A hybrid model, which combines the seasonal time series ARIMA (SARIMA) and the multilayer feedforwardneural network to forecast time series with seasonality, is shown to outperform both twosingle models. Besides the selection of transfer functions, the determination of hidden nodes to usefor the non linear model is believed to improve the accuracy of the hybrid model. In this paper, wefocus on the selection of the appropriate number of hidden nodes on the non linear model to forecastMalaysia load. Results show that by using only one hidden node, the hybrid model of Malaysia loadperforms better than both single models with mean absolute percentage error (MAPE) of less than 1%.
Copyrights © 2010