2014 : Forecasting Electricity Demands in Java-Bali Region using Flexible Forecasting Approach

Ir. Mutiah Salamah Chamid


Abstract

This paper discusses a modeling approach for electricity consumption in East-Java Bali Region namely flexible forecasting approach. This approach offers a flexibility to cover any complexities in the seasonality pattern observed in the electricity dataset. One of the very obvious facts in the considered dataset is the presence of double seasonality. The flexible forecasting approach introduces a new setting of class of innovation state space models which is obtained by replacing the seasonal components in the model by the trigonometric seasonal formulation. The approach is also designed to automatically overcome a problem of nonstationarity in the variance term, by including Box-Cox transformation in the model. This paper proved that the model is capable to generate a good performance in modeling the electricity consumption in East-Java Bali.