文献类型：期刊文献

英文题名：One-Class Support Tensor Machine

作者：Chen, Yanyan[1,2];Wang, Kuaini[3];Zhong, Ping[1]

第一作者：Chen, Yanyan;陈艳燕

通讯作者：Zhong, P[1]

机构：[1]China Agr Univ, Coll Sci, Beijing 100083, Peoples R China;[2]Beijing Union Univ, Coll Appl Sci & Technol, Beijing 102200, Peoples R China;[3]Xian Shiyou Univ, Coll Sci, Xian 710065, Peoples R China

第一机构：China Agr Univ, Coll Sci, Beijing 100083, Peoples R China

通讯机构：[1]corresponding author), China Agr Univ, Coll Sci, Beijing 100083, Peoples R China.

年份：2016

卷号：96

起止页码：14-28

外文期刊名：KNOWLEDGE-BASED SYSTEMS

收录：;EI(收录号：20160701938249);Scopus(收录号：2-s2.0-84957866945);WOS:【SCI-EXPANDED(收录号:WOS:000370907200002)】；

基金：The work is supported by the National Science Foundation of China (Grant No. 11171346) and the "New Start" Academic Research Projects of Beijing Union University (Zk10201513). The authors also gratefully acknowledge the helpful comments and suggestions of the reviewers, which have improved the presentation.

语种：英文

外文关键词：Support vector machine; Support tensor machine; One-class classification; High dimensional and small sample size problem

摘要：In fault diagnosis, face recognition, network anomaly detection, text classification and many other fields, we often encounter one-class classification problems. The traditional vector-based one-class classification algorithms represented by One-Class Support Vector Machine (OCSVM) have limitations when tensor is considered as input data. This work addresses one-class classification problem with tensor-based maximal margin classification paradigm. To this end, we formulate the One-Class Support Tensor Machine (OCSTM), which separates most samples of interested class from the origin in the tensor space, with maximal margin. The benefits of the proposed algorithm are twofold. First, the use of direct tensor representation helps to retain the data topology more efficiently. The second benefit is that tensor representation can greatly reduce the number of parameters. It helps overcome the overfitting problem caused mostly by vector-based algorithms and especially suits for high dimensional and small sample size problem. To solve the corresponding optimization problem in OCSTM, the alternating projection method is implemented, for it is simplified by solving a typical OCSVM optimization problem at each iteration. The efficiency of the proposed method is illustrated on both vector and tensor datasets. The experimental results indicate the validity of the new method. (C) 2016 Elsevier B.V. All rights reserved.