Support vector machine and wavelet transform for short term load forecasting

For the short-term electric power load with uncertain influence factors, a novel approach is proposed by combining the wavelet transform (WT) and support vector machine (SVM). From the signal analysis point of view, load can also be considered as a linear combination of different frequencies. The proposed approach decomposes the historical load into an approximate part and several detail parts through the wavelet transform. Then based on the maximum and minimum loads of the approximate part, the similar coefficients are given. The support vector machine, trained by low frequencies and the corresponding temperature records, is used to forecast the maximum and minimum load of the forecasting day. The short term load forecasting is forecasted by summing the predicted approximate part and the weighted detail parts. The numerical simulation results show this approach is encouraging.