Using a SARIMAX time series model with pmdarima

SARIMA is an extension of the ARIMA model for univariate time series with a seasonal component.

SARIMAX is, then, the name of the model, which also supports exogenous variables.

These are the three ARIMA parameters:

• p = trend auto-regressive order
• d = trend difference order
• q = trend MA order

In addition to the preceding parameters, SARIMA introduces four more, as follows:

• P = seasonal auto-regressive order
• D = seasonal difference order
• Q = seasonal MA order
• m = the length of a single seasonal period in the number of time steps

To find these parameters manually can be time-consuming, and it may be advantageous to use an auto-ARIMA model.

In Python, auto-ARIMA modeling is provided by the pmdarima library. Its documentation is available at http://alkaline-ml.com/pmdarima/index.html.

The installation is straightforward, as can be seen here:

pip install pmdarima

The auto...