The prediction of the medium term power load based on combined model of the Bayes theory and LS-SVM

This paper presents a new time-series forecasting methods-Bayesian least squares support vector machines which applies the Bayesian evidence framework to infer automatically model parameters of LS-SVM regression. using the mixed Bayesian least squares support vector machines algorithm. Comparing with other models,we can see that this method has favorably high predicting precision and can achieve better results.