Giotto_Time_1.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Giotto Time 1.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "authorship_tag": "ABX9TyPpn9abPZTg0lxO38job8m0",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/RajkumarGalaxy/StructuredData/blob/master/Giotto_Time_1.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "TpI3cNYHQ3e-"
      },
      "source": [
        "# GIOTTO-TIME\n",
        "\n",
        "References:\n",
        "\n",
        "https://docs-time.giotto.ai/\n",
        "\n",
        "https://github.com/giotto-ai/giotto-time"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Oj7A6s4EQxWJ",
        "outputId": "986149b1-5a04-4116-8800-e651bbf81403"
      },
      "source": [
        "!pip install giotto-time\n"
      ],
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Collecting giotto-time\n",
            "\u001b[?25l  Downloading https://files.pythonhosted.org/packages/17/b2/3e52755ca838831e705587c70b5d26d002af92e2c1b19163d77078798267/giotto_time-0.2.1-py3-none-any.whl (83kB)\n",
            "\r\u001b[K     |████                            | 10kB 15.7MB/s eta 0:00:01\r\u001b[K     |███████▉                        | 20kB 12.6MB/s eta 0:00:01\r\u001b[K     |███████████▊                    | 30kB 12.1MB/s eta 0:00:01\r\u001b[K     |███████████████▊                | 40kB 8.7MB/s eta 0:00:01\r\u001b[K     |███████████████████▋            | 51kB 5.4MB/s eta 0:00:01\r\u001b[K     |███████████████████████▌        | 61kB 5.4MB/s eta 0:00:01\r\u001b[K     |███████████████████████████▌    | 71kB 6.0MB/s eta 0:00:01\r\u001b[K     |███████████████████████████████▍| 81kB 6.6MB/s eta 0:00:01\r\u001b[K     |████████████████████████████████| 92kB 3.7MB/s \n",
            "\u001b[?25hRequirement already satisfied: scikit-learn>=0.22.0 in /usr/local/lib/python3.7/dist-packages (from giotto-time) (0.22.2.post1)\n",
            "Requirement already satisfied: scipy>=0.17.0 in /usr/local/lib/python3.7/dist-packages (from giotto-time) (1.4.1)\n",
            "Collecting workalendar>=7.1.1\n",
            "\u001b[?25l  Downloading https://files.pythonhosted.org/packages/97/e0/d70a24739f5f357dba42e71b56776e0699cd0509c4821f2901cbf31dc96e/workalendar-15.1.0-py3-none-any.whl (190kB)\n",
            "\u001b[K     |████████████████████████████████| 194kB 8.4MB/s \n",
            "\u001b[?25hRequirement already satisfied: pandas>=0.25.3 in /usr/local/lib/python3.7/dist-packages (from giotto-time) (1.1.5)\n",
            "Requirement already satisfied: joblib>=0.11 in /usr/local/lib/python3.7/dist-packages (from scikit-learn>=0.22.0->giotto-time) (1.0.1)\n",
            "Requirement already satisfied: numpy>=1.11.0 in /usr/local/lib/python3.7/dist-packages (from scikit-learn>=0.22.0->giotto-time) (1.19.5)\n",
            "Collecting skyfield\n",
            "\u001b[?25l  Downloading https://files.pythonhosted.org/packages/64/c3/d9d51275751d649aead4a107c9cc9a2745097281776c90a41ed36e4a83b2/skyfield-1.39.tar.gz (373kB)\n",
            "\u001b[K     |████████████████████████████████| 378kB 9.1MB/s \n",
            "\u001b[?25hRequirement already satisfied: python-dateutil in /usr/local/lib/python3.7/dist-packages (from workalendar>=7.1.1->giotto-time) (2.8.1)\n",
            "Collecting lunardate\n",
            "  Downloading https://files.pythonhosted.org/packages/4e/7e/377a3cbba646ec0cf79433ef858881d809a3b87eb887b0901cb83c66a758/lunardate-0.2.0-py3-none-any.whl\n",
            "Collecting skyfield-data\n",
            "\u001b[?25l  Downloading https://files.pythonhosted.org/packages/61/ea/9f32d84708069f4cef56d682ace8d0c3f7f4ee4550bba295e49ff4cf7590/skyfield_data-3.0.0-py2.py3-none-any.whl (16.9MB)\n",
            "\u001b[K     |████████████████████████████████| 16.9MB 215kB/s \n",
            "\u001b[?25hCollecting backports.zoneinfo; python_version < \"3.9\"\n",
            "\u001b[?25l  Downloading https://files.pythonhosted.org/packages/4c/7e/ed8af95bed90eeccfb4a4fe6ec424bc7a79e1aa983e54dd1d9062d9fa20b/backports.zoneinfo-0.2.1-cp37-cp37m-manylinux1_x86_64.whl (70kB)\n",
            "\u001b[K     |████████████████████████████████| 71kB 5.4MB/s \n",
            "\u001b[?25hCollecting pyluach\n",
            "  Downloading https://files.pythonhosted.org/packages/35/e1/0feb20f4e2cdc879843673a295dc6e5f91af8efdc6ba8b6f028f1b1ddb21/pyluach-1.2.1-py3-none-any.whl\n",
            "Requirement already satisfied: setuptools>=1.0 in /usr/local/lib/python3.7/dist-packages (from workalendar>=7.1.1->giotto-time) (54.2.0)\n",
            "Collecting pyCalverter\n",
            "  Downloading https://files.pythonhosted.org/packages/4f/5c/57c6853f7a5bc41fc9da7651ae67b9c76381083742613faa7381724081e9/pyCalverter-1.6.1.tar.gz\n",
            "Requirement already satisfied: pytz>=2017.2 in /usr/local/lib/python3.7/dist-packages (from pandas>=0.25.3->giotto-time) (2018.9)\n",
            "Requirement already satisfied: certifi>=2017.4.17 in /usr/local/lib/python3.7/dist-packages (from skyfield->workalendar>=7.1.1->giotto-time) (2020.12.5)\n",
            "Collecting jplephem>=2.13\n",
            "\u001b[?25l  Downloading https://files.pythonhosted.org/packages/c9/f0/e57456436531333455f40864622c5ae359654ccede8fe1156c3c2ac6c492/jplephem-2.15.tar.gz (40kB)\n",
            "\u001b[K     |████████████████████████████████| 40kB 4.1MB/s \n",
            "\u001b[?25hCollecting sgp4>=2.2\n",
            "\u001b[?25l  Downloading https://files.pythonhosted.org/packages/9a/0a/ef3ac594e96a33929234abe7e52e0f95e66c67e9a1503b6eb63b62121da0/sgp4-2.18-cp37-cp37m-manylinux2010_x86_64.whl (255kB)\n",
            "\u001b[K     |████████████████████████████████| 256kB 51.1MB/s \n",
            "\u001b[?25hRequirement already satisfied: six>=1.5 in /usr/local/lib/python3.7/dist-packages (from python-dateutil->workalendar>=7.1.1->giotto-time) (1.15.0)\n",
            "Building wheels for collected packages: skyfield, pyCalverter, jplephem\n",
            "  Building wheel for skyfield (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "  Created wheel for skyfield: filename=skyfield-1.39-cp37-none-any.whl size=417567 sha256=ab128ab807d0b873862c94fde7b1183badef3fe9c19e4ed464cb923424817007\n",
            "  Stored in directory: /root/.cache/pip/wheels/bb/a4/a6/2f7e3bc8ec08ffc7485c92d18dd726e0dbafc0f7a8b1519518\n",
            "  Building wheel for pyCalverter (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "  Created wheel for pyCalverter: filename=pyCalverter-1.6.1-cp37-none-any.whl size=4201 sha256=548fe0a031c89acfca1a2fa0cd14368acb21a80a4147addf9fa26541c833d6a9\n",
            "  Stored in directory: /root/.cache/pip/wheels/77/4d/86/db4ff4eca6178dbbd5a365d095f97f6021e2a48f09908be79c\n",
            "  Building wheel for jplephem (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "  Created wheel for jplephem: filename=jplephem-2.15-cp37-none-any.whl size=46020 sha256=0ba676f57cac3fdb47406e783b09f79f4b40abdb6a87bb42d12ea11eef9ac08c\n",
            "  Stored in directory: /root/.cache/pip/wheels/01/71/9c/beaa33beb607ef67b62765324a10cd3584e30fbeef8856b7a8\n",
            "Successfully built skyfield pyCalverter jplephem\n",
            "Installing collected packages: jplephem, sgp4, skyfield, lunardate, skyfield-data, backports.zoneinfo, pyluach, pyCalverter, workalendar, giotto-time\n",
            "Successfully installed backports.zoneinfo-0.2.1 giotto-time-0.2.1 jplephem-2.15 lunardate-0.2.0 pyCalverter-1.6.1 pyluach-1.2.1 sgp4-2.18 skyfield-1.39 skyfield-data-3.0.0 workalendar-15.1.0\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "JENPksuaDoa0"
      },
      "source": [
        "# Basic Time-series Forecasting"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8qB6XrSNRNg8"
      },
      "source": [
        "import pandas as pd\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "from gtime.preprocessing import TimeSeriesPreparation\n",
        "from gtime.compose import FeatureCreation\n",
        "from gtime.feature_extraction import Shift, MovingAverage\n",
        "from gtime.feature_generation import PeriodicSeasonal, Constant, Calendar\n",
        "from gtime.model_selection import horizon_shift, FeatureSplitter\n",
        "from gtime.forecasting import GAR"
      ],
      "execution_count": 3,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "NoSeBl4sdja_"
      },
      "source": [
        "from sklearn.linear_model import LinearRegression"
      ],
      "execution_count": 4,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "09b1Tx6lgf2m"
      },
      "source": [
        "def test_time_series():\n",
        "    from pandas.util import testing as testing\n",
        "\n",
        "    testing.N, testing.K = 500, 1\n",
        "    df = testing.makeTimeDataFrame( freq=\"D\" )\n",
        "    return df"
      ],
      "execution_count": 5,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "a3ap3odygwQk",
        "outputId": "f22e92a5-cc2d-4b7d-846c-ae8c231de243"
      },
      "source": [
        "time_series = test_time_series()\n",
        "print(f'Time series shape: {time_series.shape}')\n",
        "print(f'Time series index type: {time_series.index.__class__}')"
      ],
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Time series shape: (30, 4)\n",
            "Time series index type: <class 'pandas.core.indexes.datetimes.DatetimeIndex'>\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/pandas/util/__init__.py:12: FutureWarning: pandas.util.testing is deprecated. Use the functions in the public API at pandas.testing instead.\n",
            "  import pandas.util.testing\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "c3A957QwhZBP",
        "outputId": "9b39f1d7-0993-429d-c3b6-468487a83603"
      },
      "source": [
        "time_series_preparation = TimeSeriesPreparation()\n",
        "period_index_time_series = time_series_preparation.transform(time_series)\n",
        "print(f'Time series index type after the preprocessing: \\n {period_index_time_series.index.__class__}')"
      ],
      "execution_count": 7,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Time series index type after the preprocessing: \n",
            " <class 'pandas.core.indexes.period.PeriodIndex'>\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 355
        },
        "id": "pbq19COEhj7F",
        "outputId": "3d4bb6e6-054a-43bb-8fa3-a750ef2953cc"
      },
      "source": [
        "period_index_time_series.plot(figsize=(10, 5))\n",
        "plt.show()"
      ],
      "execution_count": 18,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x360 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "k2Ja9634jDkG"
      },
      "source": [
        "\n",
        "cal = Calendar(\n",
        "    start_date=\"ignored\",\n",
        "    end_date=\"ignored\",\n",
        "    region=\"america\",\n",
        "    country=\"Brazil\",\n",
        "    kernel=np.array([0, 1]),\n",
        ")\n",
        "# New API \n",
        "dft = FeatureCreation(\n",
        "    [('s0', Shift(0), ['time_series']), \n",
        "     ('s1', Shift(1), ['time_series']),\n",
        "     ('ma3', MovingAverage(window_size=3), ['time_series']),\n",
        "    ])"
      ],
      "execution_count": 19,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 235
        },
        "id": "zRcLoYkjjIUk",
        "outputId": "2229500f-0425-4a7b-e360-a852b71185d8"
      },
      "source": [
        "X = dft.fit_transform(period_index_time_series)\n",
        "X.head(6)\n"
      ],
      "execution_count": 21,
      "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>s0__time_series__Shift</th>\n",
              "      <th>s1__time_series__Shift</th>\n",
              "      <th>ma3__time_series__MovingAverage</th>\n",
              "      <th>cal__Calendar</th>\n",
              "      <th>ct__Constant</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>2000-01-25</th>\n",
              "      <td>0.772001</td>\n",
              "      <td>1.477305</td>\n",
              "      <td>1.601956</td>\n",
              "      <td>0.0</td>\n",
              "      <td>2.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2000-01-26</th>\n",
              "      <td>-0.142176</td>\n",
              "      <td>0.772001</td>\n",
              "      <td>0.702377</td>\n",
              "      <td>0.0</td>\n",
              "      <td>2.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2000-01-27</th>\n",
              "      <td>-2.427659</td>\n",
              "      <td>-0.142176</td>\n",
              "      <td>-0.599278</td>\n",
              "      <td>0.0</td>\n",
              "      <td>2.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2000-01-28</th>\n",
              "      <td>0.329416</td>\n",
              "      <td>-2.427659</td>\n",
              "      <td>-0.746806</td>\n",
              "      <td>0.0</td>\n",
              "      <td>2.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2000-01-29</th>\n",
              "      <td>-0.948387</td>\n",
              "      <td>0.329416</td>\n",
              "      <td>-1.015543</td>\n",
              "      <td>0.0</td>\n",
              "      <td>2.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2000-01-30</th>\n",
              "      <td>1.989901</td>\n",
              "      <td>-0.948387</td>\n",
              "      <td>0.456977</td>\n",
              "      <td>0.0</td>\n",
              "      <td>2.0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "            s0__time_series__Shift  ...  ct__Constant\n",
              "2000-01-25                0.772001  ...           2.0\n",
              "2000-01-26               -0.142176  ...           2.0\n",
              "2000-01-27               -2.427659  ...           2.0\n",
              "2000-01-28                0.329416  ...           2.0\n",
              "2000-01-29               -0.948387  ...           2.0\n",
              "2000-01-30                1.989901  ...           2.0\n",
              "\n",
              "[6 rows x 5 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 21
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 204
        },
        "id": "9q3XWzR3jMPl",
        "outputId": "79d5fd31-41cc-4ed6-c46e-2a79b54d05f6"
      },
      "source": [
        "y = horizon_shift(period_index_time_series, horizon=3)\n",
        "y.head()"
      ],
      "execution_count": 11,
      "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>y_1</th>\n",
              "      <th>y_2</th>\n",
              "      <th>y_3</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>2000-01-01</th>\n",
              "      <td>-1.171712</td>\n",
              "      <td>0.729488</td>\n",
              "      <td>-1.423077</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2000-01-02</th>\n",
              "      <td>0.729488</td>\n",
              "      <td>-1.423077</td>\n",
              "      <td>-0.611391</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2000-01-03</th>\n",
              "      <td>-1.423077</td>\n",
              "      <td>-0.611391</td>\n",
              "      <td>-0.428489</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2000-01-04</th>\n",
              "      <td>-0.611391</td>\n",
              "      <td>-0.428489</td>\n",
              "      <td>0.857282</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2000-01-05</th>\n",
              "      <td>-0.428489</td>\n",
              "      <td>0.857282</td>\n",
              "      <td>-0.164129</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                 y_1       y_2       y_3\n",
              "2000-01-01 -1.171712  0.729488 -1.423077\n",
              "2000-01-02  0.729488 -1.423077 -0.611391\n",
              "2000-01-03 -1.423077 -0.611391 -0.428489\n",
              "2000-01-04 -0.611391 -0.428489  0.857282\n",
              "2000-01-05 -0.428489  0.857282 -0.164129"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 11
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 204
        },
        "id": "LPwecJOqjRkx",
        "outputId": "79ea9668-9fd5-41af-ac12-ec46c716aea7"
      },
      "source": [
        "feature_splitter = FeatureSplitter()\n",
        "X_train, y_train, X_test, y_test = feature_splitter.transform(X, y)\n",
        "X_train.tail()"
      ],
      "execution_count": 12,
      "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>s0__time_series__Shift</th>\n",
              "      <th>s1__time_series__Shift</th>\n",
              "      <th>ma3__time_series__MovingAverage</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>2000-01-23</th>\n",
              "      <td>2.556562</td>\n",
              "      <td>-0.611135</td>\n",
              "      <td>1.032918</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2000-01-24</th>\n",
              "      <td>1.477305</td>\n",
              "      <td>2.556562</td>\n",
              "      <td>1.140911</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2000-01-25</th>\n",
              "      <td>0.772001</td>\n",
              "      <td>1.477305</td>\n",
              "      <td>1.601956</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2000-01-26</th>\n",
              "      <td>-0.142176</td>\n",
              "      <td>0.772001</td>\n",
              "      <td>0.702377</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2000-01-27</th>\n",
              "      <td>-2.427659</td>\n",
              "      <td>-0.142176</td>\n",
              "      <td>-0.599278</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "            s0__time_series__Shift  ...  ma3__time_series__MovingAverage\n",
              "2000-01-23                2.556562  ...                         1.032918\n",
              "2000-01-24                1.477305  ...                         1.140911\n",
              "2000-01-25                0.772001  ...                         1.601956\n",
              "2000-01-26               -0.142176  ...                         0.702377\n",
              "2000-01-27               -2.427659  ...                        -0.599278\n",
              "\n",
              "[5 rows x 3 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 12
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 142
        },
        "id": "UlX5a4gtjXq9",
        "outputId": "848f2756-7bf6-479b-8a0b-221933c3d431"
      },
      "source": [
        "X_test\n"
      ],
      "execution_count": 13,
      "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>s0__time_series__Shift</th>\n",
              "      <th>s1__time_series__Shift</th>\n",
              "      <th>ma3__time_series__MovingAverage</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>2000-01-28</th>\n",
              "      <td>0.329416</td>\n",
              "      <td>-2.427659</td>\n",
              "      <td>-0.746806</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2000-01-29</th>\n",
              "      <td>-0.948387</td>\n",
              "      <td>0.329416</td>\n",
              "      <td>-1.015543</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2000-01-30</th>\n",
              "      <td>1.989901</td>\n",
              "      <td>-0.948387</td>\n",
              "      <td>0.456977</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "            s0__time_series__Shift  ...  ma3__time_series__MovingAverage\n",
              "2000-01-28                0.329416  ...                        -0.746806\n",
              "2000-01-29               -0.948387  ...                        -1.015543\n",
              "2000-01-30                1.989901  ...                         0.456977\n",
              "\n",
              "[3 rows x 3 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 13
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 204
        },
        "id": "Bb5247yHjcXu",
        "outputId": "80452cea-36e6-4628-c6df-24fe25df7ea3"
      },
      "source": [
        "y_train.tail()\n"
      ],
      "execution_count": 14,
      "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>y_1</th>\n",
              "      <th>y_2</th>\n",
              "      <th>y_3</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>2000-01-23</th>\n",
              "      <td>1.477305</td>\n",
              "      <td>0.772001</td>\n",
              "      <td>-0.142176</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2000-01-24</th>\n",
              "      <td>0.772001</td>\n",
              "      <td>-0.142176</td>\n",
              "      <td>-2.427659</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2000-01-25</th>\n",
              "      <td>-0.142176</td>\n",
              "      <td>-2.427659</td>\n",
              "      <td>0.329416</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2000-01-26</th>\n",
              "      <td>-2.427659</td>\n",
              "      <td>0.329416</td>\n",
              "      <td>-0.948387</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2000-01-27</th>\n",
              "      <td>0.329416</td>\n",
              "      <td>-0.948387</td>\n",
              "      <td>1.989901</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                 y_1       y_2       y_3\n",
              "2000-01-23  1.477305  0.772001 -0.142176\n",
              "2000-01-24  0.772001 -0.142176 -2.427659\n",
              "2000-01-25 -0.142176 -2.427659  0.329416\n",
              "2000-01-26 -2.427659  0.329416 -0.948387\n",
              "2000-01-27  0.329416 -0.948387  1.989901"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 14
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 142
        },
        "id": "DWxP4IZPjfk6",
        "outputId": "e7faea32-fa5f-4e68-a292-c355c8c85435"
      },
      "source": [
        "y_test"
      ],
      "execution_count": 15,
      "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>y_1</th>\n",
              "      <th>y_2</th>\n",
              "      <th>y_3</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>2000-01-28</th>\n",
              "      <td>-0.948387</td>\n",
              "      <td>1.989901</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2000-01-29</th>\n",
              "      <td>1.989901</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2000-01-30</th>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                 y_1       y_2  y_3\n",
              "2000-01-28 -0.948387  1.989901  NaN\n",
              "2000-01-29  1.989901       NaN  NaN\n",
              "2000-01-30       NaN       NaN  NaN"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 15
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "YmqH-Ex3jiBk"
      },
      "source": [
        "lr = LinearRegression()\n",
        "model = GAR(lr)\n",
        "model = model.fit(X_train, y_train)"
      ],
      "execution_count": 16,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 142
        },
        "id": "MUnG55lrjtib",
        "outputId": "528802c2-27b0-4bbc-e5e7-5a27d7c06d16"
      },
      "source": [
        "predictions = model.predict(X_test)\n",
        "predictions"
      ],
      "execution_count": 17,
      "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>y_1</th>\n",
              "      <th>y_2</th>\n",
              "      <th>y_3</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>2000-01-28</th>\n",
              "      <td>-0.081028</td>\n",
              "      <td>0.820356</td>\n",
              "      <td>0.555104</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2000-01-29</th>\n",
              "      <td>0.710167</td>\n",
              "      <td>0.466966</td>\n",
              "      <td>-0.085525</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2000-01-30</th>\n",
              "      <td>0.022575</td>\n",
              "      <td>0.640558</td>\n",
              "      <td>-0.235207</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                 y_1       y_2       y_3\n",
              "2000-01-28 -0.081028  0.820356  0.555104\n",
              "2000-01-29  0.710167  0.466966 -0.085525\n",
              "2000-01-30  0.022575  0.640558 -0.235207"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 17
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "c32J8uv3DjSS"
      },
      "source": [
        "### Thank you for your time"
      ]
    }
  ]
}