Adaptive neural network short term load forecasting with wavelet decompositions

This paper proposes a time series load forecast model suited to competitive electricity markets. The forecast model is based on wavelet multi-resolution decomposition and the neural network modeling of wavelet coefficients. A Bayesian method automatic relevance determination (ARD) model is used to choose the optimal neural network size. The individual wavelet domain neural network forecasts are recombined to form the accurate overall forecast. The proposed method is tested using Queensland electricity demand data from the Australian National Electricity Market.