Least-Squares Support Vector Machine Based on Improved Imperialist Competitive Algorithm in a Short-Term Load Forecasting Model

Abstract

To improve the accuracy of short-term load forecasting, a least-squares support vector machine (LSSVM) method based on improved imperialist competitive algorithm through differential evolution algorithm (ICADE) is proposed in this paper. Optimizing the regularization parameter and kernel parameter of the LSSVM through ICADE, a short-term load forecasting model that can take load-affected factors such as meteorology, weather, and date types into account is built. The proposed method is proved by implementing short-term load forecasting on the real historical data of the Yangquan power system in China. The result shows the proposed method improves the least-squares support vector machine capacity and overcomes the traditional imperialist competitive algorithm and least-squares support vector machine that exist in some of the shortcomings. The mean absolute percentage error is less than 1.5%, which demonstrates that the proposed model can be used in the short-term forecasting of the power system more efficiently.