Short-Term Electricity Demand Forecasting with Seasonal and Interactions of Variables for Thailand

Abstract

The quality of short term electricity demand forecasting is essential for all the energy market players for operation and trading activities. Electricity demand is significantly affected by non linear factors such as climatic condition, calendar and other seasonality have been widely reported in literature. Seasonality exists when electricity load influenced by seasonal factors such as types of days, months or variation of temperature. Therefore, seasonality is considered as one of the main challenging variable in forecasting. Many authors use multiple variables to capture all kinds of seasonal information. In this paper, we address this issue with very simple, but impressive technique on the historical load data called as interaction of variables and construct univariate multiple linear regression very carefully. This model can capture double seasonality: daily and weekly seasonal effect of electricity load and forecasting performance improve significantly. For this purpose, we select two models-model A that can capture double seasonality, and model B is constructed adding some interaction terms among variables so that it can capture more information of electricity patterns. And obviously model B perform better performance because of interaction terms included in model.