Electric load forecasting using Bayesian Least Squares Support Vector Machine

Electric load forecasting has received increasing attention over the years by academic and industrial researchers due to its major role for the effective and economic operation of power utilities. Least Support Vector Machine (LS SVM) is a new learning machine based on the statistical learning theory. A modelling approach based on least squares support vector machine (LS SVM) within the Bayesian evidence framework for short-term load forecasting is proposed. Under the evidence framework, the regularization and kernel parameters can be adjusted automatically, which can achieve a fine tradeoff between the minimum error and model´s complexities. The proposed approach is tested using actual power load data sets. Experimental results show that the proposed approach has better generalization performance and yields lower prediction error compared with LS SVM using the same test data set.