Using least squares support vector machines in short-term electrical load forecasting

Abstract

This paper deals with the application of a least squares support vector machine (LS-SVM) in short-time load forecasting (STLF). The objective of this paper is to examine the feasibility of SVM in STLF by comparing it with a artificial neural network (ANN). The experiment shows that LS-SVM outperforms the ANN based on the criteria of mean absolute error (MAE), mean absolute percent error (MAPE), mean squared error(MSE)and root mean square error(RMSE). Analysis of the experimental results proved that it is advantageous to apply LS-SVM to STLF.