Overview for ARIMA

Use ARIMA to specify a model for time series data that can contain autoregressive, differencing, and moving average components. You can use the model to generate forecasts.

ARIMA (autoregressive integrated moving average) fits a Box-Jenkins ARIMA model to a time series. Each term in an ARIMA model represents steps that are taken in the model construction until only random noise remains. Unlike other time series methods, ARIMA modeling uses correlational techniques. You can use ARIMA to model patterns that may not be visible in plotted data.

For example, a budget planner for a business office uses an ARIMA model to predict water and electricity costs for the next three periods.