文献类型：期刊文献

英文题名：Improved multi-kernel classification machine with Nystr?m approximation technique

作者：Zhu, Changming[1]; Gao, Daqi[1]

机构：[1] Department of Computer Science and Engineering, East China University of Science and Technology, Shanghai, 200237, China

年份：2015

卷号：48

期号：4

起止页码：1490

外文期刊名：Pattern Recognition

收录：EI(收录号：20144700227119)

语种：英文

外文关键词：Approximation algorithms - Risk assessment

摘要：Kernelized modification of Ho-Kashyap algorithm with squared approximation of the misclassification errors (KMHKS) is an effective algorithm for nonlinearly separable classification problems. While KMHKS only adopts one kernel function. So a multi-kernel classification machine with reduced complexity named Nystr?m approximation matrix with Multiple KMHKSs (NMKMHKS) has been developed. But NMKMHKS has to initialize many parameters and has not an ability to deal with noise well. To this end, we propose an improved multi-kernel classification machine with Nystr?m approximation technique (INMKMHKS). INMKMHKS is based on a new way of generating kernel functions and a new Nystr?m approximation technique. The contributions of INMKMHKS are that (1) avoiding the problem of setting too many parameters; (2) keeping comparable space and computational complexities after comparing with NMKMHKS; (3) having a tighter generalization risk bound in terms of Rademacher complexity analysis; (4) having a better recognition than NMKMHKS on average; (5) possessing an ability to deal with noise and practical images. ? 2014 Elsevier Ltd. All rights reserved.