time_series_handbook/01_AutoRegressiveIntegratedMovingAverage

Chapter 1: AutoRegressive Integrated Moving Average (ARIMA)

In this notebook, we will introduce our first approach to time-series forecasting which is ARIMA or AutoRegressive Integrated Moving Average. ARIMA, or AutoRegressive Integrated Moving Average, is a set of models that explains a time series using its own previous values given by the lags (AutoRegressive) and lagged errors (Moving Average) while considering stationarity corrected by differencing (oppossite of Integration.) In other words, ARIMA assumes that the time series is described by autocorrelations in the data rather than trends and seasonality.

This notebook will discuss:

1. Definition and Formulation of ARIMA models

2. Model Parameters (p, d, and q) and Special Cases of ARIMA models

3. Model Statistics and How to Interpret

4. Implementation and Forecasting using ARIMA

Datasets used:

• Synthetic Data (Filename: ../data/wwwusage.csv)

• Climate Data (Filename: ../data/jena_climate_2009_2016.csv")