计算机工程与应用 ›› 2020, Vol. 56 ›› Issue (17): 24-32.DOI: 10.3778/j.issn.1002-8331.2005-0089

• 热点与综述 • 上一篇    下一篇

结构化支持向量机研究综述

王霞,董永权,于巧,耿娜   

  1. 1.江苏师范大学 计算机科学与技术学院,江苏 徐州 221116
    2.江苏师范大学 电气工程及自动化学院,江苏 徐州 221116
  • 出版日期:2020-09-01 发布日期:2020-08-31

Review of Structural Support Vector Machines

WANG Xia, DONG Yongquan, YU Qiao, GENG Na   

  1. 1.School of Computer Science & Technology, Jiangsu Normal University, Xuzhou, Jiangsu 221116, China
    2.School of Electrical Engineering & Automation, Jiangsu Normal University, Xuzhou, Jiangsu 221116, China
  • Online:2020-09-01 Published:2020-08-31

摘要:

结构化支持向量机(Structural Support Vector Machine,SSVM)是支持向量机(Support Vector Machine,SVM)的变体算法,被广泛应用于多个领域。阐述了SSVM的发展过程,详细分析了SSVM各种具体实现算法的思想及表现上的优劣;并通过实验的对比讨论,发现了SSVM的各种具体实现算法在分类性能和分类效率上优于其他SVM算法,而在稳定性上则逊于后者;基于此,给出了SSVM的后续研究方向。

关键词: 结构化支持向量机, 结构粒度, 聚类技术, 结构化孪生支持向量机, 结构化非平行支持向量机

Abstract:

Structural Support Vector Machine(SSVM) is a variant algorithm of Support Vector Machine(SVM), which is widely used in many fields. The development process of SSVM is elaborated, and the thoughts and advantages and disadvantages of various specific implementation algorithms of SSVM are analyzed in detail. And through the comparison and discussion of experiments, it is found that various specific implementation algorithms of SSVM are superior to other SVM algorithms in classification performance and classification efficiency, but inferior to the latter in stability. Based on this, the future research direction of SSVM is given.

Key words: structural support vector machine, structural granularity, clustering technology, structural twin support vector machine, structural nonparallel support vector machine