摘要
支持向量机和粗糙集理论是两种分类技术。前者寻求最大化两类间隔的最优分类超平面,后者用逻辑规则解释分类。基于两者的关系,提出了一种复合算法,且将其推广到回归。新算法在一定程度降低了计算复杂度,且适用于软间隔分类。数值实验表明新算法是有效可行的。
Support vector machine （SVM） and rough set （RS） theory are two classification techniques. The former attempts to find an optimal hyperplane that maximize margin between two classes, and the later are designed to provide an explanation of the classification using logical rules. A compact algorithm is proposed based on the relationship between their principles. This new algorithm can reduce the complexity of computation to a certain degree and is especially useful for soft margin classifier; also it is generalized to regression. Numerical results illustrate that the algorithm is feasible and effective.
出处
《计算机工程与设计》
CSCD
北大核心
2007年第19期4729-4731,共3页
Computer Engineering and Design
基金
国家自然科学基金项目(60574075)
关键词
支持向量机
粗糙集
分类
最大化间隔
逻辑规则
软间隔
support vector machine
rough set theory
classification
maximize margin
logical rules
soft margin
