计算机工程与应用 ›› 2023, Vol. 59 ›› Issue (3): 23-32.DOI: 10.3778/j.issn.1002-8331.2206-0230

• 热点与综述 • 上一篇    下一篇

多目标优化特征选择研究综述

张梦婷,杜建强,罗计根,聂斌,熊旺平,刘明,赵书含   

  1. 江西中医药大学 计算机学院,南昌 330004
  • 出版日期:2023-02-01 发布日期:2023-02-01

Research on Feature Selection of Multi-Objective Optimization

ZHANG Mengting, DU Jianqiang, LUO Jigen, NIE Bin, XIONG Wangping, LIU Ming, ZHAO Shuhan   

  1. College of Computer Science, Jiangxi University of Traditional Chinese Medicine, Nanchang 330004, China
  • Online:2023-02-01 Published:2023-02-01

摘要: 特征选择是模式识别领域中有效的降维方法,当特征选择涉及到的多个目标彼此冲突,难以平衡时,将特征选择视为多目标优化问题是时下的研究热点。为方便研究者系统地了解多目标特征选择领域的研究现状和发展趋势,对多目标特征选择方法进行综述。阐明了特征选择和多目标优化的本质;根据多目标优化方法的区别和特点,重点对比剖析各类多目标优化特征选择方法的优劣势;讨论现有多目标优化特征选择研究方法存在的问题以及对未来的展望。

关键词: 特征选择, 多目标优化, 进化算法

Abstract: Feature selection is an effective dimensionality reduction method in the field of pattern recognition. When multiple objectives involved in feature selection conflict with each other and are difficult to balance, it is a research hotspot to regard feature selection as a multi-objective optimization problem. In order to facilitate researchers to systematically understand the research status and development trend in the field of multi-objective feature selection, this paper reviews the methods of multi-objective feature selection. This paper clarifies the nature of feature selection and multi-objective optimization. According to the differences and characteristics of multi-objective optimization methods, it focuses on the advantages and disadvantages of various multi-objective optimization feature selection methods. It discusses the existing multi-objective optimization feature selection research methods problems and prospects for the future.

Key words: feature selection, multi-objective optimization, evolutionary algorithms