ajaytiwari-isb/Time_Series_Analysis,_Modeling_and_Validation.ipynb

## Time_Series_Analysis,_Modeling_and_Validation.ipynb
{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Time Series Analysis, Modeling and Validation",
      "provenance": [],
      "collapsed_sections": [],
      "toc_visible": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "yF7PsICquCBN",
        "colab_type": "text"
      },
      "source": [
        "# A Project on Time Series Analysis, Modeling and Validation\n",
        "\n",
        "\n",
        "1.   Data Preparation & Analysis\n",
        "2.   A Naive Model as a Benchmark\n",
        "3.   ARIMA Model on Original Series\n",
        "4.   ARIMA Model on Differenced Series\n",
        "5.   Grid Search ARIMA Parameters\n",
        "6.   Residual Analysis\n",
        "7.   Bias Corrected Model\n",
        "7.   Saving the Model File\n",
        "2.   Model Validation\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "GskWJCBGt4xH",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "#Disable warnings\n",
        "import warnings\n",
        "warnings.filterwarnings('ignore')\n",
        "warnings.simplefilter('ignore')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8ROwxtwrLi25",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Import required libraries\n",
        "import numpy\n",
        "from pandas import read_csv\n",
        "from sklearn.metrics import mean_squared_error\n",
        "from math import sqrt\n",
        "from math import log\n",
        "from math import exp\n",
        "from scipy.stats import boxcox\n",
        "from pandas import DataFrame\n",
        "from pandas import Grouper\n",
        "from pandas import Series\n",
        "from pandas import concat\n",
        "from pandas.plotting import lag_plot\n",
        "from matplotlib import pyplot\n",
        "from statsmodels.tsa.stattools import adfuller\n",
        "from statsmodels.tsa.arima_model import ARIMA\n",
        "from statsmodels.tsa.arima_model import ARIMAResults\n",
        "from statsmodels.tsa.seasonal import seasonal_decompose\n",
        "from statsmodels.graphics.tsaplots import plot_acf\n",
        "from statsmodels.graphics.tsaplots import plot_pacf\n",
        "from statsmodels.graphics.gofplots import qqplot\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "XnEzuppEj1LU",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 136
        },
        "outputId": "0de651bb-6bf5-4f43-bc71-598e36d8e1e7"
      },
      "source": [
        "# Import data\n",
        "series = read_csv('airline-passengers.csv', header=0, index_col=0, parse_dates=True, squeeze=True)\n",
        "print(series.head())\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Month\n",
            "1949-01-01    112\n",
            "1949-02-01    118\n",
            "1949-03-01    132\n",
            "1949-04-01    129\n",
            "1949-05-01    121\n",
            "Name: Passengers, dtype: int64\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "gdn7MNIxMxi7",
        "colab_type": "text"
      },
      "source": [
        "# Data Preparation & Analysis"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "eL3ZlaQztwsF",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 170
        },
        "outputId": "3a9c6f9f-59b5-4e24-9b63-258722aef07f"
      },
      "source": [
        "# summary statistics of time series\n",
        "\n",
        "print(series.describe())"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "count    144.000000\n",
            "mean     280.298611\n",
            "std      119.966317\n",
            "min      104.000000\n",
            "25%      180.000000\n",
            "50%      265.500000\n",
            "75%      360.500000\n",
            "max      622.000000\n",
            "Name: Passengers, dtype: float64\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4QcSZNyht7BC",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 279
        },
        "outputId": "075affb8-42b4-461f-bf9a-24e7cb86b740"
      },
      "source": [
        "# line plot of time series\n",
        "\n",
        "series.plot()\n",
        "pyplot.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "S-Iy6iXCuB65",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 269
        },
        "outputId": "f9800345-ae1b-4eb5-c359-5d2435e56ba3"
      },
      "source": [
        "# multiple line plots of time series\n",
        "\n",
        "groups = series['1949':'1957'].groupby(Grouper(freq='A'))\n",
        "years = DataFrame()\n",
        "pyplot.figure()\n",
        "i = 1\n",
        "n_groups = len(groups)\n",
        "for name, group in groups:\n",
        "  pyplot.subplot((n_groups*100) + 10 + i)\n",
        "  i += 1\n",
        "  pyplot.plot(group)\n",
        "pyplot.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 9 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7MkUB_oMEEli",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 265
        },
        "outputId": "80105e5b-6290-4f5d-cdba-6b97e31cd218"
      },
      "source": [
        "# density plots of time series\n",
        "pyplot.figure(1)\n",
        "pyplot.subplot(211)\n",
        "series.hist()\n",
        "pyplot.subplot(212)\n",
        "series.plot(kind='kde')\n",
        "pyplot.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Ws71DCFxEUsJ",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 265
        },
        "outputId": "aea53961-635d-4eb3-df2c-d4e9a1a3b940"
      },
      "source": [
        "# boxplots of time series\n",
        "\n",
        "groups = series['1949':'1960'].groupby(Grouper(freq='A'))\n",
        "years = DataFrame()\n",
        "for name, group in groups:\n",
        "  years[name.year] = group.values\n",
        "years.boxplot()\n",
        "pyplot.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "dsatVwLhdrDf",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 279
        },
        "outputId": "1c9ff796-0d0e-4ebe-ad86-781a8e3359c7"
      },
      "source": [
        "# create a scatter plot\n",
        "\n",
        "lag_plot(series)\n",
        "pyplot.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "N8J179g2ehQd",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 163
        },
        "outputId": "944f071f-dda9-4d89-9682-513f857c56b5"
      },
      "source": [
        "\n",
        "values = DataFrame(series.values)\n",
        "lags = 4\n",
        "columns = [values]\n",
        "for i in range(1,(lags + 1)):\n",
        "  columns.append(values.shift(i))\n",
        "dataframe = concat(columns, axis=1)\n",
        "columns = ['t']\n",
        "for i in range(1,(lags + 1)):\n",
        "  columns.append('t-' + str(i))\n",
        "dataframe.columns = columns\n",
        "pyplot.figure(1)\n",
        "for i in range(1,(lags + 1)):\n",
        "  ax = pyplot.subplot(240 + i)\n",
        "  ax.set_title('t vs t-' + str(i))\n",
        "  pyplot.scatter(x=dataframe['t'].values, y=dataframe['t-'+str(i)].values)\n",
        "pyplot.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 4 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "oVQ7fyDcg0cR",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 298
        },
        "outputId": "9c24d508-0918-4921-dfcd-39a2443057e8"
      },
      "source": [
        "result = seasonal_decompose(series, model='multiplicative')\n",
        "result.plot()\n",
        "pyplot.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 4 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Z7J4uvKLjRLe",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "2a5c6268-89f5-45d1-e34a-80eb635f79ff"
      },
      "source": [
        "# separate out a validation dataset\n",
        "series = read_csv('airline-passengers.csv', header=0, index_col=0, parse_dates=True, squeeze=True)\n",
        "split_point = len(series) - 12\n",
        "dataset, validation = series[0:split_point], series[split_point:]\n",
        "print('Train-Dataset: %d, Validation-Dataset: %d' % (len(dataset), len(validation)))\n",
        "dataset.to_csv('dataset.csv', header=False)\n",
        "validation.to_csv('validation.csv', header=False)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Train-Dataset: 132, Validation-Dataset: 12\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "jZXfHcihzNQ_",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# load and prepare data\n",
        "series = read_csv('dataset.csv', header=None, index_col=0, parse_dates=True, squeeze=True)\n",
        "X = series.values\n",
        "X = X.astype('float32')\n",
        "train_size = int(len(X) * 0.5)\n",
        "train, test = X[0:train_size], X[train_size:]"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "s1aF0xdmnzY2",
        "colab_type": "text"
      },
      "source": [
        "# A Naive Model as a Benchmark"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "BvxF8E1jrefP",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "77bc9201-685a-4c4b-f4d8-f534a32bdae1"
      },
      "source": [
        "# predict and walk forward validation\n",
        "history = [x for x in train]\n",
        "predictions = list()\n",
        "for i in range(len(test)):\n",
        "  yhat = history[-1]\n",
        "  predictions.append(yhat)\n",
        "  obs = test[i]\n",
        "  history.append(obs)\n",
        "  print('>Forecast=%.3f, Actual=%3.f' % (yhat, obs))\n",
        "# display performance report\n",
        "rmse = sqrt(mean_squared_error(test, predictions))\n",
        "print('RMSE: %.3f' % rmse)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            ">Forecast=264.000, Actual=302\n",
            ">Forecast=302.000, Actual=293\n",
            ">Forecast=293.000, Actual=259\n",
            ">Forecast=259.000, Actual=229\n",
            ">Forecast=229.000, Actual=203\n",
            ">Forecast=203.000, Actual=229\n",
            ">Forecast=229.000, Actual=242\n",
            ">Forecast=242.000, Actual=233\n",
            ">Forecast=233.000, Actual=267\n",
            ">Forecast=267.000, Actual=269\n",
            ">Forecast=269.000, Actual=270\n",
            ">Forecast=270.000, Actual=315\n",
            ">Forecast=315.000, Actual=364\n",
            ">Forecast=364.000, Actual=347\n",
            ">Forecast=347.000, Actual=312\n",
            ">Forecast=312.000, Actual=274\n",
            ">Forecast=274.000, Actual=237\n",
            ">Forecast=237.000, Actual=278\n",
            ">Forecast=278.000, Actual=284\n",
            ">Forecast=284.000, Actual=277\n",
            ">Forecast=277.000, Actual=317\n",
            ">Forecast=317.000, Actual=313\n",
            ">Forecast=313.000, Actual=318\n",
            ">Forecast=318.000, Actual=374\n",
            ">Forecast=374.000, Actual=413\n",
            ">Forecast=413.000, Actual=405\n",
            ">Forecast=405.000, Actual=355\n",
            ">Forecast=355.000, Actual=306\n",
            ">Forecast=306.000, Actual=271\n",
            ">Forecast=271.000, Actual=306\n",
            ">Forecast=306.000, Actual=315\n",
            ">Forecast=315.000, Actual=301\n",
            ">Forecast=301.000, Actual=356\n",
            ">Forecast=356.000, Actual=348\n",
            ">Forecast=348.000, Actual=355\n",
            ">Forecast=355.000, Actual=422\n",
            ">Forecast=422.000, Actual=465\n",
            ">Forecast=465.000, Actual=467\n",
            ">Forecast=467.000, Actual=404\n",
            ">Forecast=404.000, Actual=347\n",
            ">Forecast=347.000, Actual=305\n",
            ">Forecast=305.000, Actual=336\n",
            ">Forecast=336.000, Actual=340\n",
            ">Forecast=340.000, Actual=318\n",
            ">Forecast=318.000, Actual=362\n",
            ">Forecast=362.000, Actual=348\n",
            ">Forecast=348.000, Actual=363\n",
            ">Forecast=363.000, Actual=435\n",
            ">Forecast=435.000, Actual=491\n",
            ">Forecast=491.000, Actual=505\n",
            ">Forecast=505.000, Actual=404\n",
            ">Forecast=404.000, Actual=359\n",
            ">Forecast=359.000, Actual=310\n",
            ">Forecast=310.000, Actual=337\n",
            ">Forecast=337.000, Actual=360\n",
            ">Forecast=360.000, Actual=342\n",
            ">Forecast=342.000, Actual=406\n",
            ">Forecast=406.000, Actual=396\n",
            ">Forecast=396.000, Actual=420\n",
            ">Forecast=420.000, Actual=472\n",
            ">Forecast=472.000, Actual=548\n",
            ">Forecast=548.000, Actual=559\n",
            ">Forecast=559.000, Actual=463\n",
            ">Forecast=463.000, Actual=407\n",
            ">Forecast=407.000, Actual=362\n",
            ">Forecast=362.000, Actual=405\n",
            "RMSE: 40.350\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ywWyjrwww-sa",
        "colab_type": "text"
      },
      "source": [
        "# ARIMA Model on Original Series"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "lovINnIoH9Lc",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 281
        },
        "outputId": "2618f4a6-97ec-433c-9de7-f035c80829eb"
      },
      "source": [
        "# ACF and PACF plots of time series\n",
        "pyplot.figure()\n",
        "pyplot.subplot(211)\n",
        "plot_acf(dataset, lags=25, ax=pyplot.gca())\n",
        "pyplot.subplot(212)\n",
        "plot_pacf(dataset, lags=25, ax=pyplot.gca())\n",
        "pyplot.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "38a5Bx-Q_-5H",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "05c02c8e-1627-453c-b9ff-693326732263"
      },
      "source": [
        "# evaluate manually configured ARIMA model\n",
        "\n",
        "# predict and walk-forward validation\n",
        "history = [x for x in train]\n",
        "predictions = list()\n",
        "for i in range(len(test)):\n",
        "  model = ARIMA(history, order=(7,1,1))\n",
        "  model_fit = model.fit(disp=0)\n",
        "  yhat = model_fit.forecast()[0]\n",
        "  predictions.append(yhat)\n",
        "  obs = test[i]\n",
        "  history.append(obs)\n",
        "  print('>Forecast=%.3f, Actual=%.3f' % (yhat, obs))\n",
        "# display performance report\n",
        "rmse = sqrt(mean_squared_error(test, predictions))\n",
        "print('RMSE: %.3f' % rmse)\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            ">Forecast=263.821, Actual=302.000\n",
            ">Forecast=299.381, Actual=293.000\n",
            ">Forecast=285.112, Actual=259.000\n",
            ">Forecast=240.699, Actual=229.000\n",
            ">Forecast=226.367, Actual=203.000\n",
            ">Forecast=217.451, Actual=229.000\n",
            ">Forecast=246.861, Actual=242.000\n",
            ">Forecast=246.165, Actual=233.000\n",
            ">Forecast=241.847, Actual=267.000\n",
            ">Forecast=277.033, Actual=269.000\n",
            ">Forecast=280.270, Actual=270.000\n",
            ">Forecast=288.198, Actual=315.000\n",
            ">Forecast=305.473, Actual=364.000\n",
            ">Forecast=336.812, Actual=347.000\n",
            ">Forecast=311.164, Actual=312.000\n",
            ">Forecast=270.281, Actual=274.000\n",
            ">Forecast=245.273, Actual=237.000\n",
            ">Forecast=235.710, Actual=278.000\n",
            ">Forecast=283.435, Actual=284.000\n",
            ">Forecast=264.323, Actual=277.000\n",
            ">Forecast=272.016, Actual=317.000\n",
            ">Forecast=317.822, Actual=313.000\n",
            ">Forecast=317.015, Actual=318.000\n",
            ">Forecast=339.004, Actual=374.000\n",
            ">Forecast=360.040, Actual=413.000\n",
            ">Forecast=389.182, Actual=405.000\n",
            ">Forecast=378.309, Actual=355.000\n",
            ">Forecast=306.473, Actual=306.000\n",
            ">Forecast=286.561, Actual=271.000\n",
            ">Forecast=279.485, Actual=306.000\n",
            ">Forecast=327.011, Actual=315.000\n",
            ">Forecast=313.363, Actual=301.000\n",
            ">Forecast=308.040, Actual=356.000\n",
            ">Forecast=377.171, Actual=348.000\n",
            ">Forecast=361.782, Actual=355.000\n",
            ">Forecast=393.266, Actual=422.000\n",
            ">Forecast=410.961, Actual=465.000\n",
            ">Forecast=439.835, Actual=467.000\n",
            ">Forecast=441.199, Actual=404.000\n",
            ">Forecast=339.116, Actual=347.000\n",
            ">Forecast=324.797, Actual=305.000\n",
            ">Forecast=309.202, Actual=336.000\n",
            ">Forecast=358.655, Actual=340.000\n",
            ">Forecast=339.879, Actual=318.000\n",
            ">Forecast=327.136, Actual=362.000\n",
            ">Forecast=394.922, Actual=348.000\n",
            ">Forecast=380.257, Actual=363.000\n",
            ">Forecast=422.934, Actual=435.000\n",
            ">Forecast=445.130, Actual=491.000\n",
            ">Forecast=484.670, Actual=505.000\n",
            ">Forecast=483.656, Actual=404.000\n",
            ">Forecast=341.526, Actual=359.000\n",
            ">Forecast=359.556, Actual=310.000\n",
            ">Forecast=321.141, Actual=337.000\n",
            ">Forecast=377.573, Actual=360.000\n",
            ">Forecast=366.061, Actual=342.000\n",
            ">Forecast=349.273, Actual=406.000\n",
            ">Forecast=443.287, Actual=396.000\n",
            ">Forecast=422.551, Actual=420.000\n",
            ">Forecast=465.938, Actual=472.000\n",
            ">Forecast=463.322, Actual=548.000\n",
            ">Forecast=531.752, Actual=559.000\n",
            ">Forecast=509.678, Actual=463.000\n",
            ">Forecast=386.655, Actual=407.000\n",
            ">Forecast=372.811, Actual=362.000\n",
            ">Forecast=358.454, Actual=405.000\n",
            "RMSE: 31.228\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "DK5LZ91cOSyM",
        "colab_type": "text"
      },
      "source": [
        "# ARIMA Model on Differenced Series"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "WvmqcpFI11nl",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# create a differenced series\n",
        "def difference(dataset, interval=1):\n",
        "  diff = list()\n",
        "  for i in range(interval, len(dataset)):\n",
        "    value = dataset[i] - dataset[i - interval]\n",
        "    diff.append(value)\n",
        "  return Series(diff)\n",
        "X = series.values\n",
        "X = X.astype('float32')\n",
        "\n",
        "# differenced data\n",
        "months_in_year = 12\n",
        "stationary = difference(X, months_in_year)\n",
        "stationary.index = series.index[months_in_year:]"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Rctv7b7MGEJP",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 119
        },
        "outputId": "cc7db022-4d60-4a3e-9475-74c671df3d92"
      },
      "source": [
        "# check if stationary\n",
        "result = adfuller(stationary)\n",
        "print('ADF Statistic: %f' % result[0])\n",
        "print('p-value: %f' % result[1])\n",
        "print('Critical Values:')\n",
        "for key, value in result[4].items():\n",
        "  print('\\t%s: %.3f' % (key, value))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "ADF Statistic: -3.048011\n",
            "p-value: 0.030648\n",
            "Critical Values:\n",
            "\t1%: -3.488\n",
            "\t5%: -2.887\n",
            "\t10%: -2.580\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "sMC-yvdz2Cst",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 279
        },
        "outputId": "eabb224b-c077-4a3f-8677-2ca980cfe579"
      },
      "source": [
        "# plot stationary data\n",
        "stationary.plot()\n",
        "pyplot.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "JGk8UrR8Hmqk",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 281
        },
        "outputId": "d5cba1bd-ec12-470e-ae41-26b439a24f53"
      },
      "source": [
        "# ACF and PACF plots of time series\n",
        "\n",
        "pyplot.figure()\n",
        "pyplot.subplot(211)\n",
        "plot_acf(stationary, lags=25, ax=pyplot.gca())\n",
        "pyplot.subplot(212)\n",
        "plot_pacf(stationary, lags=25, ax=pyplot.gca())\n",
        "pyplot.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "LLz02T7g2gV-",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# invert differenced value\n",
        "def inverse_difference(history, yhat, interval=1):\n",
        "  return yhat + history[-interval]"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "NINfuUc1ItUZ",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "609d1b18-5adf-4578-c554-a34e1c8b509a"
      },
      "source": [
        "# predict and walk-forward validation\n",
        "history = [x for x in train]\n",
        "predictions = list()\n",
        "for i in range(len(test)):\n",
        "  # difference data\n",
        "  months_in_year = 12\n",
        "  diff = difference(history, months_in_year)\n",
        "  # predict\n",
        "  model = ARIMA(diff, order=(1,1,1))\n",
        "  model_fit = model.fit(trend='nc', disp=0)\n",
        "  yhat = model_fit.forecast()[0]\n",
        "  yhat = inverse_difference(history, yhat, months_in_year)\n",
        "  predictions.append(yhat)\n",
        "  # observation\n",
        "  obs = test[i]\n",
        "  history.append(obs)\n",
        "  print('>Forecast=%.3f, Actual=%.3f' % (yhat, obs))\n",
        "# report performance\n",
        "rmse = sqrt(mean_squared_error(test, predictions))\n",
        "print('RMSE: %.3f' % rmse)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            ">Forecast=283.486, Actual=302.000\n",
            ">Forecast=307.082, Actual=293.000\n",
            ">Forecast=263.328, Actual=259.000\n",
            ">Forecast=229.844, Actual=229.000\n",
            ">Forecast=200.635, Actual=203.000\n",
            ">Forecast=221.436, Actual=229.000\n",
            ">Forecast=232.404, Actual=242.000\n",
            ">Forecast=223.745, Actual=233.000\n",
            ">Forecast=279.973, Actual=267.000\n",
            ">Forecast=261.614, Actual=269.000\n",
            ">Forecast=272.398, Actual=270.000\n",
            ">Forecast=303.246, Actual=315.000\n",
            ">Forecast=347.806, Actual=364.000\n",
            ">Forecast=355.372, Actual=347.000\n",
            ">Forecast=314.549, Actual=312.000\n",
            ">Forecast=281.417, Actual=274.000\n",
            ">Forecast=249.932, Actual=237.000\n",
            ">Forecast=264.128, Actual=278.000\n",
            ">Forecast=287.307, Actual=284.000\n",
            ">Forecast=278.358, Actual=277.000\n",
            ">Forecast=308.929, Actual=317.000\n",
            ">Forecast=318.691, Actual=313.000\n",
            ">Forecast=315.472, Actual=318.000\n",
            ">Forecast=361.466, Actual=374.000\n",
            ">Forecast=421.387, Actual=413.000\n",
            ">Forecast=398.945, Actual=405.000\n",
            ">Forecast=366.705, Actual=355.000\n",
            ">Forecast=322.047, Actual=306.000\n",
            ">Forecast=269.585, Actual=271.000\n",
            ">Forecast=311.297, Actual=306.000\n",
            ">Forecast=313.624, Actual=315.000\n",
            ">Forecast=306.721, Actual=301.000\n",
            ">Forecast=343.078, Actual=356.000\n",
            ">Forecast=347.642, Actual=348.000\n",
            ">Forecast=355.588, Actual=355.000\n",
            ">Forecast=409.559, Actual=422.000\n",
            ">Forecast=458.779, Actual=465.000\n",
            ">Forecast=456.819, Actual=467.000\n",
            ">Forecast=414.803, Actual=404.000\n",
            ">Forecast=358.897, Actual=347.000\n",
            ">Forecast=312.443, Actual=305.000\n",
            ">Forecast=341.389, Actual=336.000\n",
            ">Forecast=345.302, Actual=340.000\n",
            ">Forecast=326.934, Actual=318.000\n",
            ">Forecast=374.273, Actual=362.000\n",
            ">Forecast=355.655, Actual=348.000\n",
            ">Forecast=355.436, Actual=363.000\n",
            ">Forecast=428.251, Actual=435.000\n",
            ">Forecast=477.812, Actual=491.000\n",
            ">Forecast=490.673, Actual=505.000\n",
            ">Forecast=440.984, Actual=404.000\n",
            ">Forecast=355.359, Actual=359.000\n",
            ">Forecast=311.384, Actual=310.000\n",
            ">Forecast=344.517, Actual=337.000\n",
            ">Forecast=340.947, Actual=360.000\n",
            ">Forecast=333.495, Actual=342.000\n",
            ">Forecast=386.062, Actual=406.000\n",
            ">Forecast=387.688, Actual=396.000\n",
            ">Forecast=411.446, Actual=420.000\n",
            ">Forecast=490.040, Actual=472.000\n",
            ">Forecast=532.834, Actual=548.000\n",
            ">Forecast=555.661, Actual=559.000\n",
            ">Forecast=460.737, Actual=463.000\n",
            ">Forecast=415.918, Actual=407.000\n",
            ">Forecast=361.278, Actual=362.000\n",
            ">Forecast=386.891, Actual=405.000\n",
            "RMSE: 10.845\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "lf0Qa4INxm_g",
        "colab_type": "text"
      },
      "source": [
        "# Grid Search ARIMA Parameters"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "baNzXpKYJcEY",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 329
        },
        "outputId": "865a32af-c4a3-43e2-8a64-0d81c8457357"
      },
      "source": [
        "# # grid search ARIMA parameters for time series\n",
        "\n",
        "# # evaluate an ARIMA model for a given order (p,d,q) and return RMSE\n",
        "# def evaluate_arima_model(X, arima_order):\n",
        "#   # prepare training dataset\n",
        "#   X = X.astype('float32')\n",
        "#   train_size = int(len(X) * 0.5)\n",
        "#   train, test = X[0:train_size], X[train_size:]\n",
        "#   history = [x for x in train]\n",
        "#   # make predictions\n",
        "#   predictions = list()\n",
        "#   for t in range(len(test)):\n",
        "#     # difference data\n",
        "#     months_in_year = 12\n",
        "#     diff = difference(history, months_in_year)\n",
        "#     model = ARIMA(diff, order=arima_order)\n",
        "#     model_fit = model.fit(trend='nc', disp=0)\n",
        "#     yhat = model_fit.forecast()[0]\n",
        "#     yhat = inverse_difference(history, yhat, months_in_year)\n",
        "#     predictions.append(yhat)\n",
        "#     history.append(test[t])\n",
        "#   # calculate out of sample error\n",
        "#   rmse = sqrt(mean_squared_error(test, predictions))\n",
        "#   return rmse\n",
        "# # evaluate combinations of p, d and q values for an ARIMA model\n",
        "# def evaluate_models(dataset, p_values, d_values, q_values):\n",
        "#   dataset = dataset.astype('float32')\n",
        "#   best_score, best_cfg = float(\"inf\"), None\n",
        "#   for p in p_values:\n",
        "#     for d in d_values:\n",
        "#       for q in q_values:\n",
        "#         order = (p,d,q)\n",
        "#         try:\n",
        "#           rmse = evaluate_arima_model(dataset, order)\n",
        "#           if rmse < best_score:\n",
        "#             best_score, best_param = rmse, order\n",
        "#           print('ARIMA%s RMSE=%.3f' % (order,rmse))\n",
        "#         except:\n",
        "#           continue\n",
        "#   print('Best ARIMA%s RMSE=%.3f' % (best_param, best_score))\n",
        "# # load dataset\n",
        "# series = read_csv('dataset.csv', header=None, index_col=0, parse_dates=True, squeeze=True)\n",
        "# # evaluate parameters\n",
        "# p_values = range(0, 5)\n",
        "# d_values = range(0, 2)\n",
        "# q_values = range(0, 2)\n",
        "# warnings.filterwarnings(\"ignore\")\n",
        "# evaluate_models(series.values, p_values, d_values, q_values)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "ARIMA(0, 0, 1) RMSE=25.203\n",
            "ARIMA(0, 1, 1) RMSE=10.977\n",
            "ARIMA(1, 0, 0) RMSE=11.172\n",
            "ARIMA(1, 0, 1) RMSE=11.115\n",
            "ARIMA(1, 1, 0) RMSE=10.845\n",
            "ARIMA(1, 1, 1) RMSE=10.845\n",
            "ARIMA(2, 0, 0) RMSE=11.011\n",
            "ARIMA(2, 0, 1) RMSE=11.011\n",
            "ARIMA(2, 1, 0) RMSE=10.897\n",
            "ARIMA(2, 1, 1) RMSE=10.925\n",
            "ARIMA(3, 0, 0) RMSE=11.043\n",
            "ARIMA(3, 0, 1) RMSE=11.079\n",
            "ARIMA(3, 1, 0) RMSE=11.077\n",
            "ARIMA(3, 1, 1) RMSE=10.993\n",
            "ARIMA(4, 0, 0) RMSE=11.208\n",
            "ARIMA(4, 1, 0) RMSE=10.968\n",
            "ARIMA(4, 1, 1) RMSE=11.039\n",
            "Best ARIMA(1, 1, 1) RMSE=10.845\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "6zYrhSu0hDQx",
        "colab_type": "text"
      },
      "source": [
        "#Residual Analysis"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "69l_zs8dNM9I",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 437
        },
        "outputId": "6bd21d9c-df7d-4ebb-a353-cff5f143adf1"
      },
      "source": [
        "# errors\n",
        "residuals = [test[i]-predictions[i] for i in range(len(test))]\n",
        "residuals = DataFrame(residuals)\n",
        "print(residuals.describe())\n",
        "# plot\n",
        "pyplot.figure()\n",
        "pyplot.subplot(211)\n",
        "residuals.hist(ax=pyplot.gca())\n",
        "pyplot.subplot(212)\n",
        "residuals.plot(kind='kde', ax=pyplot.gca())\n",
        "pyplot.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "               0\n",
            "count  66.000000\n",
            "mean    0.810541\n",
            "std    10.897172\n",
            "min   -36.983527\n",
            "25%    -7.436698\n",
            "50%     1.049093\n",
            "75%     8.541763\n",
            "max    19.937876\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "oHzaiRhmx7OS",
        "colab_type": "text"
      },
      "source": [
        "# Bias Corrected Model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "F6gVRW68ORjA",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 451
        },
        "outputId": "8b90025c-64fe-4f1e-cf0f-40fa89637851"
      },
      "source": [
        "# plots of residual errors of bias corrected forecasts\n",
        "# walk-forward validation\n",
        "history = [x for x in train]\n",
        "predictions = list()\n",
        "bias = 0.810541\n",
        "for i in range(len(test)):\n",
        "  # difference data\n",
        "  months_in_year = 12\n",
        "  diff = difference(history, months_in_year)\n",
        "  # predict\n",
        "  model = ARIMA(diff, order=(1,1,1))\n",
        "  model_fit = model.fit(trend='nc', disp=0)\n",
        "  yhat = model_fit.forecast()[0]\n",
        "  yhat = bias + inverse_difference(history, yhat, months_in_year)\n",
        "  predictions.append(yhat)\n",
        "  # observation\n",
        "  obs = test[i]\n",
        "  history.append(obs)\n",
        "# report performance\n",
        "rmse = sqrt(mean_squared_error(test, predictions))\n",
        "print('RMSE: %.3f' % rmse)\n",
        "# errors\n",
        "residuals = [test[i]-predictions[i] for i in range(len(test))]\n",
        "residuals = DataFrame(residuals)\n",
        "print(residuals.describe())\n",
        "# plot\n",
        "pyplot.figure()\n",
        "pyplot.subplot(211)\n",
        "residuals.hist(ax=pyplot.gca())\n",
        "pyplot.subplot(212)\n",
        "residuals.plot(kind='kde', ax=pyplot.gca())\n",
        "pyplot.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "RMSE: 10.814\n",
            "                  0\n",
            "count  6.600000e+01\n",
            "mean  -1.394366e-08\n",
            "std    1.089717e+01\n",
            "min   -3.779407e+01\n",
            "25%   -8.247239e+00\n",
            "50%    2.385517e-01\n",
            "75%    7.731222e+00\n",
            "max    1.912734e+01\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fO0tBQBCP62A",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 281
        },
        "outputId": "7e154c3c-6247-4c06-8170-204802a5a73e"
      },
      "source": [
        "pyplot.figure()\n",
        "pyplot.subplot(211)\n",
        "plot_acf(residuals, ax=pyplot.gca())\n",
        "pyplot.subplot(212)\n",
        "plot_pacf(residuals, ax=pyplot.gca())\n",
        "pyplot.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "PcviVxZLyDuh",
        "colab_type": "text"
      },
      "source": [
        "# Saving the Model File"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "gtZTVNqdnQXH",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# save finalized model\n",
        "# bias constant, could be calculated from in-sample mean residual\n",
        "bias = 0.810541\n",
        "# save model\n",
        "model_fit.save('model.pkl')\n",
        "numpy.save('model_bias.npy', [bias])"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "BpKvTrDMSX82",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "220103fd-8b49-4661-d96b-c52c20c232ce"
      },
      "source": [
        "# load finalized model and make a prediction\n",
        "\n",
        "series = read_csv('dataset.csv', header=None, index_col=0, parse_dates=True, squeeze=True)\n",
        "months_in_year = 12\n",
        "model_fit = ARIMAResults.load('model.pkl')\n",
        "bias = numpy.load('model_bias.npy')\n",
        "yhat = float(model_fit.forecast()[0])\n",
        "yhat = bias + inverse_difference(series.values, yhat, months_in_year)\n",
        "print('Forecast: %.3f' % yhat)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Forecast: 410.702\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "URvBhfYNyOAd",
        "colab_type": "text"
      },
      "source": [
        "# Model Validation"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8_L7IhlpSnDN",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 486
        },
        "outputId": "4b1c8301-4d8f-4efc-e2a2-a6720d9b9632"
      },
      "source": [
        "# load and evaluate the finalized model on the validation dataset\n",
        "\n",
        "validation = read_csv('validation.csv', header=None, index_col=0, parse_dates=True,\n",
        "squeeze=True)\n",
        "y = validation.values.astype('float32')\n",
        "# load model\n",
        "model_fit = ARIMAResults.load('model.pkl')\n",
        "bias = numpy.load('model_bias.npy')\n",
        "# make first prediction\n",
        "predictions = list()\n",
        "yhat = float(model_fit.forecast()[0])\n",
        "yhat = bias + inverse_difference(history, yhat, months_in_year)\n",
        "predictions.append(yhat)\n",
        "history.append(y[0])\n",
        "print('>Forecast=%.3f, Actual=%.3f' % (yhat, y[0]))\n",
        "# rolling forecasts\n",
        "for i in range(1, len(y)):\n",
        "  # difference data\n",
        "  months_in_year = 12\n",
        "  diff = difference(history, months_in_year)\n",
        "  # predict\n",
        "  model = ARIMA(diff, order=(1,1,1))\n",
        "  model_fit = model.fit(trend='nc', disp=0)\n",
        "  yhat = model_fit.forecast()[0]\n",
        "  yhat = bias + inverse_difference(history, yhat, months_in_year)\n",
        "  predictions.append(yhat)\n",
        "  # observation\n",
        "  obs = y[i]\n",
        "  history.append(obs)\n",
        "  print('>Forecast=%.3f, Actual=%.3f' % (yhat, obs))\n",
        "# report performance\n",
        "rmse = sqrt(mean_squared_error(y, predictions))\n",
        "print('RMSE: %.3f' % rmse)\n",
        "pyplot.plot(y)\n",
        "pyplot.plot(predictions, color='red')\n",
        "pyplot.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            ">Forecast=410.702, Actual=417.000\n",
            ">Forecast=403.304, Actual=391.000\n",
            ">Forecast=456.560, Actual=419.000\n",
            ">Forecast=416.966, Actual=461.000\n",
            ">Forecast=467.411, Actual=472.000\n",
            ">Forecast=533.517, Actual=535.000\n",
            ">Forecast=605.938, Actual=622.000\n",
            ">Forecast=631.672, Actual=606.000\n",
            ">Forecast=519.549, Actual=508.000\n",
            ">Forecast=452.531, Actual=461.000\n",
            ">Forecast=414.033, Actual=390.000\n",
            ">Forecast=442.122, Actual=432.000\n",
            "RMSE: 21.147\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    }
  ]
}