摘要
1引言
Vapnik等人根据统计学习理论提出的支持向量机学习方法[1],近年来受到了国际学术界的广泛重视.支持向量机的最大特点是根据Vapnik结构风险最小化原则,尽量提高学习机的泛化能力,即由有限的训练集样本得到的小的误差能够保证对独立的测试集仍然保持小的误差.
Support vector regression machines based on structural risk minimization have a good generalization performance. However, its effect is not good if there exists heterogeneity of variance in the regression models. In order to solve the problem, a kind of weighting support vector regression is proposed in this paper. The results of simulation experiments show the feasibility and effectiveness of the method.
出处
《计算机科学》
CSCD
北大核心
2003年第11期38-39,共2页
Computer Science
基金
广东自然科学基金(021349)
