Relevance–redundancy feature selection based on ant colony optimization
作者：
Highlights：
• New unsupervised feature selection methods using ant colony optimization are proposed.
• A new heuristic information measure is defined to enhance the accuracy of the methods.
• The proposed methods can efficiently handle both irrelevant and redundant features.
• The methods are compared to the well-known univariate and multivariate filter methods.
• The results show the efficiency and effectiveness of the proposed methods.
摘要
•New unsupervised feature selection methods using ant colony optimization are proposed.•A new heuristic information measure is defined to enhance the accuracy of the methods.•The proposed methods can efficiently handle both irrelevant and redundant features.•The methods are compared to the well-known univariate and multivariate filter methods.•The results show the efficiency and effectiveness of the proposed methods.
论文关键词：Pattern recognition,Curse of dimensionality,Feature selection,Multivariate technique,Filter model,Ant colony optimization
论文评审过程：Received 13 July 2014, Revised 5 February 2015, Accepted 25 March 2015, Available online 8 April 2015, Version of Record 16 May 2015.
论文官网地址：https://doi.org/10.1016/j.patcog.2015.03.020
