摘要
基于粗糙集(RS)理论和支持向量回归(SVR)方法,提出一种电力系统短期负荷预测方法.采用粗糙集理论对影响负荷预测的各因素进行约简,将约简后得到的最小条件属性集,以此确定输入样本的维数并构造训练样本,作为支持向量回归机的输入进行训练预测.在此基础上,利用已知历史负荷数据构造训练样本群,作为SVR的输入进行训练,采用训练完毕后的SVR模型进行负荷预测.实验结果表明,与神经网络方法和标准SVR方法相比,集成粗糙集和支持向量回归的负荷预测方法,可以在缩短训练时间的前提下获得较高的预测精度.
A novel method to short-term load forecasting of power system based on rough set theory (RST) and support vector regression (SVR) is presented. Firstly, the factors that affect the load forecasting are reduced using RST method. Then a SVR module is trained with the historical load data whose dimension is constructed according to the minimum attributes set acquired by RST. Finally the trained SVR module is used to forecast the future short-term load. The experimental results show that, when compared against both neural network method and standard SVR method, the proposed method can forecast more accurate results while shortening the training time.
出处
《华侨大学学报(自然科学版)》
CAS
北大核心
2007年第3期252-255,共4页
Journal of Huaqiao University(Natural Science)
基金
国家自然科学基金资助项目(50477010)
关键词
电力系统
训练样本
短期负荷预测
粗糙集
支持向量回归
power system
trained sample
short-term load forecasting
rough set
support vector regression