文献类型：期刊文献

英文题名：Robust discriminant feature selection via joint L-2,L-1-norm distance minimization and maximization

作者：Yang, Zhangjing[1] Ye, Qiaolin[2,3] Chen, Qiao[4,5] Ma, Xu[2] Fu, Liyong[4,5] Yang, Guowei[1] Yan, He[2] Liu, Fan[6]

第一作者：Yang, Zhangjing

通信作者：Fu, LY[1];Fu, LY[2]

机构：[1]Nanjing Audit Univ, Sch Informat Engn, Jiangsu Key Lab Auditing Informat Engn, Nanjing 211815, Peoples R China;[2]Nanjing Forestry Univ, Coll Informat Sci & Technol, Nanjing 210037, Jiangsu, Peoples R China;[3]Huaiyin Inst Technol, Lab Internet Things & Mobile Internet Technol Jia, Huaian 223003, Peoples R China;[4]Chinese Acad Forestry, Inst Forest Resource Informat Tech, Beijing 100091, Peoples R China;[5]Natl Forestry & Grassland Adm, Key Lab Forest Management & Growth Modeling, Beijing 100091, Peoples R China;[6]Hohai Univ, Coll Comp & Informat, Nanjing 210098, Peoples R China

年份：2020

卷号：207

外文期刊名：KNOWLEDGE-BASED SYSTEMS

收录：;EI(收录号：20203409088535);Scopus(收录号：2-s2.0-85089603936);WOS:【SCI-EXPANDED(收录号:WOS:000574944400013)】；

基金：This work was supported in part by the Central Public-interest Scientific Institution Basal Research Fund under grant CAFYBB2019QD003, the National Natural Science Foundation of China under Grants U1831127, 61876213, and 61772277, the Natural Science Foundation of Jiangsu Province under Grants BK20171453, and 61871444, the Program of Collaborative Innovation Center of IoT Industrialization and Intelligent Production (Minjiang University) under Grant IIC1705, and the Qinglan and Six Talent Peaks Projects of Jiangsu Province.

语种：英文

外文关键词：Feature selection; L-2,L-1-norm; Discriminative Feature Selection; Robustness

摘要：Discriminative Feature Selection (DFS) is an algorithm, proposed recently for effective feature selection by considering both joint linear discriminant analysis and row sparsity regularization. However, this method is not robust enough to protect the data from outliers, because it utilizes the squared L-2-norm distance metric. To overcome this problem, we present in this paper, a novel discriminative feature selection algorithm, which uses the robust L-2,L-1-norm for measuring the distances in DFS. Although the algorithm is apparently simple, it should not be considered trivial because of its non-convexity. Also, we present an analysis of the convergence, both theoretical and empirical. More importantly, we proposed an iterative algorithm to achieve optimal results. Experimental results, using various data sets, demonstrate the effectiveness of the proposed method. (C) 2020 Elsevier B.V. All rights reserved.