基于支持向量机的短期负荷预测的方法改进

An Improved Method for Short-Term Load Forecasting Based on Support Vector Machines

在对支持向量机(Support Vector Machines,SVM)方法的参数性能进行分析的基础上,提出了将Grid-search方法引入至基于支持向量机的短期负荷预测算法中,以解决支持向量机方法的参数选择问题。该参数选择方法减少了参数选择的盲目性,提高了SVM的预测精度。通过在East-Slovakia Power Distribution Company提供的电网运行数据上验算,证明了该改进方法的正确性和有效性。

Grid-search method,which aims to solve the parameters selection problems for SVM method,is introduced to short-term load forecasting based on support vector machines after the parameter performance of support vector machines(SVM) is analyzed.This method reduces the blindness for the choice of an appropriate set of parameters,and it is helpful to improve the prediction precision of SVM method.Experiments are carried out on the basis of the history data offered by East-Slovakia Power Distribution Company.Simulation results show that this improved algorithm is valid and effective.