Title :
Multiple SVM classification syatem based on Choquet integral with respect to composed measure of L-measure and Delta-measure
Author :
Lin, Wen-chih ; Huang, Chih-sheng ; Yih, Jeng-Ming ; Wu, Der-Bang ; Yu, Yen-kuei
Author_Institution :
Dept. of Comput. Sci. & Inf. Eng., Asia Univ., Wufeng, Taiwan
Abstract :
In order to overcome the situation that interactions exist between all classifiers from multiple classification system. In this study, we fuse the multiple SVM classifiers by fuzzy fusion algorithm with respect to a novel composed measure of L-measure and Delta (δ)-measure. We expect to gain a more accurate classification than single SVM and other combination method, like majority vote. From the experiment results, the fusion method based on the fuzzy fusion algorithm with respect to composed measure obtains advancement in terms of the performance of classification.
Keywords :
fuzzy set theory; pattern classification; support vector machines; δ-measure; Choquet integral; Delta-measure; L-measure; fuzzy fusion algorithm; multiple SVM classification system; Accuracy; Classification algorithms; Kernel; Machine learning; Polynomials; Support vector machines; Training; Fuzzy fusion; Fuzzy integral; L-measure; SVM;
Conference_Titel :
Machine Learning and Cybernetics (ICMLC), 2010 International Conference on
Conference_Location :
Qingdao
Print_ISBN :
978-1-4244-6526-2
DOI :
10.1109/ICMLC.2010.5580701