Short-term electricity price modeling and forecasting using wavelets and multivariate time series

Abstract

This work presents a new method to model and forecast the short-term electricity prices. The historical price and load data are first decomposed by wavelet transform, then multivariate time series is applied to model and forecast the wavelet coefficients of next day electricity price. The forecasted price is obtained by reconstructing the wavelet coefficients. The numerical examples of Pennsylvania-New Jersey-Maryland (PJM) spot market data are presented.