LOAD FORECASTING VIA SUBOPTIMAL SEASONAL AUTOREGRESSIVE MODELS AND ITERATIVELY REWEIGHTED LEAST-SQUARES ESTIMATION

We propose suboptimal least squares or IRWLS procedures for estimating the parameters of a seasonal multiplicative AR model encountered duri ng power system load forecasting. The proposed method involves using a n interactive computer environment to estimate the parameters of a sea sonal multiplicative AR process. The method comprises five major compu tational steps. The first determines the order of the seasonal multipl icative AR process, and the second uses the le-ast squares or the IRWL S to estimate the optimal nonseasonal AR model parameters. In the thir d step one obtains the intermediate series by back forecast, which is followed by using the least squares or the IRWLS to estimate the optim al seasonal AR parameters. The final step uses the estimated parameter s to forecast future load. The method is applied to predict the Nova S cotia Power Corporation's 168 lead time hourly load. The results obtai ned are documented and compared with results based on the Box and Jenk ins method.