A study of short-term load forecasting based on ARIMA-ANN

Abstract

ARIMA and ANN are very practical forecasting technology in the electric short-term load forecasting fields. ANN is extensively applied to electric load forecasting especially in recent years. Both ARIMA and ANN have different characteristics. ARIMA is suitable for linear prediction and ANN is suitable for nonlinear prediction. Because of the complexity of the historical load data and the randomness of a lot of uncertain factors influence, the observed data include the linear and nonlinear parts. The choice of the forecasting model becomes the important influence factor how to improve load forecasting accuracy. A combined model of ARIMA-ANN is proposed in the text. The linear part of the historical load data can be dealt with ARIMA, and ANN model can deal with the nonlinear part of historical load data. Empirical results indicate that a hybrid ARIMA-ANN model can improve the load forecasting accuracy.