notebook-pH-raw_data.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Forecasting in Aquaponics"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Libraries and setup"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib as mpl\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "import tensorflow as tf\n",
    "import os\n",
    "import sys\n",
    "\n",
    "os.environ['KMP_DUPLICATE_LIB_OK']='True'\n",
    "os.environ[\"CUDA_VISIBLE_DEVICES\"] = \"3\"\n",
    "\n",
    "# np.set_printoptions(threshold=sys.maxsize)\n",
    "\n",
    "mpl.rcParams['figure.figsize'] = (14, 8)\n",
    "mpl.rcParams['axes.grid'] = True\n",
    "\n",
    "font = {'size'   : 18}\n",
    "\n",
    "mpl.rc('font', **font)\n",
    "\n",
    "np.random.seed(10)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Data adquisition"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Data adquisition"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Dataset ready\n"
     ]
    }
   ],
   "source": [
    "def read_file(textfile):\n",
    "    df = pd.read_csv(textfile, compression='zip', sep=\"\\t\")\n",
    "    df.drop(df.columns[0], axis=1, inplace=True)\n",
    "    print('Dataset ready')\n",
    "    return df\n",
    "\n",
    "textfile = \"dataset_aquaponics_03012021_05222021.txt.zip\"\n",
    "df = read_file(textfile)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Organize data in columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Label                      1         2    3       4       5          6    \\\n",
      "DateTime                                                                   \n",
      "2021-04-01 00:00:12  20.880000  51.34556  0.0  6.6498  6.6667  44.003500   \n",
      "2021-04-01 00:00:38  20.822498  51.09667  0.0  6.6498  6.6667  44.007060   \n",
      "2021-04-01 00:01:04  20.853998  51.17500  0.0  6.6514  6.6681  43.999874   \n",
      "2021-04-01 00:05:57  20.928003  50.80000  0.0  6.6584  6.6718  43.621853   \n",
      "2021-04-01 00:06:23  21.000000  50.68667  0.0  6.6596  6.6731  42.650578   \n",
      "\n",
      "Label                      7       8       9       10   ...       107  \\\n",
      "DateTime                                                ...             \n",
      "2021-04-01 00:00:12  49.062016  99.262  95.587  11.690  ...  50.53125   \n",
      "2021-04-01 00:00:38  49.063637  99.262  95.587  11.690  ...  50.53125   \n",
      "2021-04-01 00:01:04  49.068360  99.255  95.599  11.687  ...  52.50001   \n",
      "2021-04-01 00:05:57  49.068610  99.289  95.774  11.615  ...  52.50001   \n",
      "2021-04-01 00:06:23  49.073135  99.273  95.758  11.613  ...  52.50001   \n",
      "\n",
      "Label                     108        109        110   111   112   113   114  \\\n",
      "DateTime                                                                      \n",
      "2021-04-01 00:00:12  77.43749  45.281254  97.781235  10.0  10.0  10.0  10.0   \n",
      "2021-04-01 00:00:38  77.43749  45.281254  97.781235  10.0  10.0  10.0  10.0   \n",
      "2021-04-01 00:01:04  76.78126  45.281254  97.781235  10.0  10.0  10.0  10.0   \n",
      "2021-04-01 00:05:57  76.78126  45.937490  97.125000  10.0  10.0  10.0  10.0   \n",
      "2021-04-01 00:06:23  76.78126  45.937490  97.781235  10.0  10.0  10.0  10.0   \n",
      "\n",
      "Label                115  116  \n",
      "DateTime                       \n",
      "2021-04-01 00:00:12  0.0  0.0  \n",
      "2021-04-01 00:00:38  0.0  0.0  \n",
      "2021-04-01 00:01:04  0.0  0.0  \n",
      "2021-04-01 00:05:57  0.0  0.0  \n",
      "2021-04-01 00:06:23  0.0  0.0  \n",
      "\n",
      "[5 rows x 114 columns]\n"
     ]
    }
   ],
   "source": [
    "\n",
    "df_grouped = pd.pivot_table(df, index = 'DateTime', columns = 'Label', values = 'Value')\n",
    "df_grouped.index = pd.to_datetime(df_grouped.index)\n",
    "\n",
    "start_date = pd.to_datetime('2021-04-01 00:00:00')\n",
    "end_date = pd.to_datetime('2021-04-17 23:59:59')\n",
    "\n",
    "df_grouped = df_grouped.loc[start_date:end_date]\n",
    "print(df_grouped.head())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Data visualization"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1008x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sample_sensor = 5\n",
    "\n",
    "df_grouped[sample_sensor].plot(x = 'DateTime', y = 'Value', kind = 'line')\n",
    "plt.xticks(rotation=20)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Data Pre-processing"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Selecting target "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Target variables:\n",
    "\n",
    "* 4     = pH sump B\n",
    "* 5     = pH sump A\n",
    "* 8     = % oxygen B\n",
    "* 9     = % oxygen A\n",
    "* 93*   = Weight cell 1\n",
    "* 99*   = Nitrate B \n",
    "* 100*  = Ammonia B\n",
    "* 103*  = Nitrate A \n",
    "* 104*  = Ammonia A\n",
    "* sensors are not online yet*"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "label_target = {\n",
    "#4: 'pH sump B',}\n",
    "5: 'pH sump A',}\n",
    "# 8: '% oxygen B',}\n",
    "# 9: '% oxygen A'}"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Select features - Human knowledge"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Label                      1         2       4       5          6    \\\n",
      "DateTime                                                              \n",
      "2021-04-01 00:00:12  20.880000  51.34556  6.6498  6.6667  44.003500   \n",
      "2021-04-01 00:00:38  20.822498  51.09667  6.6498  6.6667  44.007060   \n",
      "2021-04-01 00:01:04  20.853998  51.17500  6.6514  6.6681  43.999874   \n",
      "2021-04-01 00:05:57  20.928003  50.80000  6.6584  6.6718  43.621853   \n",
      "2021-04-01 00:06:23  21.000000  50.68667  6.6596  6.6731  42.650578   \n",
      "\n",
      "Label                      7       8       9       10      11   ...    91   \\\n",
      "DateTime                                                        ...          \n",
      "2021-04-01 00:00:12  49.062016  99.262  95.587  11.690  13.552  ...  406.0   \n",
      "2021-04-01 00:00:38  49.063637  99.262  95.587  11.690  13.552  ...  406.0   \n",
      "2021-04-01 00:01:04  49.068360  99.255  95.599  11.687  13.528  ...  406.0   \n",
      "2021-04-01 00:05:57  49.068610  99.289  95.774  11.615  13.512  ...  406.0   \n",
      "2021-04-01 00:06:23  49.073135  99.273  95.758  11.613  13.502  ...  406.0   \n",
      "\n",
      "Label                  92    102    103       107       108        109  \\\n",
      "DateTime                                                                 \n",
      "2021-04-01 00:00:12  406.0  8.98  322.2  50.53125  77.43749  45.281254   \n",
      "2021-04-01 00:00:38  406.0  8.98  322.2  50.53125  77.43749  45.281254   \n",
      "2021-04-01 00:01:04  406.0  9.06  324.3  52.50001  76.78126  45.281254   \n",
      "2021-04-01 00:05:57  406.0  9.01  322.0  52.50001  76.78126  45.937490   \n",
      "2021-04-01 00:06:23  406.0  8.95  322.0  52.50001  76.78126  45.937490   \n",
      "\n",
      "Label                      110  115  116  \n",
      "DateTime                                  \n",
      "2021-04-01 00:00:12  97.781235  0.0  0.0  \n",
      "2021-04-01 00:00:38  97.781235  0.0  0.0  \n",
      "2021-04-01 00:01:04  97.781235  0.0  0.0  \n",
      "2021-04-01 00:05:57  97.125000  0.0  0.0  \n",
      "2021-04-01 00:06:23  97.781235  0.0  0.0  \n",
      "\n",
      "[5 rows x 46 columns]\n"
     ]
    }
   ],
   "source": [
    "drop_signal = [3, 12, 13, 14, 15, 22, 23, 24, 25, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 42, 43, 44, 45, 46, 47, 48, 49, 56, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 93, 94, 95, 97, 98, 99, 100, 101, 104, 105, 106, 111, 112, 113, 114]\n",
    "df_grouped  = df_grouped.drop(columns = drop_signal)\n",
    "print(df_grouped.head())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Denoising dataframe - FFT filter"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Check sampling frecuency"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0 days 00:00:26\n"
     ]
    }
   ],
   "source": [
    "sampling_time = df_grouped.index.to_series().diff().median()\n",
    "print(sampling_time)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Design sample filter"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1008x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "def f_filtering(df, cutt_fc):\n",
    "    \n",
    "    new_df = df.copy() \n",
    "    index = df.index\n",
    "    \n",
    "    dt = 25 # seconds\n",
    "    n_samples = new_df.shape[0]\n",
    "    \n",
    "    # Fast Fourier transform\n",
    "    f_hat = np.fft.fft(df.to_numpy())\n",
    "    f_mag_2 = np.real(f_hat * np.conj(f_hat)) / n_samples\n",
    "    fc_df = (1 / (dt * n_samples)) * np.arange(n_samples)\n",
    "    \n",
    "    # Cutting frecuency\n",
    "    ind_filter = f_mag_2 > cutt_fc\n",
    "    \n",
    "    # Filter signal\n",
    "    f_hat_clean = f_hat * ind_filter\n",
    "    filter_signal = np.fft.ifft(f_hat_clean)\n",
    "    \n",
    "    \n",
    "    filter_signal = np.real(filter_signal)\n",
    "    \n",
    "    return filter_signal\n",
    "\n",
    "df_grouped[sample_sensor].plot(x = 'DateTime', y = 'Value', kind = 'line')\n",
    "plt.xticks(rotation=20)\n",
    "plt.title('Raw data')\n",
    "plt.show()\n",
    "\n",
    "# Cutting frecuency\n",
    "cutt_fc = {\n",
    "1: 5,   \n",
    "2: 150,     \n",
    "4: 0.001,   \n",
    "5: 0.005,   \n",
    "6: 25,   \n",
    "7: 25,   \n",
    "8: 1,   \n",
    "9: 1,  \n",
    "10: 0.1,  \n",
    "11: 0.1,  \n",
    "16: 0.1,  \n",
    "17: 0.1,\n",
    "18: 500,  \n",
    "19: 5,  \n",
    "20: 20,  \n",
    "21: 20, \n",
    "102: 10, \n",
    "103: 50}\n",
    "\n",
    "df_clean = df_grouped.copy()\n",
    "\n",
    "# for sensor in cutt_fc:\n",
    "#     df_clean[sensor] = f_filtering(df_grouped[sensor], cutt_fc[sensor])\n",
    "    \n",
    "# print(df_clean.shape)\n",
    "\n",
    "# df_clean[sample_sensor].plot(x = 'DateTime', y = 'Value', kind = 'line')\n",
    "# plt.xticks(rotation=20)\n",
    "# plt.title('Filtered data')\n",
    "# plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Feature selection via correlation analysis"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Describe  dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "         count         mean           std          min          25%  \\\n",
      "Label                                                                 \n",
      "1      28636.0    19.391888      1.699759    14.888001    18.188002   \n",
      "2      28636.0    43.098705     11.559430    15.972223    33.881387   \n",
      "4      28636.0     6.664216      0.051004     6.210000     6.636400   \n",
      "5      28636.0     6.675076      0.034890     6.116000     6.666700   \n",
      "6      28636.0    46.009057      2.698169    33.928627    43.997658   \n",
      "7      28636.0    47.539641      3.026140    30.119194    45.693269   \n",
      "8      28636.0   101.171154      2.667098    94.335000    99.433000   \n",
      "9      28636.0    96.048075      4.579891    88.220000    92.955000   \n",
      "10     28636.0    13.012148      0.706207    11.117000    12.511000   \n",
      "11     28636.0    13.447863      0.572697    12.027000    13.062000   \n",
      "16     28636.0    10.730241      0.314434     9.751400    10.513000   \n",
      "17     28636.0    10.101525      0.501803     9.000000     9.753000   \n",
      "18     28636.0   431.811461     29.766308   330.000000   404.000000   \n",
      "19     28636.0    21.442294      3.359788    15.059999    18.970000   \n",
      "20     28636.0    10.902294      7.754490     0.000000     4.774928   \n",
      "21     28636.0     6.610800      6.628146     0.000000     3.374934   \n",
      "26     28636.0   233.983133    238.683192     0.000000     0.000000   \n",
      "27     28636.0   233.883468    238.600936     0.000000     0.000000   \n",
      "28     28636.0   246.265540    251.500245     0.000000     0.000000   \n",
      "29     28636.0   233.921637    238.641360     0.000000     0.000000   \n",
      "50     28636.0     5.498602      5.533782     0.000000     0.090384   \n",
      "51     28636.0     6.626879      5.990154     0.000000     0.000000   \n",
      "52     28636.0     8.003685      6.897884     0.000001     0.136293   \n",
      "53     28636.0     7.245122      6.393222     0.000000     0.107661   \n",
      "54     28636.0     4.780400      4.743538     0.000000     0.000000   \n",
      "55     28636.0     6.638935      5.957586     0.000001     0.422319   \n",
      "57     28636.0    20.759848     54.345074     0.000000     0.000000   \n",
      "58     28636.0    11.714066     41.152462     0.000000     0.000000   \n",
      "59     28636.0    31.409415     90.516194     0.000000     0.000000   \n",
      "60     28636.0   121.599944    188.810537     0.000000     0.000000   \n",
      "61     28636.0   180.245425    233.206339     0.000000     0.000000   \n",
      "62     28636.0   180.245425    233.206339     0.000000     0.000000   \n",
      "87     28636.0  1356.293721     10.510800  1339.000000  1347.000000   \n",
      "88     28636.0   350.358535      6.128813   341.000000   345.000000   \n",
      "89     28636.0   386.400894      6.437937   377.000000   381.000000   \n",
      "90     28636.0   345.357592      5.874463   336.000000   340.000000   \n",
      "91     28636.0   414.654945      5.023185   406.000000   410.000000   \n",
      "92     28636.0   414.654945      5.023185   406.000000   410.000000   \n",
      "102    28636.0     9.565989      5.035761     0.000000     6.180000   \n",
      "103    28636.0   694.495027  59092.066275   241.800000   298.400000   \n",
      "107    28636.0    51.575167      2.098091    35.437496    50.531250   \n",
      "108    28636.0    77.352089      3.954382    57.093742    74.812490   \n",
      "109    28636.0    40.128718      3.871128    32.812500    38.062496   \n",
      "110    28636.0    94.370148      3.347678    68.250000    93.187510   \n",
      "115    28636.0    30.259825     51.008317     0.000000     0.000000   \n",
      "116    28636.0    12.959481     29.156032     0.000000     0.000000   \n",
      "\n",
      "               50%          75%           max  \n",
      "Label                                          \n",
      "1        19.650002    20.592499  2.444600e+01  \n",
      "2        44.373890    52.030000  6.875000e+01  \n",
      "4         6.657500     6.680800  7.116800e+00  \n",
      "5         6.674900     6.686500  6.958000e+00  \n",
      "6        45.693154    47.379129  5.388481e+01  \n",
      "7        47.690038    49.095533  5.416806e+01  \n",
      "8       100.780000   103.130000  1.085500e+02  \n",
      "9        94.880000    97.112250  1.089800e+02  \n",
      "10       13.027000    13.576000  1.438400e+01  \n",
      "11       13.466000    13.866000  1.511200e+01  \n",
      "16       10.718000    10.952000  1.172100e+01  \n",
      "17        9.987850    10.256000  1.156200e+01  \n",
      "18      435.000000   457.000000  5.390000e+02  \n",
      "19       21.140000    23.060000  3.203000e+01  \n",
      "20       10.649920    15.899897  6.000000e+01  \n",
      "21        6.274939     8.324981  6.000000e+01  \n",
      "26      161.000000   466.000000  6.440000e+02  \n",
      "27      161.000000   466.000000  6.440000e+02  \n",
      "28      171.000000   491.000000  6.930000e+02  \n",
      "29      161.000000   466.000000  6.440000e+02  \n",
      "50        4.083759    10.172269  1.671268e+01  \n",
      "51        6.397726    13.177115  1.551657e+01  \n",
      "52        8.298773    14.725478  1.996974e+01  \n",
      "53        6.814311    13.513293  2.084017e+01  \n",
      "54        3.612955     9.256068  1.330080e+01  \n",
      "55        5.917156    11.754891  1.995870e+01  \n",
      "57        0.000000     0.000000  3.820000e+02  \n",
      "58        0.000000     0.000000  3.360000e+02  \n",
      "59        0.000000     0.000000  7.120000e+02  \n",
      "60        0.000000   214.000000  6.640000e+02  \n",
      "61        0.000000   360.000000  7.200000e+02  \n",
      "62        0.000000   360.000000  7.200000e+02  \n",
      "87     1356.000000  1366.000000  1.374000e+03  \n",
      "88      351.000000   356.000000  3.620000e+02  \n",
      "89      386.000000   392.000000  3.990000e+02  \n",
      "90      345.000000   350.000000  3.560000e+02  \n",
      "91      415.000000   419.000000  4.230000e+02  \n",
      "92      415.000000   419.000000  4.230000e+02  \n",
      "102       8.350000    11.560000  3.497000e+01  \n",
      "103     326.200000   366.800000  9.999999e+06  \n",
      "107      51.843746    52.500010  6.168750e+01  \n",
      "108      76.781260    77.437490  9.121874e+01  \n",
      "109      38.718760    44.624990  4.659375e+01  \n",
      "110      95.156235    97.125000  1.135312e+02  \n",
      "115       0.000000    57.558613  1.798057e+02  \n",
      "116       0.000000     0.000000  1.137255e+02  \n"
     ]
    }
   ],
   "source": [
    "print(df_clean.describe().transpose())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Split dataset: train, validation and test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "total_data = len(df_clean)\n",
    "\n",
    "train_perc, val_perc, test_perc = 0.7, 0.2, 0.1\n",
    "\n",
    "train_df    = df_clean[0:int(total_data*train_perc)]\n",
    "val_df      = df_clean[int(total_data*train_perc):int(total_data*(train_perc + val_perc))]\n",
    "test_df     = df_clean[int(total_data*(train_perc + val_perc)):total_data]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Data normalization"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Min-max normalization"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "def f_normalize_minmax(no_norm_df, df):\n",
    "    norm_df = (no_norm_df - df.min()) / (df.max() - df.min())\n",
    "    return norm_df\n",
    "\n",
    "norm_train_df   = f_normalize_minmax(train_df, df_clean)\n",
    "norm_val_df     = f_normalize_minmax(val_df, df_clean)\n",
    "norm_test_df    = f_normalize_minmax(test_df, df_clean)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Show tails with box plot"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1008x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "melt_train_df = norm_train_df.melt(var_name = 'Label', value_name = 'Normalized value')\n",
    "plt.figure()\n",
    "ax = sns.violinplot(x = 'Label', y = 'Normalized value', data = melt_train_df)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Plot normalized data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1008x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "norm_train_df[sample_sensor].plot(x = 'DateTime', y = 'Value', kind = 'line')\n",
    "plt.xticks(rotation=20)\n",
    "plt.title('Normalized data')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Create input and target dataframe - Window generation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "n_hour_data = 65 # 65 entries ~ 1 hour - There is a variable sampling frecuency\n",
    "n_hours_in  = 3\n",
    "n_hours_out = 1\n",
    "\n",
    "input_width     = n_hour_data * n_hours_in \n",
    "label_width     = n_hour_data * n_hours_out \n",
    "offset_width    = 1 # 1 as default\n",
    "\n",
    "def f_window_gen(df, label_target, input_width, label_width, offset_width):\n",
    "\n",
    "    time_range = df.index\n",
    "    data = []\n",
    "    target = []\n",
    "\n",
    "    data_time = []\n",
    "    target_time = []\n",
    "\n",
    "    for i in range(len(time_range) - (input_width + offset_width + label_width)):\n",
    "        range_data = df.loc[time_range[i]:time_range[input_width + i]].values\n",
    "        data.append(range_data)\n",
    "        data_time.append(time_range[i:input_width + i])\n",
    "\n",
    "        range_target = df[label_target].loc[time_range[input_width + offset_width + i: input_width + offset_width + label_width + i]].values\n",
    "        target.append(range_target.flatten('F'))\n",
    "        target_time.append(time_range[input_width + offset_width + i: input_width + offset_width + label_width + i])\n",
    "\n",
    "    data = np.array(data)\n",
    "    target = np.array(target)\n",
    "\n",
    "    return data, target, data_time, target_time\n",
    "\n",
    "train_data, train_target, train_data_time, train_target_time = f_window_gen(norm_train_df, label_target, input_width, label_width, offset_width)\n",
    "val_data, val_target, val_data_time, val_target_time = f_window_gen(norm_val_df, label_target, input_width, label_width, offset_width)\n",
    "test_data, test_target, test_data_time, test_target_time = f_window_gen(norm_test_df, label_target, input_width, label_width, offset_width)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Plotting - Normalized values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot_results(label_target, label_time, label, results, label_width):\n",
    "    n_label = np.int(len(label_target))\n",
    "\n",
    "    sensor_tag = list(label_target.values())\n",
    "    \n",
    "    fig, ax = plt.subplots(n_label, sharex = 'all')\n",
    "    \n",
    "    for i in range(n_label):\n",
    "\n",
    "        label_plot = label[((i) * label_width): ((i + 1)* label_width)]\n",
    "        ax.plot(label_time, label_plot, color = 'blue', label = 'Real')\n",
    "        \n",
    "        results_plot = results[((i) * label_width): ((i + 1)* label_width)]\n",
    "        ax.plot(label_time, results_plot, color = 'orange', label = 'Prediction')\n",
    "\n",
    "        ax.set_ylabel(sensor_tag[i])\n",
    "\n",
    "    lines, labels = fig.axes[-1].get_legend_handles_labels()\n",
    "    fig.legend(lines, labels, loc = 'center right')\n",
    "    plt.xticks(rotation=20)\n",
    "    plt.xlabel('Day - Time')\n",
    "    plt.ylim([0.8, 0.9])\n",
    "    plt.show()\n",
    "\n",
    "    pass"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Plotting - History"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot_history(history):\n",
    "\n",
    "    acc = history.history['mean_absolute_error']\n",
    "    val_acc = history.history['val_mean_absolute_error']\n",
    "\n",
    "    loss = history.history['loss']\n",
    "    val_loss = history.history['val_loss']\n",
    "\n",
    "    # Get number of epochs\n",
    "    epochs = range(len(acc))\n",
    "\n",
    "#     fig, ax = plt.subplots(1, sharex = 'all')\n",
    "\n",
    "    # Plot training and validation accuracy per epoch\n",
    "#     ax[0].plot(epochs, acc, label = 'Training')\n",
    "#     ax[0].plot(epochs, val_acc, label = 'Validation')\n",
    "#     ax[0].title.set_text('Training and validation MAE loss')\n",
    "\n",
    "    # Plot training and validation loss per epoch\n",
    "    plt.plot(epochs, loss, label = 'Training')\n",
    "    plt.plot(epochs, val_loss, label = 'Validation')\n",
    "    plt.title('Training and validation MSE loss')\n",
    "\n",
    "#     lines, labels = fig.axes[-1].get_legend_handles_labels()\n",
    "#     fig.legend(lines, labels, loc = 'center right')\n",
    "    plt.xticks(rotation=20)\n",
    "    plt.xlabel('epochs')\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Architecture and Deployment"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Callbacks"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "def scheduler(epoch, lr):\n",
    "    if epoch < 5:\n",
    "        return lr\n",
    "    else:\n",
    "        return lr * 0.9"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Training method"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "def compile_and_fit(model, train_data, train_target, val_data, val_target, \n",
    "                    test_data, test_target, max_epochs, batch_size, test_entry, label_width):\n",
    "\n",
    "    # Add callbacks\n",
    "    early_stop = tf.keras.callbacks.EarlyStopping(monitor = \"val_loss\", \n",
    "                                                  patience = 3)\n",
    "    lr_scheduler = tf.keras.callbacks.LearningRateScheduler(scheduler, verbose = 0)\n",
    "    callbacks = [early_stop, lr_scheduler]\n",
    "\n",
    "    # Model compile\n",
    "    model.compile(loss      = tf.losses.MeanSquaredError(),\n",
    "                  optimizer = tf.keras.optimizers.Nadam(),\n",
    "                  metrics   = [tf.metrics.MeanAbsoluteError()])\n",
    "\n",
    "    # Model fit\n",
    "    history = model.fit(train_data,\n",
    "                        train_target,\n",
    "                        batch_size = batch_size,\n",
    "                        epochs  = max_epochs,\n",
    "                        verbose = 1,\n",
    "                        shuffle = False, \n",
    "                        validation_data = (val_data, val_target))\n",
    "\n",
    "    model.summary()\n",
    "    \n",
    "    # Model evaluate\n",
    "    loss_eval, metric_eval = model.evaluate(test_data, test_target, verbose = 0)\n",
    "    print(\"Evaluation loss: {} ---- Evaluation metric:{}\".format(loss_eval, metric_eval))\n",
    "\n",
    "    # Test model\n",
    "    test_data = np.reshape(test_data[test_entry], (1, test_data[0].shape[0], test_data[0].shape[1]))\n",
    "    test_prediction = model.predict(test_data)\n",
    "    \n",
    "    test_prediction = np.squeeze(test_prediction)\n",
    "    \n",
    "    # Plot results\n",
    "    plot_results(label_target, test_target_time[test_entry], test_target[test_entry], test_prediction, label_width)\n",
    "\n",
    "    plot_history(history)\n",
    "\n",
    "    return history, test_prediction\n",
    "\n",
    "test_entry = np.random.randint(0, 100)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Models"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Linear model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "def f_linear_model(label_target, label_width):\n",
    "    model = tf.keras.Sequential([\n",
    "            tf.keras.layers.Flatten(),\n",
    "        \n",
    "#             tf.keras.layers.Dense(units = 300, kernel_initializer='random_normal', bias_initializer='zeros'),\n",
    "#             tf.keras.layers.Dropout(0.2),\n",
    "            \n",
    "            tf.keras.layers.Dense(units = 400, kernel_initializer='random_normal', bias_initializer='zeros'),\n",
    "#             tf.keras.layers.Dropout(0.1),\n",
    "        \n",
    "            tf.keras.layers.Dense(units = 200, kernel_initializer='random_normal', bias_initializer='zeros'),\n",
    "#             tf.keras.layers.Dropout(0.1),\n",
    "        \n",
    "            tf.keras.layers.Dense(units = len(label_target) * label_width)])\n",
    "    return model\n",
    "\n",
    "linear_model = f_linear_model(label_target, label_width)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 19784 samples, validate on 5466 samples\n",
      "Epoch 1/30\n",
      "19784/19784 [==============================] - 3s 131us/sample - loss: 13.2752 - mean_absolute_error: 2.0396 - val_loss: 12.6879 - val_mean_absolute_error: 2.9137\n",
      "Epoch 2/30\n",
      "19784/19784 [==============================] - 1s 69us/sample - loss: 1.3339 - mean_absolute_error: 0.6330 - val_loss: 7.8537 - val_mean_absolute_error: 2.2455\n",
      "Epoch 3/30\n",
      "19784/19784 [==============================] - 1s 68us/sample - loss: 0.5656 - mean_absolute_error: 0.4292 - val_loss: 4.0185 - val_mean_absolute_error: 1.5627\n",
      "Epoch 4/30\n",
      "19784/19784 [==============================] - 1s 69us/sample - loss: 0.3271 - mean_absolute_error: 0.3507 - val_loss: 4.4508 - val_mean_absolute_error: 1.7638\n",
      "Epoch 5/30\n",
      "19784/19784 [==============================] - 1s 68us/sample - loss: 0.2508 - mean_absolute_error: 0.3177 - val_loss: 4.3723 - val_mean_absolute_error: 1.7423\n",
      "Epoch 6/30\n",
      "19784/19784 [==============================] - 1s 69us/sample - loss: 0.1454 - mean_absolute_error: 0.2597 - val_loss: 1.4380 - val_mean_absolute_error: 0.9480\n",
      "Epoch 7/30\n",
      "19784/19784 [==============================] - 1s 70us/sample - loss: 0.1926 - mean_absolute_error: 0.2667 - val_loss: 1.7140 - val_mean_absolute_error: 1.0163\n",
      "Epoch 8/30\n",
      "19784/19784 [==============================] - 1s 71us/sample - loss: 0.1389 - mean_absolute_error: 0.2369 - val_loss: 3.0487 - val_mean_absolute_error: 1.3922\n",
      "Epoch 9/30\n",
      "19784/19784 [==============================] - 2s 79us/sample - loss: 0.1104 - mean_absolute_error: 0.2173 - val_loss: 2.7425 - val_mean_absolute_error: 1.3208\n",
      "Epoch 10/30\n",
      "19784/19784 [==============================] - 1s 72us/sample - loss: 0.0767 - mean_absolute_error: 0.1865 - val_loss: 1.1846 - val_mean_absolute_error: 0.8857\n",
      "Epoch 11/30\n",
      "19784/19784 [==============================] - 1s 74us/sample - loss: 0.0673 - mean_absolute_error: 0.1745 - val_loss: 0.7773 - val_mean_absolute_error: 0.7057\n",
      "Epoch 12/30\n",
      "19784/19784 [==============================] - 1s 73us/sample - loss: 0.0605 - mean_absolute_error: 0.1698 - val_loss: 0.9669 - val_mean_absolute_error: 0.7846\n",
      "Epoch 13/30\n",
      "19784/19784 [==============================] - 1s 69us/sample - loss: 0.0468 - mean_absolute_error: 0.1533 - val_loss: 0.2968 - val_mean_absolute_error: 0.4195\n",
      "Epoch 14/30\n",
      "19784/19784 [==============================] - 1s 68us/sample - loss: 0.0385 - mean_absolute_error: 0.1397 - val_loss: 0.3992 - val_mean_absolute_error: 0.5052\n",
      "Epoch 15/30\n",
      "19784/19784 [==============================] - 1s 69us/sample - loss: 0.0331 - mean_absolute_error: 0.1282 - val_loss: 0.2776 - val_mean_absolute_error: 0.4174\n",
      "Epoch 16/30\n",
      "19784/19784 [==============================] - 1s 67us/sample - loss: 0.0289 - mean_absolute_error: 0.1212 - val_loss: 0.3390 - val_mean_absolute_error: 0.4678\n",
      "Epoch 17/30\n",
      "19784/19784 [==============================] - 1s 67us/sample - loss: 0.0294 - mean_absolute_error: 0.1212 - val_loss: 0.1999 - val_mean_absolute_error: 0.3649\n",
      "Epoch 18/30\n",
      "19784/19784 [==============================] - 1s 74us/sample - loss: 0.0275 - mean_absolute_error: 0.1156 - val_loss: 0.2653 - val_mean_absolute_error: 0.4134\n",
      "Epoch 19/30\n",
      "19784/19784 [==============================] - 1s 72us/sample - loss: 0.0238 - mean_absolute_error: 0.1102 - val_loss: 0.2082 - val_mean_absolute_error: 0.3648\n",
      "Epoch 20/30\n",
      "19784/19784 [==============================] - 1s 70us/sample - loss: 0.0275 - mean_absolute_error: 0.1176 - val_loss: 0.1640 - val_mean_absolute_error: 0.3159\n",
      "Epoch 21/30\n",
      "19784/19784 [==============================] - 1s 71us/sample - loss: 0.0220 - mean_absolute_error: 0.1059 - val_loss: 0.1502 - val_mean_absolute_error: 0.3076\n",
      "Epoch 22/30\n",
      "19784/19784 [==============================] - 1s 68us/sample - loss: 0.0239 - mean_absolute_error: 0.1059 - val_loss: 0.1137 - val_mean_absolute_error: 0.2681\n",
      "Epoch 23/30\n",
      "19784/19784 [==============================] - 1s 70us/sample - loss: 0.0212 - mean_absolute_error: 0.0997 - val_loss: 0.1730 - val_mean_absolute_error: 0.3262\n",
      "Epoch 24/30\n",
      "19784/19784 [==============================] - 1s 70us/sample - loss: 0.0171 - mean_absolute_error: 0.0914 - val_loss: 0.1254 - val_mean_absolute_error: 0.2756\n",
      "Epoch 25/30\n",
      "19784/19784 [==============================] - 1s 71us/sample - loss: 402.8141 - mean_absolute_error: 6.4745 - val_loss: 258.2062 - val_mean_absolute_error: 11.8940\n",
      "Epoch 26/30\n",
      "19784/19784 [==============================] - 1s 73us/sample - loss: 9.8567 - mean_absolute_error: 2.0407 - val_loss: 1.4093 - val_mean_absolute_error: 0.9283\n",
      "Epoch 27/30\n",
      "19784/19784 [==============================] - 1s 72us/sample - loss: 1.0641 - mean_absolute_error: 0.7354 - val_loss: 1.9117 - val_mean_absolute_error: 1.1067\n",
      "Epoch 28/30\n",
      "19784/19784 [==============================] - 1s 71us/sample - loss: 0.5413 - mean_absolute_error: 0.5398 - val_loss: 1.7317 - val_mean_absolute_error: 1.0035\n",
      "Epoch 29/30\n",
      "19784/19784 [==============================] - 1s 71us/sample - loss: 0.3302 - mean_absolute_error: 0.4218 - val_loss: 2.0598 - val_mean_absolute_error: 1.0869\n",
      "Epoch 30/30\n",
      "19784/19784 [==============================] - 1s 66us/sample - loss: 0.2536 - mean_absolute_error: 0.3655 - val_loss: 1.2971 - val_mean_absolute_error: 0.8849\n",
      "Model: \"sequential\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "flatten (Flatten)            multiple                  0         \n",
      "_________________________________________________________________\n",
      "dense (Dense)                multiple                  3606800   \n",
      "_________________________________________________________________\n",
      "dense_1 (Dense)              multiple                  80200     \n",
      "_________________________________________________________________\n",
      "dense_2 (Dense)              multiple                  13065     \n",
      "=================================================================\n",
      "Total params: 3,700,065\n",
      "Trainable params: 3,700,065\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n",
      "Evaluation loss: 2.1554867218331197 ---- Evaluation metric:1.1439435482025146\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "max_epochs = 30\n",
    "batch_size = 100\n",
    "\n",
    "linear_history, linear_prediction = compile_and_fit(linear_model, train_data, train_target, val_data, val_target, \n",
    "                                                    test_data, test_target, max_epochs, batch_size, test_entry, label_width)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## MLP model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "def f_mlp_model(label_target, label_width):\n",
    "    model = tf.keras.Sequential([\n",
    "            tf.keras.layers.Flatten(),\n",
    "        \n",
    "#             tf.keras.layers.Dense(units = 800, activation = 'relu', kernel_initializer='random_normal', bias_initializer='zeros'),\n",
    "#             tf.keras.layers.Dropout(0.2),\n",
    "        \n",
    "            tf.keras.layers.Dense(units = 200, activation = 'relu', kernel_initializer='random_normal', bias_initializer='zeros'),\n",
    "            tf.keras.layers.Dropout(0.2),\n",
    "        \n",
    "            tf.keras.layers.Dense(units = 100, activation = 'relu', kernel_initializer='random_normal', bias_initializer='zeros'),\n",
    "            tf.keras.layers.Dropout(0.2),\n",
    "\n",
    "        \n",
    "            tf.keras.layers.Dense(units = len(label_target) * label_width)])\n",
    "    return model\n",
    "    \n",
    "mlp_model = f_mlp_model(label_target, label_width)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 19784 samples, validate on 5466 samples\n",
      "Epoch 1/30\n",
      "19784/19784 [==============================] - 2s 121us/sample - loss: 0.1297 - mean_absolute_error: 0.2624 - val_loss: 0.0127 - val_mean_absolute_error: 0.0974\n",
      "Epoch 2/30\n",
      "19784/19784 [==============================] - 1s 65us/sample - loss: 0.0242 - mean_absolute_error: 0.1221 - val_loss: 0.0141 - val_mean_absolute_error: 0.0977\n",
      "Epoch 3/30\n",
      "19784/19784 [==============================] - 1s 63us/sample - loss: 0.0154 - mean_absolute_error: 0.0967 - val_loss: 0.0117 - val_mean_absolute_error: 0.0873\n",
      "Epoch 4/30\n",
      "19784/19784 [==============================] - 1s 65us/sample - loss: 0.0124 - mean_absolute_error: 0.0865 - val_loss: 0.0044 - val_mean_absolute_error: 0.0528\n",
      "Epoch 5/30\n",
      "19784/19784 [==============================] - 1s 65us/sample - loss: 0.0074 - mean_absolute_error: 0.0661 - val_loss: 0.0022 - val_mean_absolute_error: 0.0445\n",
      "Epoch 6/30\n",
      "19784/19784 [==============================] - 1s 70us/sample - loss: 0.0055 - mean_absolute_error: 0.0548 - val_loss: 0.0029 - val_mean_absolute_error: 0.0513\n",
      "Epoch 7/30\n",
      "19784/19784 [==============================] - 1s 72us/sample - loss: 0.0049 - mean_absolute_error: 0.0514 - val_loss: 0.0028 - val_mean_absolute_error: 0.0503\n",
      "Epoch 8/30\n",
      "19784/19784 [==============================] - 1s 68us/sample - loss: 0.0045 - mean_absolute_error: 0.0485 - val_loss: 0.0028 - val_mean_absolute_error: 0.0512\n",
      "Epoch 9/30\n",
      "19784/19784 [==============================] - 1s 66us/sample - loss: 0.0042 - mean_absolute_error: 0.0467 - val_loss: 0.0028 - val_mean_absolute_error: 0.0510\n",
      "Epoch 10/30\n",
      "19784/19784 [==============================] - 1s 66us/sample - loss: 0.0041 - mean_absolute_error: 0.0453 - val_loss: 0.0028 - val_mean_absolute_error: 0.0503\n",
      "Epoch 11/30\n",
      "19784/19784 [==============================] - 1s 64us/sample - loss: 0.0039 - mean_absolute_error: 0.0439 - val_loss: 0.0027 - val_mean_absolute_error: 0.0497\n",
      "Epoch 12/30\n",
      "19784/19784 [==============================] - 1s 64us/sample - loss: 0.0037 - mean_absolute_error: 0.0424 - val_loss: 0.0028 - val_mean_absolute_error: 0.0508\n",
      "Epoch 13/30\n",
      "19784/19784 [==============================] - 1s 67us/sample - loss: 0.0036 - mean_absolute_error: 0.0415 - val_loss: 0.0027 - val_mean_absolute_error: 0.0499\n",
      "Epoch 14/30\n",
      "19784/19784 [==============================] - 1s 66us/sample - loss: 0.0035 - mean_absolute_error: 0.0407 - val_loss: 0.0029 - val_mean_absolute_error: 0.0520\n",
      "Epoch 15/30\n",
      "19784/19784 [==============================] - 1s 67us/sample - loss: 0.0034 - mean_absolute_error: 0.0396 - val_loss: 0.0030 - val_mean_absolute_error: 0.0525\n",
      "Epoch 16/30\n",
      "19784/19784 [==============================] - 1s 66us/sample - loss: 0.0033 - mean_absolute_error: 0.0388 - val_loss: 0.0030 - val_mean_absolute_error: 0.0527\n",
      "Epoch 17/30\n",
      "19784/19784 [==============================] - 1s 67us/sample - loss: 0.0032 - mean_absolute_error: 0.0378 - val_loss: 0.0030 - val_mean_absolute_error: 0.0531\n",
      "Epoch 18/30\n",
      "19784/19784 [==============================] - 1s 65us/sample - loss: 0.0031 - mean_absolute_error: 0.0370 - val_loss: 0.0030 - val_mean_absolute_error: 0.0530\n",
      "Epoch 19/30\n",
      "19784/19784 [==============================] - 1s 66us/sample - loss: 0.0031 - mean_absolute_error: 0.0363 - val_loss: 0.0030 - val_mean_absolute_error: 0.0525\n",
      "Epoch 20/30\n",
      "19784/19784 [==============================] - 1s 67us/sample - loss: 0.0030 - mean_absolute_error: 0.0357 - val_loss: 0.0032 - val_mean_absolute_error: 0.0542\n",
      "Epoch 21/30\n",
      "19784/19784 [==============================] - 1s 67us/sample - loss: 0.0030 - mean_absolute_error: 0.0351 - val_loss: 0.0030 - val_mean_absolute_error: 0.0531\n",
      "Epoch 22/30\n",
      "19784/19784 [==============================] - 1s 66us/sample - loss: 0.0029 - mean_absolute_error: 0.0343 - val_loss: 0.0032 - val_mean_absolute_error: 0.0546\n",
      "Epoch 23/30\n",
      "19784/19784 [==============================] - 1s 65us/sample - loss: 0.0029 - mean_absolute_error: 0.0340 - val_loss: 0.0030 - val_mean_absolute_error: 0.0523\n",
      "Epoch 24/30\n",
      "19784/19784 [==============================] - 1s 66us/sample - loss: 0.0028 - mean_absolute_error: 0.0335 - val_loss: 0.0032 - val_mean_absolute_error: 0.0545\n",
      "Epoch 25/30\n",
      "19784/19784 [==============================] - 1s 70us/sample - loss: 0.0028 - mean_absolute_error: 0.0330 - val_loss: 0.0031 - val_mean_absolute_error: 0.0541\n",
      "Epoch 26/30\n",
      "19784/19784 [==============================] - 1s 70us/sample - loss: 0.0028 - mean_absolute_error: 0.0329 - val_loss: 0.0031 - val_mean_absolute_error: 0.0539\n",
      "Epoch 27/30\n",
      "19784/19784 [==============================] - 1s 70us/sample - loss: 0.0027 - mean_absolute_error: 0.0319 - val_loss: 0.0030 - val_mean_absolute_error: 0.0528\n",
      "Epoch 28/30\n",
      "19784/19784 [==============================] - 1s 68us/sample - loss: 0.0027 - mean_absolute_error: 0.0320 - val_loss: 0.0031 - val_mean_absolute_error: 0.0539\n",
      "Epoch 29/30\n",
      "19784/19784 [==============================] - 1s 69us/sample - loss: 0.0027 - mean_absolute_error: 0.0315 - val_loss: 0.0031 - val_mean_absolute_error: 0.0535\n",
      "Epoch 30/30\n",
      "19784/19784 [==============================] - 1s 66us/sample - loss: 0.0026 - mean_absolute_error: 0.0312 - val_loss: 0.0031 - val_mean_absolute_error: 0.0540\n",
      "Model: \"sequential_1\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "flatten_1 (Flatten)          multiple                  0         \n",
      "_________________________________________________________________\n",
      "dense_3 (Dense)              multiple                  1803400   \n",
      "_________________________________________________________________\n",
      "dropout (Dropout)            multiple                  0         \n",
      "_________________________________________________________________\n",
      "dense_4 (Dense)              multiple                  20100     \n",
      "_________________________________________________________________\n",
      "dropout_1 (Dropout)          multiple                  0         \n",
      "_________________________________________________________________\n",
      "dense_5 (Dense)              multiple                  6565      \n",
      "=================================================================\n",
      "Total params: 1,830,065\n",
      "Trainable params: 1,830,065\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n",
      "Evaluation loss: 0.0019622195291658846 ---- Evaluation metric:0.04136662930250168\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "max_epochs = 30\n",
    "batch_size = 100\n",
    "mlp_history, mlp_prediction = compile_and_fit(mlp_model, train_data, train_target, val_data, val_target, \n",
    "                                              test_data, test_target, max_epochs, batch_size, test_entry, label_width)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## LSTM RNN model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "def f_lstm_model(input_shape, output_shape):\n",
    "    model = tf.keras.Sequential([\n",
    "            \n",
    "            tf.keras.layers.LSTM(units = 800, input_shape = input_shape, kernel_initializer='random_normal', bias_initializer='zeros', dropout=0.2, return_sequences=True),\n",
    "    \n",
    "            tf.keras.layers.LSTM(units = 50, input_shape = input_shape, kernel_initializer='random_normal', bias_initializer='zeros', dropout=0.2, return_sequences=False),\n",
    "            tf.keras.layers.Dense(output_shape)])\n",
    "    return model\n",
    "\n",
    "input_shape = train_data[0].shape\n",
    "output_shape = train_target.shape[1]\n",
    "lstm_model = f_lstm_model(input_shape, output_shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 19784 samples, validate on 5466 samples\n",
      "Epoch 1/20\n",
      "19784/19784 [==============================] - 35s 2ms/sample - loss: 0.0066 - mean_absolute_error: 0.0301 - val_loss: 8.4272e-04 - val_mean_absolute_error: 0.0256\n",
      "Epoch 2/20\n",
      "19784/19784 [==============================] - 29s 1ms/sample - loss: 9.5025e-04 - mean_absolute_error: 0.0110 - val_loss: 6.8954e-04 - val_mean_absolute_error: 0.0231\n",
      "Epoch 3/20\n",
      "19784/19784 [==============================] - 29s 1ms/sample - loss: 9.2630e-04 - mean_absolute_error: 0.0105 - val_loss: 6.2664e-04 - val_mean_absolute_error: 0.0220\n",
      "Epoch 4/20\n",
      "19784/19784 [==============================] - 29s 1ms/sample - loss: 9.2278e-04 - mean_absolute_error: 0.0104 - val_loss: 5.9901e-04 - val_mean_absolute_error: 0.0215\n",
      "Epoch 5/20\n",
      "19784/19784 [==============================] - 29s 1ms/sample - loss: 9.2903e-04 - mean_absolute_error: 0.0105 - val_loss: 5.8612e-04 - val_mean_absolute_error: 0.0213\n",
      "Epoch 6/20\n",
      "19784/19784 [==============================] - 29s 1ms/sample - loss: 9.4121e-04 - mean_absolute_error: 0.0106 - val_loss: 5.8460e-04 - val_mean_absolute_error: 0.0213\n",
      "Epoch 7/20\n",
      "19784/19784 [==============================] - 29s 1ms/sample - loss: 9.5769e-04 - mean_absolute_error: 0.0108 - val_loss: 5.8832e-04 - val_mean_absolute_error: 0.0213\n",
      "Epoch 8/20\n",
      "19784/19784 [==============================] - 29s 1ms/sample - loss: 9.7726e-04 - mean_absolute_error: 0.0110 - val_loss: 5.9669e-04 - val_mean_absolute_error: 0.0215\n",
      "Epoch 9/20\n",
      "19784/19784 [==============================] - 29s 1ms/sample - loss: 9.9968e-04 - mean_absolute_error: 0.0112 - val_loss: 6.0799e-04 - val_mean_absolute_error: 0.0217\n",
      "Epoch 10/20\n",
      "19784/19784 [==============================] - 30s 1ms/sample - loss: 0.0010 - mean_absolute_error: 0.0115 - val_loss: 6.2317e-04 - val_mean_absolute_error: 0.0220\n",
      "Epoch 11/20\n",
      "19784/19784 [==============================] - 29s 1ms/sample - loss: 0.0011 - mean_absolute_error: 0.0117 - val_loss: 6.4021e-04 - val_mean_absolute_error: 0.0223\n",
      "Epoch 12/20\n",
      "19784/19784 [==============================] - 30s 1ms/sample - loss: 0.0011 - mean_absolute_error: 0.0120 - val_loss: 6.5846e-04 - val_mean_absolute_error: 0.0227\n",
      "Epoch 13/20\n",
      "19784/19784 [==============================] - 29s 1ms/sample - loss: 0.0011 - mean_absolute_error: 0.0123 - val_loss: 6.7831e-04 - val_mean_absolute_error: 0.0231\n",
      "Epoch 14/20\n",
      "19784/19784 [==============================] - 29s 1ms/sample - loss: 0.0011 - mean_absolute_error: 0.0126 - val_loss: 6.9860e-04 - val_mean_absolute_error: 0.0234\n",
      "Epoch 15/20\n",
      "19784/19784 [==============================] - 29s 1ms/sample - loss: 0.0012 - mean_absolute_error: 0.0129 - val_loss: 7.1998e-04 - val_mean_absolute_error: 0.0238\n",
      "Epoch 16/20\n",
      "19784/19784 [==============================] - 30s 1ms/sample - loss: 0.0012 - mean_absolute_error: 0.0133 - val_loss: 7.4059e-04 - val_mean_absolute_error: 0.0242\n",
      "Epoch 17/20\n",
      "19784/19784 [==============================] - 29s 1ms/sample - loss: 0.0012 - mean_absolute_error: 0.0136 - val_loss: 7.6293e-04 - val_mean_absolute_error: 0.0246\n",
      "Epoch 18/20\n",
      "19784/19784 [==============================] - 29s 1ms/sample - loss: 0.0013 - mean_absolute_error: 0.0139 - val_loss: 7.8429e-04 - val_mean_absolute_error: 0.0250\n",
      "Epoch 19/20\n",
      "19784/19784 [==============================] - 29s 1ms/sample - loss: 0.0013 - mean_absolute_error: 0.0142 - val_loss: 8.0759e-04 - val_mean_absolute_error: 0.0254\n",
      "Epoch 20/20\n",
      "19784/19784 [==============================] - 30s 1ms/sample - loss: 0.0014 - mean_absolute_error: 0.0154 - val_loss: 8.7705e-04 - val_mean_absolute_error: 0.0265\n",
      "Model: \"sequential_2\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "lstm (LSTM)                  (None, 196, 800)          2710400   \n",
      "_________________________________________________________________\n",
      "lstm_1 (LSTM)                (None, 50)                170200    \n",
      "_________________________________________________________________\n",
      "dense_6 (Dense)              (None, 65)                3315      \n",
      "=================================================================\n",
      "Total params: 2,883,915\n",
      "Trainable params: 2,883,915\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n",
      "Evaluation loss: 0.0003773080524574333 ---- Evaluation metric:0.016086453571915627\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "max_epochs = 20\n",
    "batch_size = 50\n",
    "lstm_history, lstm_prediction = compile_and_fit(lstm_model, train_data, train_target, val_data, val_target, \n",
    "                                                test_data, test_target, max_epochs, batch_size, test_entry, label_width)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## GRU RNN model "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "def f_gru_model(input_shape, output_shape):\n",
    "    model = tf.keras.Sequential([\n",
    "            tf.keras.layers.GRU(units = 800, input_shape = input_shape, kernel_initializer='random_normal', bias_initializer='zeros', dropout=0.2, return_sequences=True),\n",
    "        \n",
    "            tf.keras.layers.GRU(units = 300, input_shape = input_shape, kernel_initializer='random_normal', bias_initializer='zeros', dropout=0.2, return_sequences=True),\n",
    "        \n",
    "            tf.keras.layers.GRU(units = 60, input_shape = input_shape, kernel_initializer='random_normal', bias_initializer='zeros', dropout=0.2, return_sequences=False),\n",
    "\n",
    "            tf.keras.layers.Dense(output_shape)])\n",
    "    return model\n",
    "\n",
    "input_shape = train_data[0].shape\n",
    "output_shape = train_target.shape[1]\n",
    "gru_model = f_gru_model(input_shape, output_shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 19784 samples, validate on 5466 samples\n",
      "Epoch 1/20\n",
      "19784/19784 [==============================] - 40s 2ms/sample - loss: 0.0071 - mean_absolute_error: 0.0333 - val_loss: 8.4678e-04 - val_mean_absolute_error: 0.0248\n",
      "Epoch 2/20\n",
      "19784/19784 [==============================] - 34s 2ms/sample - loss: 8.6369e-04 - mean_absolute_error: 0.0111 - val_loss: 7.0068e-04 - val_mean_absolute_error: 0.0226\n",
      "Epoch 3/20\n",
      "19784/19784 [==============================] - 34s 2ms/sample - loss: 8.8543e-04 - mean_absolute_error: 0.0103 - val_loss: 6.1911e-04 - val_mean_absolute_error: 0.0213\n",
      "Epoch 4/20\n",
      "19784/19784 [==============================] - 34s 2ms/sample - loss: 7.6964e-04 - mean_absolute_error: 0.0092 - val_loss: 6.1893e-04 - val_mean_absolute_error: 0.0212\n",
      "Epoch 5/20\n",
      "19784/19784 [==============================] - 34s 2ms/sample - loss: 7.7426e-04 - mean_absolute_error: 0.0092 - val_loss: 6.6929e-04 - val_mean_absolute_error: 0.0220\n",
      "Epoch 6/20\n",
      "19784/19784 [==============================] - 34s 2ms/sample - loss: 8.0659e-04 - mean_absolute_error: 0.0092 - val_loss: 6.3599e-04 - val_mean_absolute_error: 0.0213\n",
      "Epoch 7/20\n",
      "19784/19784 [==============================] - 34s 2ms/sample - loss: 8.0482e-04 - mean_absolute_error: 0.0090 - val_loss: 5.5303e-04 - val_mean_absolute_error: 0.0198\n",
      "Epoch 8/20\n",
      "19784/19784 [==============================] - 35s 2ms/sample - loss: 7.7026e-04 - mean_absolute_error: 0.0088 - val_loss: 9.2579e-04 - val_mean_absolute_error: 0.0261\n",
      "Epoch 9/20\n",
      "19784/19784 [==============================] - 34s 2ms/sample - loss: 8.0798e-04 - mean_absolute_error: 0.0091 - val_loss: 5.2069e-04 - val_mean_absolute_error: 0.0194\n",
      "Epoch 10/20\n",
      "19784/19784 [==============================] - 34s 2ms/sample - loss: 8.7000e-04 - mean_absolute_error: 0.0091 - val_loss: 3.9130e-04 - val_mean_absolute_error: 0.0168\n",
      "Epoch 11/20\n",
      "19784/19784 [==============================] - 34s 2ms/sample - loss: 8.1901e-04 - mean_absolute_error: 0.0087 - val_loss: 6.2225e-04 - val_mean_absolute_error: 0.0210\n",
      "Epoch 12/20\n",
      "19784/19784 [==============================] - 34s 2ms/sample - loss: 8.0940e-04 - mean_absolute_error: 0.0089 - val_loss: 5.2996e-04 - val_mean_absolute_error: 0.0196\n",
      "Epoch 13/20\n",
      "19784/19784 [==============================] - 34s 2ms/sample - loss: 0.0011 - mean_absolute_error: 0.0107 - val_loss: 3.5058e-04 - val_mean_absolute_error: 0.0159\n",
      "Epoch 14/20\n",
      "19784/19784 [==============================] - 34s 2ms/sample - loss: 8.5033e-04 - mean_absolute_error: 0.0091 - val_loss: 3.5319e-04 - val_mean_absolute_error: 0.0160\n",
      "Epoch 15/20\n",
      "19784/19784 [==============================] - 34s 2ms/sample - loss: 8.6528e-04 - mean_absolute_error: 0.0092 - val_loss: 3.5882e-04 - val_mean_absolute_error: 0.0161\n",
      "Epoch 16/20\n",
      "19784/19784 [==============================] - 34s 2ms/sample - loss: 8.8156e-04 - mean_absolute_error: 0.0094 - val_loss: 3.6321e-04 - val_mean_absolute_error: 0.0162\n",
      "Epoch 17/20\n",
      "19784/19784 [==============================] - 34s 2ms/sample - loss: 8.9905e-04 - mean_absolute_error: 0.0096 - val_loss: 3.7159e-04 - val_mean_absolute_error: 0.0164\n",
      "Epoch 18/20\n",
      "19784/19784 [==============================] - 34s 2ms/sample - loss: 9.1518e-04 - mean_absolute_error: 0.0098 - val_loss: 3.6975e-04 - val_mean_absolute_error: 0.0163\n",
      "Epoch 19/20\n",
      "19784/19784 [==============================] - 34s 2ms/sample - loss: 9.4329e-04 - mean_absolute_error: 0.0099 - val_loss: 3.7258e-04 - val_mean_absolute_error: 0.0164\n",
      "Epoch 20/20\n",
      "19784/19784 [==============================] - 34s 2ms/sample - loss: 9.4912e-04 - mean_absolute_error: 0.0100 - val_loss: 3.7389e-04 - val_mean_absolute_error: 0.0164\n",
      "Model: \"sequential_3\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "gru (GRU)                    (None, 196, 800)          2035200   \n",
      "_________________________________________________________________\n",
      "gru_1 (GRU)                  (None, 196, 300)          991800    \n",
      "_________________________________________________________________\n",
      "gru_2 (GRU)                  (None, 60)                65160     \n",
      "_________________________________________________________________\n",
      "dense_7 (Dense)              (None, 65)                3965      \n",
      "=================================================================\n",
      "Total params: 3,096,125\n",
      "Trainable params: 3,096,125\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n",
      "Evaluation loss: 0.00024015112276045198 ---- Evaluation metric:0.012474852614104748\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "max_epochs = 20\n",
    "batch_size = 50\n",
    "gru_history, gru_prediction = compile_and_fit(gru_model, train_data, train_target, val_data, val_target, \n",
    "                                              test_data, test_target, max_epochs, batch_size, test_entry, label_width)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot_results3(label_target, label_time, label, results_linear, results_mlp, results_lstm, results_gru, label_width):\n",
    "    n_label = np.int(len(label_target))\n",
    "\n",
    "    sensor_tag = list(label_target.values())\n",
    "    \n",
    "    fig, ax = plt.subplots(n_label, sharex = 'all')\n",
    "    \n",
    "    for i in range(n_label):\n",
    "\n",
    "        label_plot = label[((i) * label_width): ((i + 1)* label_width)]\n",
    "        ax.plot(label_time, label_plot, color = 'blue', label = 'Real')\n",
    "        \n",
    "#         results_plot_linear = results_linear[((i) * label_width): ((i + 1)* label_width)]\n",
    "#         ax.plot(label_time, results_plot_linear, color = 'orange', label = 'Linear')\n",
    "        \n",
    "        results_plot_mlp = results_mlp[((i) * label_width): ((i + 1)* label_width)]\n",
    "        ax.plot(label_time, results_plot_mlp, color = 'purple', label = 'MLP')\n",
    "        \n",
    "        results_plot_lstm = results_lstm[((i) * label_width): ((i + 1)* label_width)]\n",
    "        ax.plot(label_time, results_plot_lstm, color = 'black', label = 'LSTM')\n",
    "        \n",
    "        results_plot_gru = results_gru[((i) * label_width): ((i + 1)* label_width)]\n",
    "        ax.plot(label_time, results_plot_gru, color = 'green', label = 'GRU')\n",
    "\n",
    "        ax.set_ylabel(sensor_tag[i])\n",
    "\n",
    "    lines, labels = fig.axes[-1].get_legend_handles_labels()\n",
    "    fig.legend(lines, labels, loc = 'center right')\n",
    "    plt.xticks(rotation=20)\n",
    "    plt.xlabel('Day - Time')\n",
    "    plt.ylim([0.65, 0.75])\n",
    "    plt.show()\n",
    "\n",
    "    pass"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1008x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_results3(label_target, test_target_time[test_entry], test_target[test_entry], linear_prediction, mlp_prediction, lstm_prediction, gru_prediction, label_width)"
   ]
  },
  {
   "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": "3.6.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}