Ultra-short-term load forecasting based on adaptive bidirectional weighted least squares support vector machines

Weighted least squares fuzzy support vector machines method is proposed for ultra-short-term load forecasting. In order to reflect the characteristic that the nearer data have a greater impact on the predicting value,the membership distribution of time domain is introduced in bi-direction, namely, transverse (output samples) and longitudinal (training samples). To overcome the disadvantage of predicting with a fixed coefficient,fast leave-one-out method is used to adaptively optimize the parameters on-line. T he load data from a substation is used for simulating and the applications of different methods are compared. The results show that the proposed method can improve the forecasting accuracy compared with traditional methods.