Previsão de consumo de energia elétrica na região Sudeste: um estudo de caso usando SARIMA

Abstract

The present work proposes the implementation of a computational model to forecast the electric power consumption of southeastern Brazil, using stochastic models of the SARIMA Group (Integrated Moving Average Autoregressive with the Seasonal Characteristic of the series). A series with the monthly electricity consumption from 2004 to 2018 was used in this work. The criterion for choosing the model was the root mean square error, the absolute mean error, the Akaiake criterion and the U-Theil. The chosen model was the SARIMA (1,1,1) (2,1,3). The results obtained from this model demonstrate its effectiveness for a short term horizon.