A novel multi-labels classification algorithm

Abstract

For the problem of multi-labels classification, a multi-labels maximal-margin minimal-volume hypersphere support vector machine (SVM) is proposed by this paper. Use 1-a-r maximal-margin minimal-volume hypersphere SVM to train sub-classifiers, obtain membership vector of the sample to be classified according to the classifiers. At last give the labels that the sample to be classified belongs to according to the membership vector. The experimental results show that the algorithm has higher classification performance compared with the other multi-labels classification algorithms. © 2013 Springer-Verlag.