Application of SVM to power system short-term load forecast

The statistics learning method of SVM ( Support Vector Machines ) is introduced to short-term load forecast of power system.Sample data is constituted by filtering the historical data through clustering method.The object function considers both the fitness of prediction and error loss function.The large-scale optimization problem is solved by LIBSVM method.Corresponding software was developed and used to forecast the short-term load of a practical power system ,and the final forecast error is low.