APPLICATION OF SUPPORT VECTOR MACHINES BASED ON TIME SEQUENCE IN POWER SYSTEM LOAD FORECASTING
Yin He-jun · Power System Technology · 2004
Because power system load forecasting was uncertain, nonlinear, dynamic and complicated system, it wa difficult to describe such a nonlinear characteristics of thi system by traditional methods, so the load forecasting coul not be accurately forecasted. The authors presented a nove load forecasting method in which an improved Support Vecto Machines (SVM) algorithm based on time sequence wa applied and the principle of Structural Risk Minimizatio (SRM) was embedded into the SVM, therefore, on the basis o learning by fewer samples the presented method could conduc fast and accurate load forecasting with other samples fittin load forecasting. The presented method was more generalized and its dependence on experience was weakened. In the tim sequence the trend component and periodical component wer considered to make the load forecasting model more coinciden with the features of power loads. Applying the presente method to actual load forecasting, the comparison among th forecasted results and the true shows that the presented metho is feasible and effective.