A novel fuzzy compensation multi-class support vector machine

被引:24
作者
Zhang, Yong [1 ]
Chi, Zhong-xian
Liu, Xiao-dan
Wang, Xiang-hai
机构
[1] Dalian Univ Technol, Dept Comp Sci & Engn, Dalian 116024, Peoples R China
[2] Liaoning Normal Univ, Dept Comp, Dalian 116029, Peoples R China
基金
中国国家自然科学基金;
关键词
support vector machine; multi-class classification; fuzzy compensation; membership;
D O I
10.1007/s10489-006-0027-x
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
This paper presents a novel fuzzy compensation multi-class support vector machine (FCM-SVM) to improve the outlier and noise sensitivity problem of traditional support vector machine (SVM) for multi-class data classification. The basic idea is to give the dual effects to penalty term through treating every data point as both positive and negative classes, but with different memberships. We fuzzify penalty term, compensate weight to classification, reconstruct the optimization problem and its restrictions, reconstruct Lagrangian formula, and present the theoretic deduction. By this way the new fuzzy compensation multi-class support vector machine is expected to have more generalization ability while preserving the merit of insensitive to outliers. Experimental results on benchmark data set and real data set show that the proposed method reduces the effect of noise data and yields higher classification rate than traditional multi-class SVM does.
引用
收藏
页码:21 / 28
页数:8
相关论文
共 26 条
[1]  
[Anonymous], P INT C ADV NAT COMP
[2]  
[Anonymous], 1998, Encyclopedia of Biostatistics
[3]  
Boser B. E., 1992, Proceedings of the Fifth Annual ACM Workshop on Computational Learning Theory, P144, DOI 10.1145/130385.130401
[4]   Knowledge-based analysis of microarray gene expression data by using support vector machines [J].
Brown, MPS ;
Grundy, WN ;
Lin, D ;
Cristianini, N ;
Sugnet, CW ;
Furey, TS ;
Ares, M ;
Haussler, D .
PROCEEDINGS OF THE NATIONAL ACADEMY OF SCIENCES OF THE UNITED STATES OF AMERICA, 2000, 97 (01) :262-267
[5]  
BURGES JC, 1997, ADV NEURAL INFORMATI, P375
[6]   Kernel-based Support Vector Machine classifiers for early detection of myocardial infarction [J].
Conforti, D ;
Guido, R .
OPTIMIZATION METHODS & SOFTWARE, 2005, 20 (2-3) :395-407
[7]  
CORTES C, 1995, MACH LEARN, V20, P273, DOI 10.1023/A:1022627411411
[8]  
GILARDI N, 1997, ACAI99, P43
[9]  
Guyon I., 1996, DISCOVERING INFORMAT
[10]   Support vector fuzzy regression machines [J].
Hong, DH ;
Hwang, CH .
FUZZY SETS AND SYSTEMS, 2003, 138 (02) :271-281