Wavelet Support Vector Machines for Short-term Load Forecasting

On the basis of sufficient study and comparison of a variety methods of short term load forecasting, a new methodology called Wavelet Support Vector Machines (WSVM) is presented. The proposed algorithm is based on the Kernel method of Support Vector Machines (SVM) and wavelet frame, a wavelet basis was introduced to construct the kernel function of SVM. Furthermore, an algorithm for constructing and training the WSVM is presented. Analysis of the experimental results indicates that WSVM can achieve greater accuracy than the traditional methods.