",
       "      'bias_regularizer': None,\n",
       "      'activity_regularizer': None,\n",
       "      'kernel_constraint': None,\n",
       "      'recurrent_constraint': None,\n",
       "      'bias_constraint': None,\n",
       "      'dropout': 0.0,\n",
       "      'recurrent_dropout': 0.0,\n",
       "      'implementation': 2}},\n",
       "    'merge_mode': 'concat'}},\n",
       "  {'class_name': 'Bidirectional',\n",
       "   'config': {'name': 'bidirectional_2',\n",
       "    'trainable': True,\n",
       "    'dtype': 'float32',\n",
       "    'layer': {'class_name': 'LSTM',\n",
       "     'config': {'name': 'lstm_2',\n",
       "      'trainable': True,\n",
       "      'dtype': 'float32',\n",
       "      'return_sequences': True,\n",
       "      'return_state': False,\n",
       "      'go_backwards': False,\n",
       "      'stateful': False,\n",
       "      'unroll': False,\n",
       "      'time_major': False,\n",
       "      'zero_output_for_mask': True,\n",
       "      'units': 100,\n",
       "      'activation': 'relu',\n",
       "      'recurrent_activation': 'sigmoid',\n",
       "      'use_bias': True,\n",
       "      'kernel_initializer': {'class_name': 'VarianceScaling',\n",
       "       'config': {'scale': 1.0,\n",
       "        'mode': 'fan_avg',\n",
       "        'distribution': 'uniform',\n",
       "        'seed': None,\n",
       "        'dtype': 'float32'}},\n",
       "      'recurrent_initializer': {'class_name': 'Orthogonal',\n",
       "       'config': {'gain': 1.0, 'seed': None, 'dtype': 'float32'}},\n",
       "      'bias_initializer': {'class_name': 'Zeros',\n",
       "       'config': {'dtype': 'float32'}},\n",
       "      'unit_forget_bias': True,\n",
       "      'kernel_regularizer': None,\n",
       "      'recurrent_regularizer': None,\n",
       "      'bias_regularizer': None,\n",
       "      'activity_regularizer': None,\n",
       "      'kernel_constraint': None,\n",
       "      'recurrent_constraint': None,\n",
       "      'bias_constraint': None,\n",
       "      'dropout': 0.0,\n",
       "      'recurrent_dropout': 0.0,\n",
       "      'implementation': 2}},\n",
       "    'merge_mode': 'concat'}},\n",
       "  {'class_name': 'Bidirectional',\n",
       "   'config': {'name': 'bidirectional_3',\n",
       "    'trainable': True,\n",
       "    'dtype': 'float32',\n",
       "    'layer': {'class_name': 'LSTM',\n",
       "     'config': {'name': 'lstm_3',\n",
       "      'trainable': True,\n",
       "      'dtype': 'float32',\n",
       "      'return_sequences': True,\n",
       "      'return_state': False,\n",
       "      'go_backwards': False,\n",
       "      'stateful': False,\n",
       "      'unroll': False,\n",
       "      'time_major': False,\n",
       "      'zero_output_for_mask': True,\n",
       "      'units': 30,\n",
       "      'activation': 'relu',\n",
       "      'recurrent_activation': 'sigmoid',\n",
       "      'use_bias': True,\n",
       "      'kernel_initializer': {'class_name': 'VarianceScaling',\n",
       "       'config': {'scale': 1.0,\n",
       "        'mode': 'fan_avg',\n",
       "        'distribution': 'uniform',\n",
       "        'seed': None,\n",
       "        'dtype': 'float32'}},\n",
       "      'recurrent_initializer': {'class_name': 'Orthogonal',\n",
       "       'config': {'gain': 1.0, 'seed': None, 'dtype': 'float32'}},\n",
       "      'bias_initializer': {'class_name': 'Zeros',\n",
       "       'config': {'dtype': 'float32'}},\n",
       "      'unit_forget_bias': True,\n",
       "      'kernel_regularizer': None,\n",
       "      'recurrent_regularizer': None,\n",
       "      'bias_regularizer': None,\n",
       "      'activity_regularizer': None,\n",
       "      'kernel_constraint': None,\n",
       "      'recurrent_constraint': None,\n",
       "      'bias_constraint': None,\n",
       "      'dropout': 0.0,\n",
       "      'recurrent_dropout': 0.0,\n",
       "      'implementation': 2}},\n",
       "    'merge_mode': 'concat'}},\n",
       "  {'class_name': 'Dense',\n",
       "   'config': {'name': 'dense_1',\n",
       "    'trainable': True,\n",
       "    'dtype': 'float32',\n",
       "    'units': 2,\n",
       "    'activation': 'softmax',\n",
       "    'use_bias': True,\n",
       "    'kernel_initializer': {'class_name': 'VarianceScaling',\n",
       "     'config': {'scale': 1.0,\n",
       "      'mode': 'fan_avg',\n",
       "      'distribution': 'uniform',\n",
       "      'seed': None,\n",
       "      'dtype': 'float32'}},\n",
       "    'bias_initializer': {'class_name': 'Zeros',\n",
       "     'config': {'dtype': 'float32'}},\n",
       "    'kernel_regularizer': None,\n",
       "    'bias_regularizer': None,\n",
       "    'activity_regularizer': None,\n",
       "    'kernel_constraint': None,\n",
       "    'bias_constraint': None}}]}"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#show model configuration\n",
    "LSTM.get_config()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "plot_model(LSTM,show_shapes=True,show_layer_names=False,rankdir='LR')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "#make predictions\n",
    "y_pred=LSTM.predict_classes(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_pred=y_pred.ravel()\n",
    "y_test=y_test.ravel()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "scrolled": true
   },
   "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>y_test</th>\n",
       "      <th>y_pred</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>190</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>191</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>192</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>193</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>194</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>195</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>196</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>197</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>198</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>199</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>200</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>201</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>202</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>203</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>204</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>205</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>206</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>207</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>208</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>209</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>210</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>211</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>212</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>213</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>214</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>215</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>216</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>217</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>218</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>219</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>220 rows × 2 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "     y_test  y_pred\n",
       "0         1       0\n",
       "1         1       0\n",
       "2         1       0\n",
       "3         0       0\n",
       "4         0       0\n",
       "5         0       0\n",
       "6         0       0\n",
       "7         0       0\n",
       "8         1       0\n",
       "9         1       1\n",
       "10        1       1\n",
       "11        1       1\n",
       "12        1       0\n",
       "13        0       0\n",
       "14        0       0\n",
       "15        1       1\n",
       "16        0       1\n",
       "17        0       1\n",
       "18        0       1\n",
       "19        0       1\n",
       "20        0       1\n",
       "21        1       1\n",
       "22        1       1\n",
       "23        0       1\n",
       "24        0       1\n",
       "25        0       1\n",
       "26        1       0\n",
       "27        1       1\n",
       "28        1       1\n",
       "29        1       0\n",
       "..      ...     ...\n",
       "190       0       1\n",
       "191       0       1\n",
       "192       1       1\n",
       "193       1       1\n",
       "194       1       1\n",
       "195       0       1\n",
       "196       1       1\n",
       "197       1       1\n",
       "198       0       1\n",
       "199       0       1\n",
       "200       1       1\n",
       "201       1       1\n",
       "202       0       1\n",
       "203       1       1\n",
       "204       0       1\n",
       "205       0       1\n",
       "206       1       1\n",
       "207       0       1\n",
       "208       1       1\n",
       "209       1       1\n",
       "210       0       1\n",
       "211       0       1\n",
       "212       1       1\n",
       "213       1       1\n",
       "214       1       1\n",
       "215       0       1\n",
       "216       0       1\n",
       "217       1       1\n",
       "218       1       1\n",
       "219       1       1\n",
       "\n",
       "[220 rows x 2 columns]"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#show predictions in tabular format\n",
    "combine=zip(y_test,y_pred)\n",
    "pd.DataFrame(combine,columns=['y_test','y_pred'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "#calculate metrics\n",
    "accuracy=accuracy_score(y_test,y_pred)\n",
    "f1=f1_score(y_test,y_pred, average='binary') # try average='weighted' or 'binary' and take higher\n",
    "auc=roc_auc_score(y_test,y_pred)\n",
    "recall=recall_score(y_test,y_pred)\n",
    "precision=precision_score(y_test,y_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "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>0</th>\n",
       "      <th>1</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Accuracy</td>\n",
       "      <td>0.540909</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>F1-score</td>\n",
       "      <td>0.632727</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>AUC</td>\n",
       "      <td>0.517473</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Recall</td>\n",
       "      <td>0.701613</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Precision</td>\n",
       "      <td>0.576159</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "           0         1\n",
       "0   Accuracy  0.540909\n",
       "1   F1-score  0.632727\n",
       "2        AUC  0.517473\n",
       "3     Recall  0.701613\n",
       "4  Precision  0.576159"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#show metrics in tabular format\n",
    "metrics=[accuracy, f1, auc, recall, precision]\n",
    "metrics_labels=['Accuracy','F1-score','AUC','Recall','Precision']\n",
    "pd.DataFrame(zip( metrics_labels,metrics))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.9"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}