Title :
Using membership functions to improve multiclass SVM
Author :
Xiaodan, Wang ; Chongming, Wu
Author_Institution :
Missile Inst., Air Force Eng. Univ., ShaanXi, China
fDate :
31 Aug.-4 Sept. 2004
Abstract :
Support vector machine is originally designed for binary classification. One-against-one method is commonly used in multiclass classification. Based on the analysis of the decision process of max wins used in one-against-one method, new membership functions are introduced to resolve the possibly existed unclassifiable regions and a fuzzy SVM multiclass classification algorithm FSVM is proposed. Classification experiments result proves the effectiveness of FSVM.
Keywords :
fuzzy set theory; learning (artificial intelligence); pattern classification; support vector machines; FSVM; binary classification; decision process; fuzzy support vector machine; membership function; multiclass classification; Classification algorithms; Design engineering; Machine learning; Missiles; Pattern classification; Statistical learning; Support vector machine classification; Support vector machines; Testing; Voting;
Conference_Titel :
Signal Processing, 2004. Proceedings. ICSP '04. 2004 7th International Conference on
Print_ISBN :
0-7803-8406-7
DOI :
10.1109/ICOSP.2004.1441602
