biLSTM_1step_daily.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "biLSTM_1step_daily.ipynb",
      "provenance": [],
      "collapsed_sections": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "BexZygSl10H4"
      },
      "source": [
        "references:\n",
        "\n",
        "\n",
        "*   Multivariate Time Series Forecasting with LSTMs in Keras (machinelearningmastery.com)\n",
        "*   https://www.youtube.com/watch?v=4FmVIpcwl4k&t=1117s\n",
        "\n",
        "\n",
        "*   BOOK: Hands-on time series analysis\n",
        "*   List item\n",
        "\n",
        "\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Xy6iMFeZL9gB"
      },
      "source": [
        "from numpy.random import seed\n",
        "seed(1)\n",
        "import tensorflow as tf\n",
        "tf.compat.v1.set_random_seed(2)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_GY3HgVB7T8D"
      },
      "source": [
        "import numpy as np\n",
        "import pandas as pd\n",
        "import matplotlib as mpl\n",
        "import matplotlib.pyplot as plt\n",
        "from keras.constraints import maxnorm\n",
        "from sklearn.model_selection import train_test_split\n",
        "from keras.preprocessing.sequence import TimeseriesGenerator\n",
        "from sklearn.preprocessing import MinMaxScaler, StandardScaler\n",
        "import tensorflow as tf\n",
        "from keras.models import load_model\n",
        "from keras.wrappers.scikit_learn import KerasClassifier\n",
        "from sklearn.model_selection import GridSearchCV\n",
        "from sklearn.model_selection import TimeSeriesSplit\n",
        "from sklearn.model_selection import cross_val_score\n",
        "\n",
        "mpl.rcParams['figure.figsize'] = (20, 20)\n",
        "mpl.rcParams['axes.grid'] = False"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "tj0hXjtjrJYT"
      },
      "source": [
        "from math import sqrt\n",
        "from numpy import concatenate\n",
        "from matplotlib import pyplot\n",
        "from pandas import read_csv\n",
        "from pandas import DataFrame\n",
        "from pandas import concat\n",
        "from sklearn.preprocessing import MinMaxScaler\n",
        "from sklearn.preprocessing import LabelEncoder\n",
        "from sklearn.metrics import mean_squared_error\n",
        "from keras.models import Sequential\n",
        "from keras.layers import Dense,Dropout\n",
        "from keras.layers import LSTM,Bidirectional"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "lmNbWyCJ7n0a"
      },
      "source": [
        "df = pd.read_csv('/content/clean_daily_df_human_logic_supervised.csv')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4MZJsxPq76CN",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 225
        },
        "outputId": "39c5733d-a586-4b91-bc98-935a29678bd7"
      },
      "source": [
        "df.head()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>time</th>\n",
              "      <th>MIN_TEMPERATURE_C_t_7</th>\n",
              "      <th>clients_day_t_7</th>\n",
              "      <th>Holiday_Food_Related_t_7</th>\n",
              "      <th>consommation_t_7</th>\n",
              "      <th>pdl_total_daily_load_t_7</th>\n",
              "      <th>max_pdl_daily_load_t_7</th>\n",
              "      <th>weekend_t_7</th>\n",
              "      <th>MIN_TEMPERATURE_C_t_1</th>\n",
              "      <th>clients_day_t_1</th>\n",
              "      <th>Holiday_Food_Related_t_1</th>\n",
              "      <th>consommation_t_1</th>\n",
              "      <th>pdl_total_daily_load_t_1</th>\n",
              "      <th>max_pdl_daily_load_t_1</th>\n",
              "      <th>weekend_t_1</th>\n",
              "      <th>MIN_TEMPERATURE_C_t</th>\n",
              "      <th>clients_day_t</th>\n",
              "      <th>Holiday_Food_Related_t</th>\n",
              "      <th>consommation_t</th>\n",
              "      <th>pdl_total_daily_load_t</th>\n",
              "      <th>weekend_t</th>\n",
              "      <th>max_pdl_daily_load_t</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>08/01/2015</td>\n",
              "      <td>8</td>\n",
              "      <td>3970</td>\n",
              "      <td>1</td>\n",
              "      <td>7068.81</td>\n",
              "      <td>16580</td>\n",
              "      <td>1433.0</td>\n",
              "      <td>0</td>\n",
              "      <td>13</td>\n",
              "      <td>3972</td>\n",
              "      <td>0</td>\n",
              "      <td>7468.64</td>\n",
              "      <td>20928</td>\n",
              "      <td>1319.0</td>\n",
              "      <td>0</td>\n",
              "      <td>13</td>\n",
              "      <td>3972</td>\n",
              "      <td>0</td>\n",
              "      <td>7464.41</td>\n",
              "      <td>20809</td>\n",
              "      <td>0</td>\n",
              "      <td>1379.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>09/01/2015</td>\n",
              "      <td>12</td>\n",
              "      <td>3970</td>\n",
              "      <td>0</td>\n",
              "      <td>7074.65</td>\n",
              "      <td>22006</td>\n",
              "      <td>1439.0</td>\n",
              "      <td>0</td>\n",
              "      <td>13</td>\n",
              "      <td>3972</td>\n",
              "      <td>0</td>\n",
              "      <td>7464.41</td>\n",
              "      <td>20809</td>\n",
              "      <td>1379.0</td>\n",
              "      <td>0</td>\n",
              "      <td>13</td>\n",
              "      <td>3972</td>\n",
              "      <td>0</td>\n",
              "      <td>7464.41</td>\n",
              "      <td>20803</td>\n",
              "      <td>0</td>\n",
              "      <td>1370.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>10/01/2015</td>\n",
              "      <td>12</td>\n",
              "      <td>3970</td>\n",
              "      <td>1</td>\n",
              "      <td>7074.65</td>\n",
              "      <td>18629</td>\n",
              "      <td>1127.0</td>\n",
              "      <td>1</td>\n",
              "      <td>13</td>\n",
              "      <td>3972</td>\n",
              "      <td>0</td>\n",
              "      <td>7464.41</td>\n",
              "      <td>20803</td>\n",
              "      <td>1370.0</td>\n",
              "      <td>0</td>\n",
              "      <td>14</td>\n",
              "      <td>3972</td>\n",
              "      <td>0</td>\n",
              "      <td>7446.13</td>\n",
              "      <td>18059</td>\n",
              "      <td>1</td>\n",
              "      <td>1181.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>11/01/2015</td>\n",
              "      <td>15</td>\n",
              "      <td>3971</td>\n",
              "      <td>0</td>\n",
              "      <td>7075.31</td>\n",
              "      <td>18060</td>\n",
              "      <td>1099.0</td>\n",
              "      <td>1</td>\n",
              "      <td>14</td>\n",
              "      <td>3972</td>\n",
              "      <td>0</td>\n",
              "      <td>7446.13</td>\n",
              "      <td>18059</td>\n",
              "      <td>1181.0</td>\n",
              "      <td>1</td>\n",
              "      <td>16</td>\n",
              "      <td>3972</td>\n",
              "      <td>0</td>\n",
              "      <td>7058.37</td>\n",
              "      <td>17478</td>\n",
              "      <td>1</td>\n",
              "      <td>1093.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>12/01/2015</td>\n",
              "      <td>15</td>\n",
              "      <td>3971</td>\n",
              "      <td>0</td>\n",
              "      <td>7083.30</td>\n",
              "      <td>20122</td>\n",
              "      <td>1288.0</td>\n",
              "      <td>0</td>\n",
              "      <td>16</td>\n",
              "      <td>3972</td>\n",
              "      <td>0</td>\n",
              "      <td>7058.37</td>\n",
              "      <td>17478</td>\n",
              "      <td>1093.0</td>\n",
              "      <td>1</td>\n",
              "      <td>14</td>\n",
              "      <td>3979</td>\n",
              "      <td>0</td>\n",
              "      <td>7217.96</td>\n",
              "      <td>18556</td>\n",
              "      <td>0</td>\n",
              "      <td>1205.0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "         time  MIN_TEMPERATURE_C_t_7  ...  weekend_t  max_pdl_daily_load_t\n",
              "0  08/01/2015                      8  ...          0                1379.0\n",
              "1  09/01/2015                     12  ...          0                1370.0\n",
              "2  10/01/2015                     12  ...          1                1181.0\n",
              "3  11/01/2015                     15  ...          1                1093.0\n",
              "4  12/01/2015                     15  ...          0                1205.0\n",
              "\n",
              "[5 rows x 22 columns]"
            ]
          },
          "metadata": {},
          "execution_count": 5
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 225
        },
        "id": "r1Nv6sDzc6DA",
        "outputId": "b032a311-2ff3-4dca-d484-e5821cb307a0"
      },
      "source": [
        "df.tail()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>time</th>\n",
              "      <th>MIN_TEMPERATURE_C_t_7</th>\n",
              "      <th>clients_day_t_7</th>\n",
              "      <th>Holiday_Food_Related_t_7</th>\n",
              "      <th>consommation_t_7</th>\n",
              "      <th>pdl_total_daily_load_t_7</th>\n",
              "      <th>max_pdl_daily_load_t_7</th>\n",
              "      <th>weekend_t_7</th>\n",
              "      <th>MIN_TEMPERATURE_C_t_1</th>\n",
              "      <th>clients_day_t_1</th>\n",
              "      <th>Holiday_Food_Related_t_1</th>\n",
              "      <th>consommation_t_1</th>\n",
              "      <th>pdl_total_daily_load_t_1</th>\n",
              "      <th>max_pdl_daily_load_t_1</th>\n",
              "      <th>weekend_t_1</th>\n",
              "      <th>MIN_TEMPERATURE_C_t</th>\n",
              "      <th>clients_day_t</th>\n",
              "      <th>Holiday_Food_Related_t</th>\n",
              "      <th>consommation_t</th>\n",
              "      <th>pdl_total_daily_load_t</th>\n",
              "      <th>weekend_t</th>\n",
              "      <th>max_pdl_daily_load_t</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>2027</th>\n",
              "      <td>27/07/2020</td>\n",
              "      <td>24</td>\n",
              "      <td>5189</td>\n",
              "      <td>0</td>\n",
              "      <td>5719.4</td>\n",
              "      <td>12627</td>\n",
              "      <td>924.0</td>\n",
              "      <td>0</td>\n",
              "      <td>27</td>\n",
              "      <td>5189</td>\n",
              "      <td>0</td>\n",
              "      <td>5719.4</td>\n",
              "      <td>11167</td>\n",
              "      <td>781.0</td>\n",
              "      <td>1</td>\n",
              "      <td>28</td>\n",
              "      <td>5189</td>\n",
              "      <td>0</td>\n",
              "      <td>5719.4</td>\n",
              "      <td>13174</td>\n",
              "      <td>0</td>\n",
              "      <td>1010.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2028</th>\n",
              "      <td>28/07/2020</td>\n",
              "      <td>25</td>\n",
              "      <td>5189</td>\n",
              "      <td>0</td>\n",
              "      <td>5719.4</td>\n",
              "      <td>12536</td>\n",
              "      <td>935.0</td>\n",
              "      <td>0</td>\n",
              "      <td>28</td>\n",
              "      <td>5189</td>\n",
              "      <td>0</td>\n",
              "      <td>5719.4</td>\n",
              "      <td>13174</td>\n",
              "      <td>1010.0</td>\n",
              "      <td>0</td>\n",
              "      <td>28</td>\n",
              "      <td>5189</td>\n",
              "      <td>0</td>\n",
              "      <td>5719.4</td>\n",
              "      <td>13097</td>\n",
              "      <td>0</td>\n",
              "      <td>986.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2029</th>\n",
              "      <td>29/07/2020</td>\n",
              "      <td>25</td>\n",
              "      <td>5189</td>\n",
              "      <td>0</td>\n",
              "      <td>5719.4</td>\n",
              "      <td>13369</td>\n",
              "      <td>1023.0</td>\n",
              "      <td>0</td>\n",
              "      <td>28</td>\n",
              "      <td>5189</td>\n",
              "      <td>0</td>\n",
              "      <td>5719.4</td>\n",
              "      <td>13097</td>\n",
              "      <td>986.0</td>\n",
              "      <td>0</td>\n",
              "      <td>27</td>\n",
              "      <td>5189</td>\n",
              "      <td>0</td>\n",
              "      <td>5719.4</td>\n",
              "      <td>12000</td>\n",
              "      <td>0</td>\n",
              "      <td>873.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2030</th>\n",
              "      <td>30/07/2020</td>\n",
              "      <td>26</td>\n",
              "      <td>5189</td>\n",
              "      <td>0</td>\n",
              "      <td>5719.4</td>\n",
              "      <td>13587</td>\n",
              "      <td>1027.0</td>\n",
              "      <td>0</td>\n",
              "      <td>27</td>\n",
              "      <td>5189</td>\n",
              "      <td>0</td>\n",
              "      <td>5719.4</td>\n",
              "      <td>12000</td>\n",
              "      <td>873.0</td>\n",
              "      <td>0</td>\n",
              "      <td>27</td>\n",
              "      <td>5189</td>\n",
              "      <td>0</td>\n",
              "      <td>5719.4</td>\n",
              "      <td>11604</td>\n",
              "      <td>0</td>\n",
              "      <td>762.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2031</th>\n",
              "      <td>31/07/2020</td>\n",
              "      <td>27</td>\n",
              "      <td>5189</td>\n",
              "      <td>0</td>\n",
              "      <td>5719.4</td>\n",
              "      <td>13226</td>\n",
              "      <td>1017.0</td>\n",
              "      <td>0</td>\n",
              "      <td>27</td>\n",
              "      <td>5189</td>\n",
              "      <td>0</td>\n",
              "      <td>5719.4</td>\n",
              "      <td>11604</td>\n",
              "      <td>762.0</td>\n",
              "      <td>0</td>\n",
              "      <td>28</td>\n",
              "      <td>5189</td>\n",
              "      <td>1</td>\n",
              "      <td>5719.4</td>\n",
              "      <td>8935</td>\n",
              "      <td>0</td>\n",
              "      <td>528.0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "            time  MIN_TEMPERATURE_C_t_7  ...  weekend_t  max_pdl_daily_load_t\n",
              "2027  27/07/2020                     24  ...          0                1010.0\n",
              "2028  28/07/2020                     25  ...          0                 986.0\n",
              "2029  29/07/2020                     25  ...          0                 873.0\n",
              "2030  30/07/2020                     26  ...          0                 762.0\n",
              "2031  31/07/2020                     27  ...          0                 528.0\n",
              "\n",
              "[5 rows x 22 columns]"
            ]
          },
          "metadata": {},
          "execution_count": 6
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "e_gfpzvA0cJa",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "6fdd4c98-6e0b-4994-94f0-53e3691566c2"
      },
      "source": [
        "df.info()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "<class 'pandas.core.frame.DataFrame'>\n",
            "RangeIndex: 2032 entries, 0 to 2031\n",
            "Data columns (total 22 columns):\n",
            " #   Column                    Non-Null Count  Dtype  \n",
            "---  ------                    --------------  -----  \n",
            " 0   time                      2032 non-null   object \n",
            " 1   MIN_TEMPERATURE_C_t_7     2032 non-null   int64  \n",
            " 2   clients_day_t_7           2032 non-null   int64  \n",
            " 3   Holiday_Food_Related_t_7  2032 non-null   int64  \n",
            " 4   consommation_t_7          2032 non-null   float64\n",
            " 5   pdl_total_daily_load_t_7  2032 non-null   int64  \n",
            " 6   max_pdl_daily_load_t_7    2032 non-null   float64\n",
            " 7   weekend_t_7               2032 non-null   int64  \n",
            " 8   MIN_TEMPERATURE_C_t_1     2032 non-null   int64  \n",
            " 9   clients_day_t_1           2032 non-null   int64  \n",
            " 10  Holiday_Food_Related_t_1  2032 non-null   int64  \n",
            " 11  consommation_t_1          2032 non-null   float64\n",
            " 12  pdl_total_daily_load_t_1  2032 non-null   int64  \n",
            " 13  max_pdl_daily_load_t_1    2032 non-null   float64\n",
            " 14  weekend_t_1               2032 non-null   int64  \n",
            " 15  MIN_TEMPERATURE_C_t       2032 non-null   int64  \n",
            " 16  clients_day_t             2032 non-null   int64  \n",
            " 17  Holiday_Food_Related_t    2032 non-null   int64  \n",
            " 18  consommation_t            2032 non-null   float64\n",
            " 19  pdl_total_daily_load_t    2032 non-null   int64  \n",
            " 20  weekend_t                 2032 non-null   int64  \n",
            " 21  max_pdl_daily_load_t      2032 non-null   float64\n",
            "dtypes: float64(6), int64(15), object(1)\n",
            "memory usage: 349.4+ KB\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "DX4Wjmy08aWn"
      },
      "source": [
        "time = pd.to_datetime(df['time'], dayfirst=True).dropna()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "s3QqtUCsyKLa",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 319
        },
        "outputId": "6c66fc4e-a79f-4df8-adfa-d9c4f5f2e179"
      },
      "source": [
        "df.describe()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>MIN_TEMPERATURE_C_t_7</th>\n",
              "      <th>clients_day_t_7</th>\n",
              "      <th>Holiday_Food_Related_t_7</th>\n",
              "      <th>consommation_t_7</th>\n",
              "      <th>pdl_total_daily_load_t_7</th>\n",
              "      <th>max_pdl_daily_load_t_7</th>\n",
              "      <th>weekend_t_7</th>\n",
              "      <th>MIN_TEMPERATURE_C_t_1</th>\n",
              "      <th>clients_day_t_1</th>\n",
              "      <th>Holiday_Food_Related_t_1</th>\n",
              "      <th>consommation_t_1</th>\n",
              "      <th>pdl_total_daily_load_t_1</th>\n",
              "      <th>max_pdl_daily_load_t_1</th>\n",
              "      <th>weekend_t_1</th>\n",
              "      <th>MIN_TEMPERATURE_C_t</th>\n",
              "      <th>clients_day_t</th>\n",
              "      <th>Holiday_Food_Related_t</th>\n",
              "      <th>consommation_t</th>\n",
              "      <th>pdl_total_daily_load_t</th>\n",
              "      <th>weekend_t</th>\n",
              "      <th>max_pdl_daily_load_t</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>count</th>\n",
              "      <td>2032.000000</td>\n",
              "      <td>2032.000000</td>\n",
              "      <td>2032.000000</td>\n",
              "      <td>2032.000000</td>\n",
              "      <td>2032.000000</td>\n",
              "      <td>2032.000000</td>\n",
              "      <td>2032.000000</td>\n",
              "      <td>2032.000000</td>\n",
              "      <td>2032.000000</td>\n",
              "      <td>2032.000000</td>\n",
              "      <td>2032.000000</td>\n",
              "      <td>2032.000000</td>\n",
              "      <td>2032.000000</td>\n",
              "      <td>2032.000000</td>\n",
              "      <td>2032.000000</td>\n",
              "      <td>2032.000000</td>\n",
              "      <td>2032.000000</td>\n",
              "      <td>2032.000000</td>\n",
              "      <td>2032.000000</td>\n",
              "      <td>2032.000000</td>\n",
              "      <td>2032.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>mean</th>\n",
              "      <td>20.633366</td>\n",
              "      <td>4761.092028</td>\n",
              "      <td>0.113681</td>\n",
              "      <td>7498.069537</td>\n",
              "      <td>13720.760827</td>\n",
              "      <td>1082.264365</td>\n",
              "      <td>0.285433</td>\n",
              "      <td>20.678150</td>\n",
              "      <td>4764.689961</td>\n",
              "      <td>0.112697</td>\n",
              "      <td>7494.061850</td>\n",
              "      <td>13700.614665</td>\n",
              "      <td>1081.028637</td>\n",
              "      <td>0.285433</td>\n",
              "      <td>20.685531</td>\n",
              "      <td>4765.288878</td>\n",
              "      <td>0.113189</td>\n",
              "      <td>7493.201004</td>\n",
              "      <td>13694.712598</td>\n",
              "      <td>0.285433</td>\n",
              "      <td>1080.639365</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>std</th>\n",
              "      <td>5.917364</td>\n",
              "      <td>435.595628</td>\n",
              "      <td>0.317502</td>\n",
              "      <td>1189.333584</td>\n",
              "      <td>4953.021415</td>\n",
              "      <td>209.215731</td>\n",
              "      <td>0.451732</td>\n",
              "      <td>5.910439</td>\n",
              "      <td>434.079536</td>\n",
              "      <td>0.316300</td>\n",
              "      <td>1193.029654</td>\n",
              "      <td>4944.948179</td>\n",
              "      <td>209.136835</td>\n",
              "      <td>0.451732</td>\n",
              "      <td>5.910212</td>\n",
              "      <td>433.824788</td>\n",
              "      <td>0.316902</td>\n",
              "      <td>1193.678924</td>\n",
              "      <td>4943.474583</td>\n",
              "      <td>0.451732</td>\n",
              "      <td>209.429624</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>min</th>\n",
              "      <td>8.000000</td>\n",
              "      <td>3970.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>4664.350000</td>\n",
              "      <td>3197.000000</td>\n",
              "      <td>404.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>8.000000</td>\n",
              "      <td>3972.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>4664.350000</td>\n",
              "      <td>3197.000000</td>\n",
              "      <td>404.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>8.000000</td>\n",
              "      <td>3972.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>4664.350000</td>\n",
              "      <td>3197.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>404.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>25%</th>\n",
              "      <td>16.000000</td>\n",
              "      <td>4355.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>6750.780000</td>\n",
              "      <td>10185.750000</td>\n",
              "      <td>936.750000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>16.000000</td>\n",
              "      <td>4356.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>6747.520000</td>\n",
              "      <td>10185.750000</td>\n",
              "      <td>935.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>16.000000</td>\n",
              "      <td>4356.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>6747.520000</td>\n",
              "      <td>10174.500000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>935.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>50%</th>\n",
              "      <td>20.000000</td>\n",
              "      <td>4795.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>7383.670000</td>\n",
              "      <td>12869.000000</td>\n",
              "      <td>1073.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>20.000000</td>\n",
              "      <td>4799.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>7383.670000</td>\n",
              "      <td>12855.000000</td>\n",
              "      <td>1071.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>20.000000</td>\n",
              "      <td>4801.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>7381.705000</td>\n",
              "      <td>12855.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>1071.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>75%</th>\n",
              "      <td>26.000000</td>\n",
              "      <td>5189.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>8125.475000</td>\n",
              "      <td>17240.500000</td>\n",
              "      <td>1243.717949</td>\n",
              "      <td>1.000000</td>\n",
              "      <td>26.000000</td>\n",
              "      <td>5189.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>8125.475000</td>\n",
              "      <td>17198.750000</td>\n",
              "      <td>1243.000000</td>\n",
              "      <td>1.000000</td>\n",
              "      <td>26.000000</td>\n",
              "      <td>5189.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>8125.475000</td>\n",
              "      <td>17174.000000</td>\n",
              "      <td>1.000000</td>\n",
              "      <td>1242.250000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>max</th>\n",
              "      <td>34.000000</td>\n",
              "      <td>5354.000000</td>\n",
              "      <td>1.000000</td>\n",
              "      <td>10771.150000</td>\n",
              "      <td>27936.000000</td>\n",
              "      <td>1628.000000</td>\n",
              "      <td>1.000000</td>\n",
              "      <td>34.000000</td>\n",
              "      <td>5354.000000</td>\n",
              "      <td>1.000000</td>\n",
              "      <td>10771.150000</td>\n",
              "      <td>27936.000000</td>\n",
              "      <td>1628.000000</td>\n",
              "      <td>1.000000</td>\n",
              "      <td>34.000000</td>\n",
              "      <td>5354.000000</td>\n",
              "      <td>1.000000</td>\n",
              "      <td>10771.150000</td>\n",
              "      <td>27936.000000</td>\n",
              "      <td>1.000000</td>\n",
              "      <td>1628.000000</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "       MIN_TEMPERATURE_C_t_7  ...  max_pdl_daily_load_t\n",
              "count            2032.000000  ...           2032.000000\n",
              "mean               20.633366  ...           1080.639365\n",
              "std                 5.917364  ...            209.429624\n",
              "min                 8.000000  ...            404.000000\n",
              "25%                16.000000  ...            935.000000\n",
              "50%                20.000000  ...           1071.000000\n",
              "75%                26.000000  ...           1242.250000\n",
              "max                34.000000  ...           1628.000000\n",
              "\n",
              "[8 rows x 21 columns]"
            ]
          },
          "metadata": {},
          "execution_count": 9
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "uVuqU0u0dmbG",
        "outputId": "93c26fab-ae08-4a50-b10d-582875a13f32"
      },
      "source": [
        "df.columns"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Index(['time', 'MIN_TEMPERATURE_C_t_7', 'clients_day_t_7',\n",
              "       'Holiday_Food_Related_t_7', 'consommation_t_7',\n",
              "       'pdl_total_daily_load_t_7', 'max_pdl_daily_load_t_7', 'weekend_t_7',\n",
              "       'MIN_TEMPERATURE_C_t_1', 'clients_day_t_1', 'Holiday_Food_Related_t_1',\n",
              "       'consommation_t_1', 'pdl_total_daily_load_t_1',\n",
              "       'max_pdl_daily_load_t_1', 'weekend_t_1', 'MIN_TEMPERATURE_C_t',\n",
              "       'clients_day_t', 'Holiday_Food_Related_t', 'consommation_t',\n",
              "       'pdl_total_daily_load_t', 'weekend_t', 'max_pdl_daily_load_t'],\n",
              "      dtype='object')"
            ]
          },
          "metadata": {},
          "execution_count": 10
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "0YVVz0bEduD6"
      },
      "source": [
        "df_input=df[['MIN_TEMPERATURE_C_t_7',\n",
        "        'clients_day_t_7','Holiday_Food_Related_t_7', 'consommation_t_7',\n",
        "       'pdl_total_daily_load_t_7', 'max_pdl_daily_load_t_7', 'weekend_t_7',\n",
        "\n",
        "       'MIN_TEMPERATURE_C_t_1', 'clients_day_t_1', 'Holiday_Food_Related_t_1',\n",
        "       'consommation_t_1', 'pdl_total_daily_load_t_1',\n",
        "       'max_pdl_daily_load_t_1', 'weekend_t_1',\n",
        "       \n",
        "        'MIN_TEMPERATURE_C_t','clients_day_t', 'Holiday_Food_Related_t',\n",
        "        'weekend_t', 'max_pdl_daily_load_t']].dropna()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "59aKcM4mdUQU"
      },
      "source": [
        "the add of weekend feature makes the seasonal weekend decrease more visible, so increases performance"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "gdnkSpWaENYQ"
      },
      "source": [
        "scaler = MinMaxScaler(feature_range=(0, 1))\n",
        "data_scaled = scaler.fit_transform(df_input)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "rCY48qWVESHY",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "ebb9c5de-157c-4117-ecd7-f0b88925664d"
      },
      "source": [
        "data_scaled[-1]"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([0.73076923, 0.88078035, 0.        , 0.17276642, 0.4053923 ,\n",
              "       0.50081699, 0.        , 0.73076923, 0.88060781, 0.        ,\n",
              "       0.17276642, 0.3398278 , 0.29248366, 0.        , 0.76923077,\n",
              "       0.88060781, 1.        , 0.        , 0.10130719])"
            ]
          },
          "metadata": {},
          "execution_count": 13
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "IEZuMY99zfLU"
      },
      "source": [
        "features=data_scaled[:,:-1]\n",
        "target=data_scaled[:,-1]"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "6x_iIIrkomf7",
        "outputId": "b48a2bfc-dfb3-4fe8-c856-dc54c595aa95"
      },
      "source": [
        "features.shape"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(2032, 18)"
            ]
          },
          "metadata": {},
          "execution_count": 15
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "fpuzlDtTotWA",
        "outputId": "3638e6ff-bad4-4bbe-e865-30b892db8046"
      },
      "source": [
        "target.shape"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(2032,)"
            ]
          },
          "metadata": {},
          "execution_count": 16
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "nDaUCPH6njFx",
        "outputId": "2523f597-e318-4d9a-a86c-5b75bb518b7a"
      },
      "source": [
        "# split into train and test sets\n",
        "\n",
        "n_train_days = -30*7\n",
        "\n",
        "train_X,train_y = features[:n_train_days, :],target[:n_train_days]\n",
        "test_X, test_y = features[n_train_days:, :],target[n_train_days:]\n",
        "\n",
        "# reshape input to be 3D\n",
        "train_X = train_X.reshape((train_X.shape[0], train_X.shape[1],1))\n",
        "test_X = test_X.reshape((test_X.shape[0],test_X.shape[1],1))\n",
        "print(train_X.shape, train_y.shape, test_X.shape, test_y.shape)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(1822, 18, 1) (1822,) (210, 18, 1) (210,)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "kZthsjPVnjHh",
        "outputId": "574c7630-d2c7-437c-8da4-d45e85763086"
      },
      "source": [
        "def create_model(layers):\n",
        "    model = Sequential()\n",
        "    for i, nodes in enumerate(layers):\n",
        "        if i==0:\n",
        "            model.add(Bidirectional(LSTM(units=nodes,input_shape= (train_X.shape[1],1),\n",
        "                                         kernel_constraint=maxnorm(3.5))))\n",
        "            model.add(Dropout(0.3))\n",
        "        else:\n",
        "            model.add(Bidirectional(LSTM(units=nodes, kernel_constraint=maxnorm(3.5))))\n",
        "            model.add(Dropout(0.3))\n",
        "    model.add(Dense(1)) \n",
        "    \n",
        "    model.compile(loss='mse', optimizer='adam')\n",
        "    return model\n",
        "\n",
        "model = KerasClassifier(build_fn=create_model, verbose=0) #build_fn is the name of the \n",
        "#function to call to get your model.   \n",
        "tscv = TimeSeriesSplit(n_splits=4)\n",
        "layers = [[100], [100, 50], [100, 50, 25]]\n",
        "batches=[16,32]\n",
        "epochs=[100,200]\n",
        "param_grid = dict(layers=layers, batch_size = batches , epochs=epochs)\n",
        "gsearch = GridSearchCV(estimator=model, cv=tscv, param_grid=param_grid,scoring='r2')\n",
        "grid_result=gsearch.fit(train_X, train_y)\n",
        "print(\"Best: %f using %s\" % (grid_result.best_score_, grid_result.best_params_))\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/keras/engine/sequential.py:450: UserWarning: `model.predict_classes()` is deprecated and will be removed after 2021-01-01. Please use instead:* `np.argmax(model.predict(x), axis=-1)`,   if your model does multi-class classification   (e.g. if it uses a `softmax` last-layer activation).* `(model.predict(x) > 0.5).astype(\"int32\")`,   if your model does binary classification   (e.g. if it uses a `sigmoid` last-layer activation).\n",
            "  warnings.warn('`model.predict_classes()` is deprecated and '\n",
            "/usr/local/lib/python3.7/dist-packages/keras/engine/sequential.py:450: UserWarning: `model.predict_classes()` is deprecated and will be removed after 2021-01-01. Please use instead:* `np.argmax(model.predict(x), axis=-1)`,   if your model does multi-class classification   (e.g. if it uses a `softmax` last-layer activation).* `(model.predict(x) > 0.5).astype(\"int32\")`,   if your model does binary classification   (e.g. if it uses a `sigmoid` last-layer activation).\n",
            "  warnings.warn('`model.predict_classes()` is deprecated and '\n",
            "/usr/local/lib/python3.7/dist-packages/keras/engine/sequential.py:450: UserWarning: `model.predict_classes()` is deprecated and will be removed after 2021-01-01. Please use instead:* `np.argmax(model.predict(x), axis=-1)`,   if your model does multi-class classification   (e.g. if it uses a `softmax` last-layer activation).* `(model.predict(x) > 0.5).astype(\"int32\")`,   if your model does binary classification   (e.g. if it uses a `sigmoid` last-layer activation).\n",
            "  warnings.warn('`model.predict_classes()` is deprecated and '\n",
            "/usr/local/lib/python3.7/dist-packages/keras/engine/sequential.py:450: UserWarning: `model.predict_classes()` is deprecated and will be removed after 2021-01-01. Please use instead:* `np.argmax(model.predict(x), axis=-1)`,   if your model does multi-class classification   (e.g. if it uses a `softmax` last-layer activation).* `(model.predict(x) > 0.5).astype(\"int32\")`,   if your model does binary classification   (e.g. if it uses a `sigmoid` last-layer activation).\n",
            "  warnings.warn('`model.predict_classes()` is deprecated and '\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/model_selection/_validation.py:536: FitFailedWarning: Estimator fit failed. The score on this train-test partition for these parameters will be set to nan. Details: \n",
            "ValueError: in user code:\n",
            "\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:830 train_function  *\n",
            "        return step_function(self, iterator)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:813 run_step  *\n",
            "        outputs = model.train_step(data)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:770 train_step  *\n",
            "        y_pred = self(x, training=True)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/base_layer.py:989 __call__  *\n",
            "        input_spec.assert_input_compatibility(self.input_spec, inputs, self.name)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/sequential.py:389 call  *\n",
            "        outputs = layer(inputs, **kwargs)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/layers/wrappers.py:581 __call__  *\n",
            "        return super(Bidirectional, self).__call__(inputs, **kwargs)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/base_layer.py:989 __call__  *\n",
            "        input_spec.assert_input_compatibility(self.input_spec, inputs, self.name)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/input_spec.py:212 assert_input_compatibility  *\n",
            "        raise ValueError('Input ' + str(input_index) + ' of layer ' +\n",
            "\n",
            "    ValueError: Input 0 of layer bidirectional_5 is incompatible with the layer: expected ndim=3, found ndim=2. Full shape received: (None, 200)\n",
            "\n",
            "\n",
            "  FitFailedWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/model_selection/_validation.py:536: FitFailedWarning: Estimator fit failed. The score on this train-test partition for these parameters will be set to nan. Details: \n",
            "ValueError: in user code:\n",
            "\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:830 train_function  *\n",
            "        return step_function(self, iterator)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:813 run_step  *\n",
            "        outputs = model.train_step(data)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:770 train_step  *\n",
            "        y_pred = self(x, training=True)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/base_layer.py:989 __call__  *\n",
            "        input_spec.assert_input_compatibility(self.input_spec, inputs, self.name)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/sequential.py:389 call  *\n",
            "        outputs = layer(inputs, **kwargs)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/layers/wrappers.py:581 __call__  *\n",
            "        return super(Bidirectional, self).__call__(inputs, **kwargs)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/base_layer.py:989 __call__  *\n",
            "        input_spec.assert_input_compatibility(self.input_spec, inputs, self.name)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/input_spec.py:212 assert_input_compatibility  *\n",
            "        raise ValueError('Input ' + str(input_index) + ' of layer ' +\n",
            "\n",
            "    ValueError: Input 0 of layer bidirectional_7 is incompatible with the layer: expected ndim=3, found ndim=2. Full shape received: (None, 200)\n",
            "\n",
            "\n",
            "  FitFailedWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/model_selection/_validation.py:536: FitFailedWarning: Estimator fit failed. The score on this train-test partition for these parameters will be set to nan. Details: \n",
            "ValueError: in user code:\n",
            "\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:830 train_function  *\n",
            "        return step_function(self, iterator)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:813 run_step  *\n",
            "        outputs = model.train_step(data)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:770 train_step  *\n",
            "        y_pred = self(x, training=True)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/base_layer.py:989 __call__  *\n",
            "        input_spec.assert_input_compatibility(self.input_spec, inputs, self.name)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/sequential.py:389 call  *\n",
            "        outputs = layer(inputs, **kwargs)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/layers/wrappers.py:581 __call__  *\n",
            "        return super(Bidirectional, self).__call__(inputs, **kwargs)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/base_layer.py:989 __call__  *\n",
            "        input_spec.assert_input_compatibility(self.input_spec, inputs, self.name)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/input_spec.py:212 assert_input_compatibility  *\n",
            "        raise ValueError('Input ' + str(input_index) + ' of layer ' +\n",
            "\n",
            "    ValueError: Input 0 of layer bidirectional_9 is incompatible with the layer: expected ndim=3, found ndim=2. Full shape received: (None, 200)\n",
            "\n",
            "\n",
            "  FitFailedWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/model_selection/_validation.py:536: FitFailedWarning: Estimator fit failed. The score on this train-test partition for these parameters will be set to nan. Details: \n",
            "ValueError: in user code:\n",
            "\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:830 train_function  *\n",
            "        return step_function(self, iterator)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:813 run_step  *\n",
            "        outputs = model.train_step(data)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:770 train_step  *\n",
            "        y_pred = self(x, training=True)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/base_layer.py:989 __call__  *\n",
            "        input_spec.assert_input_compatibility(self.input_spec, inputs, self.name)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/sequential.py:389 call  *\n",
            "        outputs = layer(inputs, **kwargs)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/layers/wrappers.py:581 __call__  *\n",
            "        return super(Bidirectional, self).__call__(inputs, **kwargs)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/base_layer.py:989 __call__  *\n",
            "        input_spec.assert_input_compatibility(self.input_spec, inputs, self.name)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/input_spec.py:212 assert_input_compatibility  *\n",
            "        raise ValueError('Input ' + str(input_index) + ' of layer ' +\n",
            "\n",
            "    ValueError: Input 0 of layer bidirectional_11 is incompatible with the layer: expected ndim=3, found ndim=2. Full shape received: (None, 200)\n",
            "\n",
            "\n",
            "  FitFailedWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/model_selection/_validation.py:536: FitFailedWarning: Estimator fit failed. The score on this train-test partition for these parameters will be set to nan. Details: \n",
            "ValueError: in user code:\n",
            "\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:830 train_function  *\n",
            "        return step_function(self, iterator)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:813 run_step  *\n",
            "        outputs = model.train_step(data)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:770 train_step  *\n",
            "        y_pred = self(x, training=True)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/base_layer.py:989 __call__  *\n",
            "        input_spec.assert_input_compatibility(self.input_spec, inputs, self.name)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/sequential.py:389 call  *\n",
            "        outputs = layer(inputs, **kwargs)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/layers/wrappers.py:581 __call__  *\n",
            "        return super(Bidirectional, self).__call__(inputs, **kwargs)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/base_layer.py:989 __call__  *\n",
            "        input_spec.assert_input_compatibility(self.input_spec, inputs, self.name)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/input_spec.py:212 assert_input_compatibility  *\n",
            "        raise ValueError('Input ' + str(input_index) + ' of layer ' +\n",
            "\n",
            "    ValueError: Input 0 of layer bidirectional_13 is incompatible with the layer: expected ndim=3, found ndim=2. Full shape received: (None, 200)\n",
            "\n",
            "\n",
            "  FitFailedWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/model_selection/_validation.py:536: FitFailedWarning: Estimator fit failed. The score on this train-test partition for these parameters will be set to nan. Details: \n",
            "ValueError: in user code:\n",
            "\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:830 train_function  *\n",
            "        return step_function(self, iterator)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:813 run_step  *\n",
            "        outputs = model.train_step(data)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:770 train_step  *\n",
            "        y_pred = self(x, training=True)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/base_layer.py:989 __call__  *\n",
            "        input_spec.assert_input_compatibility(self.input_spec, inputs, self.name)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/sequential.py:389 call  *\n",
            "        outputs = layer(inputs, **kwargs)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/layers/wrappers.py:581 __call__  *\n",
            "        return super(Bidirectional, self).__call__(inputs, **kwargs)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/base_layer.py:989 __call__  *\n",
            "        input_spec.assert_input_compatibility(self.input_spec, inputs, self.name)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/input_spec.py:212 assert_input_compatibility  *\n",
            "        raise ValueError('Input ' + str(input_index) + ' of layer ' +\n",
            "\n",
            "    ValueError: Input 0 of layer bidirectional_16 is incompatible with the layer: expected ndim=3, found ndim=2. Full shape received: (None, 200)\n",
            "\n",
            "\n",
            "  FitFailedWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/model_selection/_validation.py:536: FitFailedWarning: Estimator fit failed. The score on this train-test partition for these parameters will be set to nan. Details: \n",
            "ValueError: in user code:\n",
            "\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:830 train_function  *\n",
            "        return step_function(self, iterator)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:813 run_step  *\n",
            "        outputs = model.train_step(data)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:770 train_step  *\n",
            "        y_pred = self(x, training=True)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/base_layer.py:989 __call__  *\n",
            "        input_spec.assert_input_compatibility(self.input_spec, inputs, self.name)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/sequential.py:389 call  *\n",
            "        outputs = layer(inputs, **kwargs)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/layers/wrappers.py:581 __call__  *\n",
            "        return super(Bidirectional, self).__call__(inputs, **kwargs)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/base_layer.py:989 __call__  *\n",
            "        input_spec.assert_input_compatibility(self.input_spec, inputs, self.name)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/input_spec.py:212 assert_input_compatibility  *\n",
            "        raise ValueError('Input ' + str(input_index) + ' of layer ' +\n",
            "\n",
            "    ValueError: Input 0 of layer bidirectional_19 is incompatible with the layer: expected ndim=3, found ndim=2. Full shape received: (None, 200)\n",
            "\n",
            "\n",
            "  FitFailedWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/model_selection/_validation.py:536: FitFailedWarning: Estimator fit failed. The score on this train-test partition for these parameters will be set to nan. Details: \n",
            "ValueError: in user code:\n",
            "\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:830 train_function  *\n",
            "        return step_function(self, iterator)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:813 run_step  *\n",
            "        outputs = model.train_step(data)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:770 train_step  *\n",
            "        y_pred = self(x, training=True)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/base_layer.py:989 __call__  *\n",
            "        input_spec.assert_input_compatibility(self.input_spec, inputs, self.name)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/sequential.py:389 call  *\n",
            "        outputs = layer(inputs, **kwargs)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/layers/wrappers.py:581 __call__  *\n",
            "        return super(Bidirectional, self).__call__(inputs, **kwargs)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/base_layer.py:989 __call__  *\n",
            "        input_spec.assert_input_compatibility(self.input_spec, inputs, self.name)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/input_spec.py:212 assert_input_compatibility  *\n",
            "        raise ValueError('Input ' + str(input_index) + ' of layer ' +\n",
            "\n",
            "    ValueError: Input 0 of layer bidirectional_22 is incompatible with the layer: expected ndim=3, found ndim=2. Full shape received: (None, 200)\n",
            "\n",
            "\n",
            "  FitFailedWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/keras/engine/sequential.py:450: UserWarning: `model.predict_classes()` is deprecated and will be removed after 2021-01-01. Please use instead:* `np.argmax(model.predict(x), axis=-1)`,   if your model does multi-class classification   (e.g. if it uses a `softmax` last-layer activation).* `(model.predict(x) > 0.5).astype(\"int32\")`,   if your model does binary classification   (e.g. if it uses a `sigmoid` last-layer activation).\n",
            "  warnings.warn('`model.predict_classes()` is deprecated and '\n",
            "/usr/local/lib/python3.7/dist-packages/keras/engine/sequential.py:450: UserWarning: `model.predict_classes()` is deprecated and will be removed after 2021-01-01. Please use instead:* `np.argmax(model.predict(x), axis=-1)`,   if your model does multi-class classification   (e.g. if it uses a `softmax` last-layer activation).* `(model.predict(x) > 0.5).astype(\"int32\")`,   if your model does binary classification   (e.g. if it uses a `sigmoid` last-layer activation).\n",
            "  warnings.warn('`model.predict_classes()` is deprecated and '\n",
            "/usr/local/lib/python3.7/dist-packages/keras/engine/sequential.py:450: UserWarning: `model.predict_classes()` is deprecated and will be removed after 2021-01-01. Please use instead:* `np.argmax(model.predict(x), axis=-1)`,   if your model does multi-class classification   (e.g. if it uses a `softmax` last-layer activation).* `(model.predict(x) > 0.5).astype(\"int32\")`,   if your model does binary classification   (e.g. if it uses a `sigmoid` last-layer activation).\n",
            "  warnings.warn('`model.predict_classes()` is deprecated and '\n",
            "/usr/local/lib/python3.7/dist-packages/keras/engine/sequential.py:450: UserWarning: `model.predict_classes()` is deprecated and will be removed after 2021-01-01. Please use instead:* `np.argmax(model.predict(x), axis=-1)`,   if your model does multi-class classification   (e.g. if it uses a `softmax` last-layer activation).* `(model.predict(x) > 0.5).astype(\"int32\")`,   if your model does binary classification   (e.g. if it uses a `sigmoid` last-layer activation).\n",
            "  warnings.warn('`model.predict_classes()` is deprecated and '\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/model_selection/_validation.py:536: FitFailedWarning: Estimator fit failed. The score on this train-test partition for these parameters will be set to nan. Details: \n",
            "ValueError: in user code:\n",
            "\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:830 train_function  *\n",
            "        return step_function(self, iterator)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:813 run_step  *\n",
            "        outputs = model.train_step(data)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:770 train_step  *\n",
            "        y_pred = self(x, training=True)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/base_layer.py:989 __call__  *\n",
            "        input_spec.assert_input_compatibility(self.input_spec, inputs, self.name)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/sequential.py:389 call  *\n",
            "        outputs = layer(inputs, **kwargs)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/layers/wrappers.py:581 __call__  *\n",
            "        return super(Bidirectional, self).__call__(inputs, **kwargs)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/base_layer.py:989 __call__  *\n",
            "        input_spec.assert_input_compatibility(self.input_spec, inputs, self.name)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/input_spec.py:212 assert_input_compatibility  *\n",
            "        raise ValueError('Input ' + str(input_index) + ' of layer ' +\n",
            "\n",
            "    ValueError: Input 0 of layer bidirectional_29 is incompatible with the layer: expected ndim=3, found ndim=2. Full shape received: (None, 200)\n",
            "\n",
            "\n",
            "  FitFailedWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/model_selection/_validation.py:536: FitFailedWarning: Estimator fit failed. The score on this train-test partition for these parameters will be set to nan. Details: \n",
            "ValueError: in user code:\n",
            "\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:830 train_function  *\n",
            "        return step_function(self, iterator)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:813 run_step  *\n",
            "        outputs = model.train_step(data)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:770 train_step  *\n",
            "        y_pred = self(x, training=True)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/base_layer.py:989 __call__  *\n",
            "        input_spec.assert_input_compatibility(self.input_spec, inputs, self.name)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/sequential.py:389 call  *\n",
            "        outputs = layer(inputs, **kwargs)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/layers/wrappers.py:581 __call__  *\n",
            "        return super(Bidirectional, self).__call__(inputs, **kwargs)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/base_layer.py:989 __call__  *\n",
            "        input_spec.assert_input_compatibility(self.input_spec, inputs, self.name)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/input_spec.py:212 assert_input_compatibility  *\n",
            "        raise ValueError('Input ' + str(input_index) + ' of layer ' +\n",
            "\n",
            "    ValueError: Input 0 of layer bidirectional_31 is incompatible with the layer: expected ndim=3, found ndim=2. Full shape received: (None, 200)\n",
            "\n",
            "\n",
            "  FitFailedWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/model_selection/_validation.py:536: FitFailedWarning: Estimator fit failed. The score on this train-test partition for these parameters will be set to nan. Details: \n",
            "ValueError: in user code:\n",
            "\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:830 train_function  *\n",
            "        return step_function(self, iterator)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:813 run_step  *\n",
            "        outputs = model.train_step(data)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:770 train_step  *\n",
            "        y_pred = self(x, training=True)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/base_layer.py:989 __call__  *\n",
            "        input_spec.assert_input_compatibility(self.input_spec, inputs, self.name)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/sequential.py:389 call  *\n",
            "        outputs = layer(inputs, **kwargs)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/layers/wrappers.py:581 __call__  *\n",
            "        return super(Bidirectional, self).__call__(inputs, **kwargs)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/base_layer.py:989 __call__  *\n",
            "        input_spec.assert_input_compatibility(self.input_spec, inputs, self.name)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/input_spec.py:212 assert_input_compatibility  *\n",
            "        raise ValueError('Input ' + str(input_index) + ' of layer ' +\n",
            "\n",
            "    ValueError: Input 0 of layer bidirectional_33 is incompatible with the layer: expected ndim=3, found ndim=2. Full shape received: (None, 200)\n",
            "\n",
            "\n",
            "  FitFailedWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/model_selection/_validation.py:536: FitFailedWarning: Estimator fit failed. The score on this train-test partition for these parameters will be set to nan. Details: \n",
            "ValueError: in user code:\n",
            "\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:830 train_function  *\n",
            "        return step_function(self, iterator)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:813 run_step  *\n",
            "        outputs = model.train_step(data)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:770 train_step  *\n",
            "        y_pred = self(x, training=True)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/base_layer.py:989 __call__  *\n",
            "        input_spec.assert_input_compatibility(self.input_spec, inputs, self.name)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/sequential.py:389 call  *\n",
            "        outputs = layer(inputs, **kwargs)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/layers/wrappers.py:581 __call__  *\n",
            "        return super(Bidirectional, self).__call__(inputs, **kwargs)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/base_layer.py:989 __call__  *\n",
            "        input_spec.assert_input_compatibility(self.input_spec, inputs, self.name)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/input_spec.py:212 assert_input_compatibility  *\n",
            "        raise ValueError('Input ' + str(input_index) + ' of layer ' +\n",
            "\n",
            "    ValueError: Input 0 of layer bidirectional_35 is incompatible with the layer: expected ndim=3, found ndim=2. Full shape received: (None, 200)\n",
            "\n",
            "\n",
            "  FitFailedWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/model_selection/_validation.py:536: FitFailedWarning: Estimator fit failed. The score on this train-test partition for these parameters will be set to nan. Details: \n",
            "ValueError: in user code:\n",
            "\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:830 train_function  *\n",
            "        return step_function(self, iterator)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:813 run_step  *\n",
            "        outputs = model.train_step(data)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:770 train_step  *\n",
            "        y_pred = self(x, training=True)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/base_layer.py:989 __call__  *\n",
            "        input_spec.assert_input_compatibility(self.input_spec, inputs, self.name)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/sequential.py:389 call  *\n",
            "        outputs = layer(inputs, **kwargs)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/layers/wrappers.py:581 __call__  *\n",
            "        return super(Bidirectional, self).__call__(inputs, **kwargs)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/base_layer.py:989 __call__  *\n",
            "        input_spec.assert_input_compatibility(self.input_spec, inputs, self.name)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/input_spec.py:212 assert_input_compatibility  *\n",
            "        raise ValueError('Input ' + str(input_index) + ' of layer ' +\n",
            "\n",
            "    ValueError: Input 0 of layer bidirectional_37 is incompatible with the layer: expected ndim=3, found ndim=2. Full shape received: (None, 200)\n",
            "\n",
            "\n",
            "  FitFailedWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/model_selection/_validation.py:536: FitFailedWarning: Estimator fit failed. The score on this train-test partition for these parameters will be set to nan. Details: \n",
            "ValueError: in user code:\n",
            "\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:830 train_function  *\n",
            "        return step_function(self, iterator)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:813 run_step  *\n",
            "        outputs = model.train_step(data)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:770 train_step  *\n",
            "        y_pred = self(x, training=True)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/base_layer.py:989 __call__  *\n",
            "        input_spec.assert_input_compatibility(self.input_spec, inputs, self.name)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/sequential.py:389 call  *\n",
            "        outputs = layer(inputs, **kwargs)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/layers/wrappers.py:581 __call__  *\n",
            "        return super(Bidirectional, self).__call__(inputs, **kwargs)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/base_layer.py:989 __call__  *\n",
            "        input_spec.assert_input_compatibility(self.input_spec, inputs, self.name)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/input_spec.py:212 assert_input_compatibility  *\n",
            "        raise ValueError('Input ' + str(input_index) + ' of layer ' +\n",
            "\n",
            "    ValueError: Input 0 of layer bidirectional_40 is incompatible with the layer: expected ndim=3, found ndim=2. Full shape received: (None, 200)\n",
            "\n",
            "\n",
            "  FitFailedWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/model_selection/_validation.py:536: FitFailedWarning: Estimator fit failed. The score on this train-test partition for these parameters will be set to nan. Details: \n",
            "ValueError: in user code:\n",
            "\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:830 train_function  *\n",
            "        return step_function(self, iterator)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:813 run_step  *\n",
            "        outputs = model.train_step(data)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:770 train_step  *\n",
            "        y_pred = self(x, training=True)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/base_layer.py:989 __call__  *\n",
            "        input_spec.assert_input_compatibility(self.input_spec, inputs, self.name)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/sequential.py:389 call  *\n",
            "        outputs = layer(inputs, **kwargs)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/layers/wrappers.py:581 __call__  *\n",
            "        return super(Bidirectional, self).__call__(inputs, **kwargs)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/base_layer.py:989 __call__  *\n",
            "        input_spec.assert_input_compatibility(self.input_spec, inputs, self.name)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/input_spec.py:212 assert_input_compatibility  *\n",
            "        raise ValueError('Input ' + str(input_index) + ' of layer ' +\n",
            "\n",
            "    ValueError: Input 0 of layer bidirectional_43 is incompatible with the layer: expected ndim=3, found ndim=2. Full shape received: (None, 200)\n",
            "\n",
            "\n",
            "  FitFailedWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/model_selection/_validation.py:536: FitFailedWarning: Estimator fit failed. The score on this train-test partition for these parameters will be set to nan. Details: \n",
            "ValueError: in user code:\n",
            "\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:830 train_function  *\n",
            "        return step_function(self, iterator)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:813 run_step  *\n",
            "        outputs = model.train_step(data)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:770 train_step  *\n",
            "        y_pred = self(x, training=True)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/base_layer.py:989 __call__  *\n",
            "        input_spec.assert_input_compatibility(self.input_spec, inputs, self.name)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/sequential.py:389 call  *\n",
            "        outputs = layer(inputs, **kwargs)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/layers/wrappers.py:581 __call__  *\n",
            "        return super(Bidirectional, self).__call__(inputs, **kwargs)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/base_layer.py:989 __call__  *\n",
            "        input_spec.assert_input_compatibility(self.input_spec, inputs, self.name)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/input_spec.py:212 assert_input_compatibility  *\n",
            "        raise ValueError('Input ' + str(input_index) + ' of layer ' +\n",
            "\n",
            "    ValueError: Input 0 of layer bidirectional_46 is incompatible with the layer: expected ndim=3, found ndim=2. Full shape received: (None, 200)\n",
            "\n",
            "\n",
            "  FitFailedWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/keras/engine/sequential.py:450: UserWarning: `model.predict_classes()` is deprecated and will be removed after 2021-01-01. Please use instead:* `np.argmax(model.predict(x), axis=-1)`,   if your model does multi-class classification   (e.g. if it uses a `softmax` last-layer activation).* `(model.predict(x) > 0.5).astype(\"int32\")`,   if your model does binary classification   (e.g. if it uses a `sigmoid` last-layer activation).\n",
            "  warnings.warn('`model.predict_classes()` is deprecated and '\n",
            "/usr/local/lib/python3.7/dist-packages/keras/engine/sequential.py:450: UserWarning: `model.predict_classes()` is deprecated and will be removed after 2021-01-01. Please use instead:* `np.argmax(model.predict(x), axis=-1)`,   if your model does multi-class classification   (e.g. if it uses a `softmax` last-layer activation).* `(model.predict(x) > 0.5).astype(\"int32\")`,   if your model does binary classification   (e.g. if it uses a `sigmoid` last-layer activation).\n",
            "  warnings.warn('`model.predict_classes()` is deprecated and '\n",
            "/usr/local/lib/python3.7/dist-packages/keras/engine/sequential.py:450: UserWarning: `model.predict_classes()` is deprecated and will be removed after 2021-01-01. Please use instead:* `np.argmax(model.predict(x), axis=-1)`,   if your model does multi-class classification   (e.g. if it uses a `softmax` last-layer activation).* `(model.predict(x) > 0.5).astype(\"int32\")`,   if your model does binary classification   (e.g. if it uses a `sigmoid` last-layer activation).\n",
            "  warnings.warn('`model.predict_classes()` is deprecated and '\n",
            "/usr/local/lib/python3.7/dist-packages/keras/engine/sequential.py:450: UserWarning: `model.predict_classes()` is deprecated and will be removed after 2021-01-01. Please use instead:* `np.argmax(model.predict(x), axis=-1)`,   if your model does multi-class classification   (e.g. if it uses a `softmax` last-layer activation).* `(model.predict(x) > 0.5).astype(\"int32\")`,   if your model does binary classification   (e.g. if it uses a `sigmoid` last-layer activation).\n",
            "  warnings.warn('`model.predict_classes()` is deprecated and '\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/model_selection/_validation.py:536: FitFailedWarning: Estimator fit failed. The score on this train-test partition for these parameters will be set to nan. Details: \n",
            "ValueError: in user code:\n",
            "\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:830 train_function  *\n",
            "        return step_function(self, iterator)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:813 run_step  *\n",
            "        outputs = model.train_step(data)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:770 train_step  *\n",
            "        y_pred = self(x, training=True)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/base_layer.py:989 __call__  *\n",
            "        input_spec.assert_input_compatibility(self.input_spec, inputs, self.name)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/sequential.py:389 call  *\n",
            "        outputs = layer(inputs, **kwargs)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/layers/wrappers.py:581 __call__  *\n",
            "        return super(Bidirectional, self).__call__(inputs, **kwargs)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/base_layer.py:989 __call__  *\n",
            "        input_spec.assert_input_compatibility(self.input_spec, inputs, self.name)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/input_spec.py:212 assert_input_compatibility  *\n",
            "        raise ValueError('Input ' + str(input_index) + ' of layer ' +\n",
            "\n",
            "    ValueError: Input 0 of layer bidirectional_53 is incompatible with the layer: expected ndim=3, found ndim=2. Full shape received: (None, 200)\n",
            "\n",
            "\n",
            "  FitFailedWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/model_selection/_validation.py:536: FitFailedWarning: Estimator fit failed. The score on this train-test partition for these parameters will be set to nan. Details: \n",
            "ValueError: in user code:\n",
            "\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:830 train_function  *\n",
            "        return step_function(self, iterator)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:813 run_step  *\n",
            "        outputs = model.train_step(data)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:770 train_step  *\n",
            "        y_pred = self(x, training=True)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/base_layer.py:989 __call__  *\n",
            "        input_spec.assert_input_compatibility(self.input_spec, inputs, self.name)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/sequential.py:389 call  *\n",
            "        outputs = layer(inputs, **kwargs)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/layers/wrappers.py:581 __call__  *\n",
            "        return super(Bidirectional, self).__call__(inputs, **kwargs)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/base_layer.py:989 __call__  *\n",
            "        input_spec.assert_input_compatibility(self.input_spec, inputs, self.name)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/input_spec.py:212 assert_input_compatibility  *\n",
            "        raise ValueError('Input ' + str(input_index) + ' of layer ' +\n",
            "\n",
            "    ValueError: Input 0 of layer bidirectional_55 is incompatible with the layer: expected ndim=3, found ndim=2. Full shape received: (None, 200)\n",
            "\n",
            "\n",
            "  FitFailedWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/model_selection/_validation.py:536: FitFailedWarning: Estimator fit failed. The score on this train-test partition for these parameters will be set to nan. Details: \n",
            "ValueError: in user code:\n",
            "\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:830 train_function  *\n",
            "        return step_function(self, iterator)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:813 run_step  *\n",
            "        outputs = model.train_step(data)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:770 train_step  *\n",
            "        y_pred = self(x, training=True)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/base_layer.py:989 __call__  *\n",
            "        input_spec.assert_input_compatibility(self.input_spec, inputs, self.name)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/sequential.py:389 call  *\n",
            "        outputs = layer(inputs, **kwargs)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/layers/wrappers.py:581 __call__  *\n",
            "        return super(Bidirectional, self).__call__(inputs, **kwargs)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/base_layer.py:989 __call__  *\n",
            "        input_spec.assert_input_compatibility(self.input_spec, inputs, self.name)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/input_spec.py:212 assert_input_compatibility  *\n",
            "        raise ValueError('Input ' + str(input_index) + ' of layer ' +\n",
            "\n",
            "    ValueError: Input 0 of layer bidirectional_57 is incompatible with the layer: expected ndim=3, found ndim=2. Full shape received: (None, 200)\n",
            "\n",
            "\n",
            "  FitFailedWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/model_selection/_validation.py:536: FitFailedWarning: Estimator fit failed. The score on this train-test partition for these parameters will be set to nan. Details: \n",
            "ValueError: in user code:\n",
            "\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:830 train_function  *\n",
            "        return step_function(self, iterator)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:813 run_step  *\n",
            "        outputs = model.train_step(data)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:770 train_step  *\n",
            "        y_pred = self(x, training=True)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/base_layer.py:989 __call__  *\n",
            "        input_spec.assert_input_compatibility(self.input_spec, inputs, self.name)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/sequential.py:389 call  *\n",
            "        outputs = layer(inputs, **kwargs)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/layers/wrappers.py:581 __call__  *\n",
            "        return super(Bidirectional, self).__call__(inputs, **kwargs)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/base_layer.py:989 __call__  *\n",
            "        input_spec.assert_input_compatibility(self.input_spec, inputs, self.name)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/input_spec.py:212 assert_input_compatibility  *\n",
            "        raise ValueError('Input ' + str(input_index) + ' of layer ' +\n",
            "\n",
            "    ValueError: Input 0 of layer bidirectional_59 is incompatible with the layer: expected ndim=3, found ndim=2. Full shape received: (None, 200)\n",
            "\n",
            "\n",
            "  FitFailedWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/model_selection/_validation.py:536: FitFailedWarning: Estimator fit failed. The score on this train-test partition for these parameters will be set to nan. Details: \n",
            "ValueError: in user code:\n",
            "\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:830 train_function  *\n",
            "        return step_function(self, iterator)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:813 run_step  *\n",
            "        outputs = model.train_step(data)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:770 train_step  *\n",
            "        y_pred = self(x, training=True)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/base_layer.py:989 __call__  *\n",
            "        input_spec.assert_input_compatibility(self.input_spec, inputs, self.name)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/sequential.py:389 call  *\n",
            "        outputs = layer(inputs, **kwargs)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/layers/wrappers.py:581 __call__  *\n",
            "        return super(Bidirectional, self).__call__(inputs, **kwargs)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/base_layer.py:989 __call__  *\n",
            "        input_spec.assert_input_compatibility(self.input_spec, inputs, self.name)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/input_spec.py:212 assert_input_compatibility  *\n",
            "        raise ValueError('Input ' + str(input_index) + ' of layer ' +\n",
            "\n",
            "    ValueError: Input 0 of layer bidirectional_61 is incompatible with the layer: expected ndim=3, found ndim=2. Full shape received: (None, 200)\n",
            "\n",
            "\n",
            "  FitFailedWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/model_selection/_validation.py:536: FitFailedWarning: Estimator fit failed. The score on this train-test partition for these parameters will be set to nan. Details: \n",
            "ValueError: in user code:\n",
            "\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:830 train_function  *\n",
            "        return step_function(self, iterator)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:813 run_step  *\n",
            "        outputs = model.train_step(data)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:770 train_step  *\n",
            "        y_pred = self(x, training=True)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/base_layer.py:989 __call__  *\n",
            "        input_spec.assert_input_compatibility(self.input_spec, inputs, self.name)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/sequential.py:389 call  *\n",
            "        outputs = layer(inputs, **kwargs)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/layers/wrappers.py:581 __call__  *\n",
            "        return super(Bidirectional, self).__call__(inputs, **kwargs)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/base_layer.py:989 __call__  *\n",
            "        input_spec.assert_input_compatibility(self.input_spec, inputs, self.name)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/input_spec.py:212 assert_input_compatibility  *\n",
            "        raise ValueError('Input ' + str(input_index) + ' of layer ' +\n",
            "\n",
            "    ValueError: Input 0 of layer bidirectional_64 is incompatible with the layer: expected ndim=3, found ndim=2. Full shape received: (None, 200)\n",
            "\n",
            "\n",
            "  FitFailedWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/model_selection/_validation.py:536: FitFailedWarning: Estimator fit failed. The score on this train-test partition for these parameters will be set to nan. Details: \n",
            "ValueError: in user code:\n",
            "\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:830 train_function  *\n",
            "        return step_function(self, iterator)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:813 run_step  *\n",
            "        outputs = model.train_step(data)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:770 train_step  *\n",
            "        y_pred = self(x, training=True)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/base_layer.py:989 __call__  *\n",
            "        input_spec.assert_input_compatibility(self.input_spec, inputs, self.name)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/sequential.py:389 call  *\n",
            "        outputs = layer(inputs, **kwargs)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/layers/wrappers.py:581 __call__  *\n",
            "        return super(Bidirectional, self).__call__(inputs, **kwargs)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/base_layer.py:989 __call__  *\n",
            "        input_spec.assert_input_compatibility(self.input_spec, inputs, self.name)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/input_spec.py:212 assert_input_compatibility  *\n",
            "        raise ValueError('Input ' + str(input_index) + ' of layer ' +\n",
            "\n",
            "    ValueError: Input 0 of layer bidirectional_67 is incompatible with the layer: expected ndim=3, found ndim=2. Full shape received: (None, 200)\n",
            "\n",
            "\n",
            "  FitFailedWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/model_selection/_validation.py:536: FitFailedWarning: Estimator fit failed. The score on this train-test partition for these parameters will be set to nan. Details: \n",
            "ValueError: in user code:\n",
            "\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:830 train_function  *\n",
            "        return step_function(self, iterator)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:813 run_step  *\n",
            "        outputs = model.train_step(data)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:770 train_step  *\n",
            "        y_pred = self(x, training=True)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/base_layer.py:989 __call__  *\n",
            "        input_spec.assert_input_compatibility(self.input_spec, inputs, self.name)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/sequential.py:389 call  *\n",
            "        outputs = layer(inputs, **kwargs)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/layers/wrappers.py:581 __call__  *\n",
            "        return super(Bidirectional, self).__call__(inputs, **kwargs)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/base_layer.py:989 __call__  *\n",
            "        input_spec.assert_input_compatibility(self.input_spec, inputs, self.name)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/input_spec.py:212 assert_input_compatibility  *\n",
            "        raise ValueError('Input ' + str(input_index) + ' of layer ' +\n",
            "\n",
            "    ValueError: Input 0 of layer bidirectional_70 is incompatible with the layer: expected ndim=3, found ndim=2. Full shape received: (None, 200)\n",
            "\n",
            "\n",
            "  FitFailedWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/keras/engine/sequential.py:450: UserWarning: `model.predict_classes()` is deprecated and will be removed after 2021-01-01. Please use instead:* `np.argmax(model.predict(x), axis=-1)`,   if your model does multi-class classification   (e.g. if it uses a `softmax` last-layer activation).* `(model.predict(x) > 0.5).astype(\"int32\")`,   if your model does binary classification   (e.g. if it uses a `sigmoid` last-layer activation).\n",
            "  warnings.warn('`model.predict_classes()` is deprecated and '\n",
            "/usr/local/lib/python3.7/dist-packages/keras/engine/sequential.py:450: UserWarning: `model.predict_classes()` is deprecated and will be removed after 2021-01-01. Please use instead:* `np.argmax(model.predict(x), axis=-1)`,   if your model does multi-class classification   (e.g. if it uses a `softmax` last-layer activation).* `(model.predict(x) > 0.5).astype(\"int32\")`,   if your model does binary classification   (e.g. if it uses a `sigmoid` last-layer activation).\n",
            "  warnings.warn('`model.predict_classes()` is deprecated and '\n",
            "/usr/local/lib/python3.7/dist-packages/keras/engine/sequential.py:450: UserWarning: `model.predict_classes()` is deprecated and will be removed after 2021-01-01. Please use instead:* `np.argmax(model.predict(x), axis=-1)`,   if your model does multi-class classification   (e.g. if it uses a `softmax` last-layer activation).* `(model.predict(x) > 0.5).astype(\"int32\")`,   if your model does binary classification   (e.g. if it uses a `sigmoid` last-layer activation).\n",
            "  warnings.warn('`model.predict_classes()` is deprecated and '\n",
            "/usr/local/lib/python3.7/dist-packages/keras/engine/sequential.py:450: UserWarning: `model.predict_classes()` is deprecated and will be removed after 2021-01-01. Please use instead:* `np.argmax(model.predict(x), axis=-1)`,   if your model does multi-class classification   (e.g. if it uses a `softmax` last-layer activation).* `(model.predict(x) > 0.5).astype(\"int32\")`,   if your model does binary classification   (e.g. if it uses a `sigmoid` last-layer activation).\n",
            "  warnings.warn('`model.predict_classes()` is deprecated and '\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/model_selection/_validation.py:536: FitFailedWarning: Estimator fit failed. The score on this train-test partition for these parameters will be set to nan. Details: \n",
            "ValueError: in user code:\n",
            "\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:830 train_function  *\n",
            "        return step_function(self, iterator)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:813 run_step  *\n",
            "        outputs = model.train_step(data)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:770 train_step  *\n",
            "        y_pred = self(x, training=True)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/base_layer.py:989 __call__  *\n",
            "        input_spec.assert_input_compatibility(self.input_spec, inputs, self.name)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/sequential.py:389 call  *\n",
            "        outputs = layer(inputs, **kwargs)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/layers/wrappers.py:581 __call__  *\n",
            "        return super(Bidirectional, self).__call__(inputs, **kwargs)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/base_layer.py:989 __call__  *\n",
            "        input_spec.assert_input_compatibility(self.input_spec, inputs, self.name)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/input_spec.py:212 assert_input_compatibility  *\n",
            "        raise ValueError('Input ' + str(input_index) + ' of layer ' +\n",
            "\n",
            "    ValueError: Input 0 of layer bidirectional_77 is incompatible with the layer: expected ndim=3, found ndim=2. Full shape received: (None, 200)\n",
            "\n",
            "\n",
            "  FitFailedWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/model_selection/_validation.py:536: FitFailedWarning: Estimator fit failed. The score on this train-test partition for these parameters will be set to nan. Details: \n",
            "ValueError: in user code:\n",
            "\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:830 train_function  *\n",
            "        return step_function(self, iterator)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:813 run_step  *\n",
            "        outputs = model.train_step(data)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:770 train_step  *\n",
            "        y_pred = self(x, training=True)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/base_layer.py:989 __call__  *\n",
            "        input_spec.assert_input_compatibility(self.input_spec, inputs, self.name)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/sequential.py:389 call  *\n",
            "        outputs = layer(inputs, **kwargs)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/layers/wrappers.py:581 __call__  *\n",
            "        return super(Bidirectional, self).__call__(inputs, **kwargs)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/base_layer.py:989 __call__  *\n",
            "        input_spec.assert_input_compatibility(self.input_spec, inputs, self.name)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/input_spec.py:212 assert_input_compatibility  *\n",
            "        raise ValueError('Input ' + str(input_index) + ' of layer ' +\n",
            "\n",
            "    ValueError: Input 0 of layer bidirectional_79 is incompatible with the layer: expected ndim=3, found ndim=2. Full shape received: (None, 200)\n",
            "\n",
            "\n",
            "  FitFailedWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/model_selection/_validation.py:536: FitFailedWarning: Estimator fit failed. The score on this train-test partition for these parameters will be set to nan. Details: \n",
            "ValueError: in user code:\n",
            "\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:830 train_function  *\n",
            "        return step_function(self, iterator)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:813 run_step  *\n",
            "        outputs = model.train_step(data)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:770 train_step  *\n",
            "        y_pred = self(x, training=True)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/base_layer.py:989 __call__  *\n",
            "        input_spec.assert_input_compatibility(self.input_spec, inputs, self.name)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/sequential.py:389 call  *\n",
            "        outputs = layer(inputs, **kwargs)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/layers/wrappers.py:581 __call__  *\n",
            "        return super(Bidirectional, self).__call__(inputs, **kwargs)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/base_layer.py:989 __call__  *\n",
            "        input_spec.assert_input_compatibility(self.input_spec, inputs, self.name)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/input_spec.py:212 assert_input_compatibility  *\n",
            "        raise ValueError('Input ' + str(input_index) + ' of layer ' +\n",
            "\n",
            "    ValueError: Input 0 of layer bidirectional_81 is incompatible with the layer: expected ndim=3, found ndim=2. Full shape received: (None, 200)\n",
            "\n",
            "\n",
            "  FitFailedWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/model_selection/_validation.py:536: FitFailedWarning: Estimator fit failed. The score on this train-test partition for these parameters will be set to nan. Details: \n",
            "ValueError: in user code:\n",
            "\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:830 train_function  *\n",
            "        return step_function(self, iterator)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:813 run_step  *\n",
            "        outputs = model.train_step(data)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:770 train_step  *\n",
            "        y_pred = self(x, training=True)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/base_layer.py:989 __call__  *\n",
            "        input_spec.assert_input_compatibility(self.input_spec, inputs, self.name)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/sequential.py:389 call  *\n",
            "        outputs = layer(inputs, **kwargs)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/layers/wrappers.py:581 __call__  *\n",
            "        return super(Bidirectional, self).__call__(inputs, **kwargs)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/base_layer.py:989 __call__  *\n",
            "        input_spec.assert_input_compatibility(self.input_spec, inputs, self.name)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/input_spec.py:212 assert_input_compatibility  *\n",
            "        raise ValueError('Input ' + str(input_index) + ' of layer ' +\n",
            "\n",
            "    ValueError: Input 0 of layer bidirectional_83 is incompatible with the layer: expected ndim=3, found ndim=2. Full shape received: (None, 200)\n",
            "\n",
            "\n",
            "  FitFailedWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/model_selection/_validation.py:536: FitFailedWarning: Estimator fit failed. The score on this train-test partition for these parameters will be set to nan. Details: \n",
            "ValueError: in user code:\n",
            "\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:830 train_function  *\n",
            "        return step_function(self, iterator)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:813 run_step  *\n",
            "        outputs = model.train_step(data)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:770 train_step  *\n",
            "        y_pred = self(x, training=True)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/base_layer.py:989 __call__  *\n",
            "        input_spec.assert_input_compatibility(self.input_spec, inputs, self.name)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/sequential.py:389 call  *\n",
            "        outputs = layer(inputs, **kwargs)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/layers/wrappers.py:581 __call__  *\n",
            "        return super(Bidirectional, self).__call__(inputs, **kwargs)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/base_layer.py:989 __call__  *\n",
            "        input_spec.assert_input_compatibility(self.input_spec, inputs, self.name)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/input_spec.py:212 assert_input_compatibility  *\n",
            "        raise ValueError('Input ' + str(input_index) + ' of layer ' +\n",
            "\n",
            "    ValueError: Input 0 of layer bidirectional_85 is incompatible with the layer: expected ndim=3, found ndim=2. Full shape received: (None, 200)\n",
            "\n",
            "\n",
            "  FitFailedWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/model_selection/_validation.py:536: FitFailedWarning: Estimator fit failed. The score on this train-test partition for these parameters will be set to nan. Details: \n",
            "ValueError: in user code:\n",
            "\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:830 train_function  *\n",
            "        return step_function(self, iterator)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:813 run_step  *\n",
            "        outputs = model.train_step(data)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:770 train_step  *\n",
            "        y_pred = self(x, training=True)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/base_layer.py:989 __call__  *\n",
            "        input_spec.assert_input_compatibility(self.input_spec, inputs, self.name)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/sequential.py:389 call  *\n",
            "        outputs = layer(inputs, **kwargs)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/layers/wrappers.py:581 __call__  *\n",
            "        return super(Bidirectional, self).__call__(inputs, **kwargs)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/base_layer.py:989 __call__  *\n",
            "        input_spec.assert_input_compatibility(self.input_spec, inputs, self.name)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/input_spec.py:212 assert_input_compatibility  *\n",
            "        raise ValueError('Input ' + str(input_index) + ' of layer ' +\n",
            "\n",
            "    ValueError: Input 0 of layer bidirectional_88 is incompatible with the layer: expected ndim=3, found ndim=2. Full shape received: (None, 200)\n",
            "\n",
            "\n",
            "  FitFailedWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/model_selection/_validation.py:536: FitFailedWarning: Estimator fit failed. The score on this train-test partition for these parameters will be set to nan. Details: \n",
            "ValueError: in user code:\n",
            "\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:830 train_function  *\n",
            "        return step_function(self, iterator)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:813 run_step  *\n",
            "        outputs = model.train_step(data)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:770 train_step  *\n",
            "        y_pred = self(x, training=True)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/base_layer.py:989 __call__  *\n",
            "        input_spec.assert_input_compatibility(self.input_spec, inputs, self.name)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/sequential.py:389 call  *\n",
            "        outputs = layer(inputs, **kwargs)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/layers/wrappers.py:581 __call__  *\n",
            "        return super(Bidirectional, self).__call__(inputs, **kwargs)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/base_layer.py:989 __call__  *\n",
            "        input_spec.assert_input_compatibility(self.input_spec, inputs, self.name)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/input_spec.py:212 assert_input_compatibility  *\n",
            "        raise ValueError('Input ' + str(input_index) + ' of layer ' +\n",
            "\n",
            "    ValueError: Input 0 of layer bidirectional_91 is incompatible with the layer: expected ndim=3, found ndim=2. Full shape received: (None, 200)\n",
            "\n",
            "\n",
            "  FitFailedWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/model_selection/_validation.py:536: FitFailedWarning: Estimator fit failed. The score on this train-test partition for these parameters will be set to nan. Details: \n",
            "ValueError: in user code:\n",
            "\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:830 train_function  *\n",
            "        return step_function(self, iterator)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:813 run_step  *\n",
            "        outputs = model.train_step(data)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/training.py:770 train_step  *\n",
            "        y_pred = self(x, training=True)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/base_layer.py:989 __call__  *\n",
            "        input_spec.assert_input_compatibility(self.input_spec, inputs, self.name)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/sequential.py:389 call  *\n",
            "        outputs = layer(inputs, **kwargs)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/layers/wrappers.py:581 __call__  *\n",
            "        return super(Bidirectional, self).__call__(inputs, **kwargs)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/base_layer.py:989 __call__  *\n",
            "        input_spec.assert_input_compatibility(self.input_spec, inputs, self.name)\n",
            "    /usr/local/lib/python3.7/dist-packages/keras/engine/input_spec.py:212 assert_input_compatibility  *\n",
            "        raise ValueError('Input ' + str(input_index) + ' of layer ' +\n",
            "\n",
            "    ValueError: Input 0 of layer bidirectional_94 is incompatible with the layer: expected ndim=3, found ndim=2. Full shape received: (None, 200)\n",
            "\n",
            "\n",
            "  FitFailedWarning)\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Best: -9.754916 using {'batch_size': 16, 'epochs': 100, 'layers': [100]}\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "j2wZgmBbYLOS",
        "outputId": "373cd8b3-89b6-4787-d59b-718804f5faef"
      },
      "source": [
        "model = Sequential()\n",
        "model.add(Bidirectional(LSTM(units=100,input_shape= (train_X.shape[1],1))))\n",
        "model.add(Dropout(0.3))\n",
        "model.add(Dense(units=1))\n",
        "\n",
        "es = tf.keras.callbacks.EarlyStopping(monitor='val_loss',patience=20,mode='min',verbose=1)\n",
        "mc = tf.keras.callbacks.ModelCheckpoint('best_model.h5', monitor='val_loss', mode='min', verbose=1, \n",
        "                     save_best_only=True)\n",
        "model.compile(loss='mae', optimizer='adam')\n",
        "history = model.fit(train_X, train_y, epochs=100, batch_size=16,validation_data=(test_X,test_y),\n",
        "                    verbose=0,callbacks=[es,mc], shuffle=False)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "\n",
            "Epoch 00001: val_loss improved from inf to 0.16841, saving model to best_model.h5\n",
            "\n",
            "Epoch 00002: val_loss improved from 0.16841 to 0.14742, saving model to best_model.h5\n",
            "\n",
            "Epoch 00003: val_loss improved from 0.14742 to 0.14200, saving model to best_model.h5\n",
            "\n",
            "Epoch 00004: val_loss improved from 0.14200 to 0.11482, saving model to best_model.h5\n",
            "\n",
            "Epoch 00005: val_loss improved from 0.11482 to 0.09796, saving model to best_model.h5\n",
            "\n",
            "Epoch 00006: val_loss did not improve from 0.09796\n",
            "\n",
            "Epoch 00007: val_loss did not improve from 0.09796\n",
            "\n",
            "Epoch 00008: val_loss did not improve from 0.09796\n",
            "\n",
            "Epoch 00009: val_loss did not improve from 0.09796\n",
            "\n",
            "Epoch 00010: val_loss improved from 0.09796 to 0.08866, saving model to best_model.h5\n",
            "\n",
            "Epoch 00011: val_loss did not improve from 0.08866\n",
            "\n",
            "Epoch 00012: val_loss did not improve from 0.08866\n",
            "\n",
            "Epoch 00013: val_loss did not improve from 0.08866\n",
            "\n",
            "Epoch 00014: val_loss improved from 0.08866 to 0.08507, saving model to best_model.h5\n",
            "\n",
            "Epoch 00015: val_loss did not improve from 0.08507\n",
            "\n",
            "Epoch 00016: val_loss did not improve from 0.08507\n",
            "\n",
            "Epoch 00017: val_loss improved from 0.08507 to 0.08181, saving model to best_model.h5\n",
            "\n",
            "Epoch 00018: val_loss did not improve from 0.08181\n",
            "\n",
            "Epoch 00019: val_loss improved from 0.08181 to 0.08065, saving model to best_model.h5\n",
            "\n",
            "Epoch 00020: val_loss did not improve from 0.08065\n",
            "\n",
            "Epoch 00021: val_loss did not improve from 0.08065\n",
            "\n",
            "Epoch 00022: val_loss improved from 0.08065 to 0.07902, saving model to best_model.h5\n",
            "\n",
            "Epoch 00023: val_loss improved from 0.07902 to 0.07228, saving model to best_model.h5\n",
            "\n",
            "Epoch 00024: val_loss did not improve from 0.07228\n",
            "\n",
            "Epoch 00025: val_loss did not improve from 0.07228\n",
            "\n",
            "Epoch 00026: val_loss improved from 0.07228 to 0.07046, saving model to best_model.h5\n",
            "\n",
            "Epoch 00027: val_loss improved from 0.07046 to 0.06608, saving model to best_model.h5\n",
            "\n",
            "Epoch 00028: val_loss improved from 0.06608 to 0.06564, saving model to best_model.h5\n",
            "\n",
            "Epoch 00029: val_loss did not improve from 0.06564\n",
            "\n",
            "Epoch 00030: val_loss did not improve from 0.06564\n",
            "\n",
            "Epoch 00031: val_loss did not improve from 0.06564\n",
            "\n",
            "Epoch 00032: val_loss did not improve from 0.06564\n",
            "\n",
            "Epoch 00033: val_loss improved from 0.06564 to 0.06344, saving model to best_model.h5\n",
            "\n",
            "Epoch 00034: val_loss improved from 0.06344 to 0.05799, saving model to best_model.h5\n",
            "\n",
            "Epoch 00035: val_loss did not improve from 0.05799\n",
            "\n",
            "Epoch 00036: val_loss did not improve from 0.05799\n",
            "\n",
            "Epoch 00037: val_loss did not improve from 0.05799\n",
            "\n",
            "Epoch 00038: val_loss did not improve from 0.05799\n",
            "\n",
            "Epoch 00039: val_loss did not improve from 0.05799\n",
            "\n",
            "Epoch 00040: val_loss did not improve from 0.05799\n",
            "\n",
            "Epoch 00041: val_loss did not improve from 0.05799\n",
            "\n",
            "Epoch 00042: val_loss did not improve from 0.05799\n",
            "\n",
            "Epoch 00043: val_loss did not improve from 0.05799\n",
            "\n",
            "Epoch 00044: val_loss improved from 0.05799 to 0.05726, saving model to best_model.h5\n",
            "\n",
            "Epoch 00045: val_loss did not improve from 0.05726\n",
            "\n",
            "Epoch 00046: val_loss did not improve from 0.05726\n",
            "\n",
            "Epoch 00047: val_loss did not improve from 0.05726\n",
            "\n",
            "Epoch 00048: val_loss did not improve from 0.05726\n",
            "\n",
            "Epoch 00049: val_loss did not improve from 0.05726\n",
            "\n",
            "Epoch 00050: val_loss did not improve from 0.05726\n",
            "\n",
            "Epoch 00051: val_loss did not improve from 0.05726\n",
            "\n",
            "Epoch 00052: val_loss did not improve from 0.05726\n",
            "\n",
            "Epoch 00053: val_loss did not improve from 0.05726\n",
            "\n",
            "Epoch 00054: val_loss did not improve from 0.05726\n",
            "\n",
            "Epoch 00055: val_loss did not improve from 0.05726\n",
            "\n",
            "Epoch 00056: val_loss did not improve from 0.05726\n",
            "\n",
            "Epoch 00057: val_loss did not improve from 0.05726\n",
            "\n",
            "Epoch 00058: val_loss did not improve from 0.05726\n",
            "\n",
            "Epoch 00059: val_loss did not improve from 0.05726\n",
            "\n",
            "Epoch 00060: val_loss did not improve from 0.05726\n",
            "\n",
            "Epoch 00061: val_loss did not improve from 0.05726\n",
            "\n",
            "Epoch 00062: val_loss did not improve from 0.05726\n",
            "\n",
            "Epoch 00063: val_loss did not improve from 0.05726\n",
            "\n",
            "Epoch 00064: val_loss did not improve from 0.05726\n",
            "Epoch 00064: early stopping\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "S2DtZsevRZXZ",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 795
        },
        "outputId": "f9f5654f-63d1-49a6-c126-0a1c5b588e90"
      },
      "source": [
        "# plot history\n",
        "pyplot.plot(history.history['loss'], label='train')\n",
        "pyplot.plot(history.history['val_loss'], label='test')\n",
        "pyplot.legend()\n",
        "pyplot.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1440x1440 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "w8CVRM2TS92m"
      },
      "source": [
        "better stability with df_logic"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "S_KMCaeInjNn",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "3e333d7c-56b2-46eb-ddf2-6aa2df207970"
      },
      "source": [
        "  # load the saved model\n",
        "model = load_model('best_model.h5')\n",
        "model.summary()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"sequential\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "bidirectional (Bidirectional (None, 200)               81600     \n",
            "_________________________________________________________________\n",
            "dropout (Dropout)            (None, 200)               0         \n",
            "_________________________________________________________________\n",
            "dense (Dense)                (None, 1)                 201       \n",
            "=================================================================\n",
            "Total params: 81,801\n",
            "Trainable params: 81,801\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Hsk_n93VtEES"
      },
      "source": [
        "# make a prediction\n",
        "yhat = model.predict(test_X) #shape=(14,1)\n",
        "test_X = test_X.reshape((test_X.shape[0], test_X.shape[1]))\n",
        "# invert scaling for forecast\n",
        "#inv_yhat = concatenate((test_X[:, :-1],yhat), axis=1)\n",
        "inv_yhat = concatenate((test_X,yhat), axis=1)\n",
        "inv_yhat = scaler.inverse_transform(inv_yhat)\n",
        "inv_yhat = inv_yhat[:,-1]\n",
        "# invert scaling for actual\n",
        "test_y = test_y.reshape((len(test_y), 1))\n",
        "#inv_y = concatenate((test_X[:, :-1],test_y), axis=1)\n",
        "inv_y = concatenate((test_X,test_y), axis=1)\n",
        "inv_y = scaler.inverse_transform(inv_y)\n",
        "inv_y = inv_y[:,-1]"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "CejIp8s7rvO6"
      },
      "source": [
        "def timeseries_evaluation_metrics_func(y_true, y_pred):\n",
        "  def mean_absolute_percentage_error(y_true, y_pred):\n",
        "    y_true, y_pred = np.array(y_true), np.array(y_pred)\n",
        "    return np.mean(np.abs((y_true - y_pred) / y_true)) * 100\n",
        "  def smape(y_true,y_pred):\n",
        "    y_true, y_pred = np.array(y_true), np.array(y_pred)\n",
        "    return 2/y_true.shape[0]*np.sum(np.abs(y_true-y_pred))/np.sum((np.abs(y_true)+np.abs(y_pred)))*100\n",
        "  def mase(y_true,y_pred):\n",
        "    y_true, y_pred = np.array(y_true), np.array(y_pred)\n",
        "    naive_pred=y_true[1:]\n",
        "    return 1/y_true.shape[0]*np.sum(np.abs(y_true-y_pred))/np.sum(np.abs(y_true[:len(y_true)-1]-naive_pred))*(y_true.shape[0]-1)\n",
        "  #def owe(y_true,y_pred):\n",
        "   # return (mase(y_true,y_pred)+smape(y_true,y_pred))/2\n",
        "\n",
        "  print('Evaluation metric results:-')\n",
        "  print(f'MSE is : {metrics.mean_squared_error(y_true, y_pred)}')\n",
        "  print(f'MAE is : {metrics.mean_absolute_error(y_true, y_pred)}')\n",
        "  print(f'RMSE is : {np.sqrt(metrics.mean_squared_error(y_true, y_pred))}')\n",
        "  print(f'MAPE is : {mean_absolute_percentage_error(y_true, y_pred)} %')\n",
        "  #mape.append(mean_absolute_percentage_error(y_true, y_pred))\n",
        "  #rmse.append(np.sqrt(metrics.mean_squared_error(y_true, y_pred)))\n",
        "  print(f'R2 is : {metrics.r2_score(y_true, y_pred)}',end='\\n\\n')\n",
        "  print('SMAPE is : '+str(smape(y_true,y_pred))+'\\n')\n",
        "  print('MASE is : '+str(mase(y_true,y_pred)))\n",
        "  return"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "U-yxUOktrvTD",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "c2e307b5-b2bd-49f8-c96a-0cfcb73ae9fa"
      },
      "source": [
        "from sklearn import metrics\n",
        "import numpy as np\n",
        "\n",
        "timeseries_evaluation_metrics_func(inv_y, inv_yhat)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Evaluation metric results:-\n",
            "MSE is : 10910.490366684702\n",
            "MAE is : 70.09039416653769\n",
            "RMSE is : 104.45329275175916\n",
            "MAPE is : 6.591499943222379 %\n",
            "R2 is : 0.7192013341129104\n",
            "\n",
            "SMAPE is : 0.027731674065364195\n",
            "\n",
            "MASE is : 0.8037801031992527\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "smzoQvxrrvUh",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 413
        },
        "outputId": "4ff645b9-2052-4739-aa8f-cf8ae26185de"
      },
      "source": [
        "from matplotlib.pyplot import figure\n",
        "\n",
        "figure(figsize=(8, 6), dpi=80)\n",
        "plt.plot(time[train_X.shape[0]:len(time)],inv_y,label='actual')\n",
        "plt.plot(time[train_X.shape[0]:len(time)],inv_yhat,label='Pred')\n",
        "plt.legend()\n",
        "#plt.savefig('xgboost_test')\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "U2XJYkXANBWI"
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "creouAuRJ2qH"
      },
      "source": [
        "# peak period"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Owa5qPllJ4_7",
        "outputId": "5b7a0289-2fbe-4c85-9f44-a39f7d556ae2"
      },
      "source": [
        "timeseries_evaluation_metrics_func(inv_y[:90], inv_yhat[:90])"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Evaluation metric results:-\n",
            "MSE is : 10169.996740088805\n",
            "MAE is : 70.62585487895542\n",
            "RMSE is : 100.84640172107682\n",
            "MAPE is : 5.6264433016445645 %\n",
            "R2 is : 0.39118358883188875\n",
            "\n",
            "SMAPE is : 0.05922127037615825\n",
            "\n",
            "MASE is : 0.7054658904856375\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 413
        },
        "id": "Yfk5WcvcNBlt",
        "outputId": "d1d02e54-e1fe-4993-a997-b64c19178aca"
      },
      "source": [
        "from matplotlib.pyplot import figure\n",
        "\n",
        "figure(figsize=(8, 6), dpi=80)\n",
        "plt.plot(time[train_X.shape[0]:train_X.shape[0]+90],inv_y[:90],label='actual')\n",
        "plt.plot(time[train_X.shape[0]:train_X.shape[0]+90],inv_yhat[:90],label='Pred')\n",
        "plt.legend()\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    }
  ]
}