三峡大学学报（自然科学版）2012，Vol.34Issue(6)：39-43,5.
基于非线性IRWLS-SVM短期电价预测的改进方法
Short Period Eelectricity Price Forecast Based on Nonlinear Irwls-svm Improved Method
马历 1马克雄2
作者信息
- 1. 中国长江三峡集团公司市场营销部,北京 100038
- 2. 三峡大学电气与新能源学院,湖北宜昌443002
- 折叠
摘要
Abstract
This paper has introduced a short period electricity price forecast model based on nonlinear iterative re-weighted least squares(IRWLS) support vector machine(SVM). First of all. it has simulated with different cost functions； and the result has indicated that Huber cost function has robustness, and it can adapt to small changes of the model. Then, this paper has compared with three kernel functions, i. e. linear kernel function, radial basis kernel function and polynomial kernel function. The simulation results show that the polynomial one has the best regression result. At last, the paper proposes an improved algorithm which can improve the local forecasting precision.关键词
支持向量机/非线性/短期电价预测Key words
support vector machine(SVM)/nonlinearing/short period electricity price forecast分类
信息技术与安全科学引用本文复制引用
马历,马克雄..基于非线性IRWLS-SVM短期电价预测的改进方法[J].三峡大学学报（自然科学版）,2012,34(6):39-43,5.
