文献类型：期刊文献

英文题名：Ramp loss least squares support vector machine

作者：Liu, Dalian[1,2];Shi, Yong[1,3,4,5];Tian, Yingjie[3,4];Huang, Xiankai[6]

第一作者：Liu, Dalian

通讯作者：Shi, Y[1];Tian, YJ[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, Coll Informat Sci & Technol, Omaha, NE 68182 USA;[6]Beijing Union Univ, Tourism Inst, Beijing 100101, Peoples R China

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, Key Lab Big Data Min & Knowledge Management, Beijing 100190, Peoples R China.

年份：2016

卷号：14

起止页码：61-68

外文期刊名：JOURNAL OF COMPUTATIONAL SCIENCE

收录：;EI(收录号：20160902033306);Scopus(收录号：2-s2.0-84975686927);WOS:【SCI-EXPANDED(收录号:WOS:000379560000007)】；

基金：This work has been partially supported by grants from National Natural Science Foundation of China (No. 11271361, No. 61472390, No. 71331005, No. 71373023), Major International (Regional) Joint Research Project (No. 71110107026), "New Start" Academic Research Project of Beijing Union University (No. ZK10201409) 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：Least squares support vector machine; Sparse; Ramp loss; CCCP; Classification

摘要：In this paper, we propose a novel sparse least squares support vector machine, named ramp loss least squares support vector machine (RLSSVM), for binary classification. By introducing a non-convex and non-differentiable loss function based on the 8-insensitive loss function, RLSSVM has several improved advantages compared with the plain LSSVM: firstly, it has the sparseness which is controlled by the ramp loss, leading to its better scaling properties; secondly, it can explicitly incorporate noise and outlier suppression in the training process, and thirdly, the non-convexity of RLSSVM can be efficiently solved by the Concave-Convex Procedure (CCCP). Experimental results on several benchmark datasets show the effectiveness of our method. (C) 2016 Elsevier B.V. All rights reserved.