SHORT-TERM LOAD FORECASTING USING A BAYESIAN COMBINATION METHOD

This paper presents the Bayesian Combined Predictor (BCP), a probabili stically motivated predictor for Short Term Load Forecasting (STLF) ba sed on the combination of an artificial neural network (ANN) predictor and two linear regression (LR) predictors. The method is applied to S TLF for the Creek Public Power Corporation dispatching center of the i sland of Crete, using 1994 data, and daily load profiles are obtained. Statistical analysis of prediction errors reveals that during given t ime periods the ANN predictor consistently produces better forecasts f or certain hours of the day, while the LR predictor produces better fo recasts for the rest. This relative prediction advantage may change ov er different time intervals. The combined prediction is a weighted sum of the ANN and LR predictions, where the weights are computed using a n adaptive update of the Bayesian poster for probability of each predi ctor, based on their past predictive performance. The proposed method outperforms both ANN and LR predictions. (C) 1997 Elsevier Science Ltd