Multiscale Forecasting Method using ARMAX Models

Abstract

In this paper we propose a new forecasting methodology that comprises simultaneous level wise modeling in the wavelet domain. The WAW methodology (short for wavelet armax winters) uses three modeling strategies: ARMAX models capable of incorporating external inputs and model feedbacks, trigonometric regressions sensitive to seasonality effects and Holt Winters models describing trends. The comprehensive empirical analysis of WAW models is provided, as well as the illustration on the real life problem in forecasting the natural gas price, electricity price, and customer demand for electric business.