International Journal of Engineering and Technology

This paper presents a new method. It is a combination of two algorithms that have never been presented before. The research proposes the combination of the Hodrick-Prescott (HP) filters with Multiple linear regression method (MLR). The case study is the electricity demand of Thailand. The first stage, we separate the signal from the original demand signal to a trend and detail components. After that these components will be computed the correlated value with all factors before forecasting with the multivariate or multiple linear regression method in the second stage. The results of the forecasting approach were quite good when compared to other methods in the past, and this method also makes it a relatively complex inside of the electrification of the country.