Short-Term Load Forecasting Based on the BKF-SVM
暂无分享,去创建一个
Support vector machine has been widely used in the area of load forecasting, but there are still many disadvantages that are large processed data and slow processing speed etc when training data.. According to the disadvantages, this paper proposes a kind of forecasting method of SVM based on Boolean kernel function. In order to determine the super parameters which exert a direct influence on the ability of extension of SVM, the fixed step iteration method is presented, achieving the automatic selection of super parameters. The practical example shows that the system with BKF-SVM(Boolean Kernel Functions of SVM) method, comparing with the RBF-SVM method, when being applied to short-term load-forecasting has got higher prediction accuracy with such advantages as simple structure and good generalization performance without over-fitting phenomenon.
[1] Wang Hui-qing. Study on Forecasting Approach to Short-term Load of SVM Based on Data Mining , 2006 .
[2] Yang Zhang,et al. DRC-BK: Mining Classification Rules by Using Boolean Kernels , 2005, ICCSA.