The SARIMA model, as an extension of the ARIMA model which
is the common linear approach for predicting future time-series can improve the prediction accuracy by removing the characteristics
of seasonal variation through seasonal differences. For example,
electricity consumption in educational facilities is closely related to
heating and cooling loads, which have seasonal characteristics in
one-year interval. In this case, the SARIMA model can be implemented
to remove the characteristics of seasonal variation, which
can improve the prediction accuracy of future electricity consumption
in educational facilities.