文献类型：期刊文献

英文题名：Ramp loss nonparallel support vector machine for pattern classification

作者：Liu, Dalian[1,2];Shi, Yong[1,3,4,5];Tian, Yingjie[3,4]

第一作者：Liu, Dalian

通讯作者：Shi, Y[1]

机构：[1]Beijing Jiaotong Univ, Sch Comp & Informat Technol, Beijing 100044, Peoples R China;[2]Beijing Union Univ, Dept Basic Course Teaching, Beijing 100101, Peoples R China;[3]Chinese Acad Sci, Res Ctr Fictitious Econ & Data Sci, Beijing 100190, Peoples R China;[4]Chinese Acad Sci, Key Lab Big Data Min & Knowledge Management, Beijing 100190, Peoples R China;[5]Univ Nebraska, Collegae Informat Sci & Technol, Omaha, NE 68182 USA

第一机构：Beijing Jiaotong Univ, Sch Comp & Informat Technol, Beijing 100044, Peoples R China

通讯机构：[1]corresponding author), Chinese Acad Sci, Res Ctr Fictitious Econ & Data Sci, Beijing 100190, Peoples R China.

年份：2015

卷号：85

起止页码：224-233

外文期刊名：KNOWLEDGE-BASED SYSTEMS

收录：;EI(收录号：20152300911756);Scopus(收录号：2-s2.0-84937522201);WOS:【SCI-EXPANDED(收录号:WOS:000359331000017)】；

基金：This work has been partially supported by grants from National Natural Science Foundation of China (Nos. 61472390, 11271361, 71331005, and 11226089), "New Start" Academic Research Project of Beijing Union University (No. ZK10201409), Major International (Regional) Joint Research Project (No. 71110107026) and the Ministry of Water Resources special funds for scientific research on public causes (No. 201301094).

语种：英文

外文关键词：Support vector machine; Twin support vector machine; CCCP; Ramp loss; Sparseness

摘要：In this paper, we propose a novel sparse and robust nonparallel hyperplane classifier, named Ramp loss Nonparallel Support Vector Machine (RNPSVM), for binary classification. By introducing the Ramp loss function and also proposing a new non-convex and non-differentiable loss function based on the epsilon-insensitive loss function, RNPSVM can explicitly incorporate noise and outlier suppression in the training process, has less support vectors and the increased sparsity leads to its better scaling properties. The non-convexity of RNPSVM can be efficiently solved by the Concave-Convex Procedure and experimental results on benchmark datasets confirm the effectiveness of the proposed algorithm. (C) 2015 Elsevier B.V. All rights reserved.