用FSVM解决多类分类问题中不可分区域问题剖析  

An Analysis of Resolving the Problem of Unclassifiable Regions in Multi-classification by Means of FSVM

作　　者：徐文彬[1] 陈美珠[2] 

机构地区：[1]常熟理工学院计算机科学与工程系,江苏常熟215500 [2]泰州职业技术学院,江苏泰州225300

出　　处：《常熟理工学院学报》2007年第8期84-88,共5页Journal of Changshu Institute of Technology

摘　　要：分析了支持向量机(SVM)的工作原理和将其推广到多类分类时会遇到的问题,对用模糊SVM(FSVM)解决此问题时的模糊策略作了详细论证,说明此模糊策略是非常完美的一个解决方案,并指出了进一步的研究方向。This paper analyzes the principle of support vector machine（SVM） and points out problems encountered in its extension to multi-classification.To solve this problem,fuzzy strategy of fuzzy SVM（FSVM） has been demonstrated in detail.And this vague strategy is a very good solution.The paper also suggests a direction for further research.

关 键 词：支持向量机 模糊支持向量机 分类超平面 

分 类 号：O235[理学—运筹学与控制论]