基于核密度分布的模糊双支持向量回归机  

Fuzzy Twin Support Vector Regression with Kernel Density Distribution

作　　者：付晓晓 张晓丹[1] 

机构地区：[1]北京科技大学数理学院,北京100083

出　　处：《济南大学学报（自然科学版）》2017年第6期529-535,共7页Journal of University of Jinan(Science and Technology)

基　　金：中央高校基本科研业务费资助项目(FRF-BR-12-021)

摘　　要：在双支持向量回归机的基础上,考虑到不同的样本点可能对回归函数产生不同的影响,提出一种改进的双支持向量回归机模型,即基于核密度分布的模糊正则化双支持向量回归机;该模型不仅考虑到最小化结构风险项,而且还设计一种基于核密度分布的模糊隶属度函数,给不同的样本点赋予不同的模糊隶属度。结果表明,所提出的基于核密度分布的模糊双支持向量回归机有较理想的回归效果。The different samples may have different effects on the regression function. An improved twin support vector regression,that is,a fuzzy regularized twin support vector regression with kernel density distribution was proposed based on the twin support vector regression. This model not only considered the minimization of structure risk,but also designed a fuzzy kernel the density distribution based on membership function for different samples with different fuzzy memberships. The results show that the proposed fuzzy twin support vector regression with kernel density distribution has an ideal effect on regression.

关 键 词：双支持向量机 模糊正则化 核密度分布 隶属度函数 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程]