Unsupervised probabilistic feature selection using ant colony optimization
作者：
Highlights：
• We proposed an unsupervised method to remove redundant and irrelevant features.
• The algorithm needs no learning algorithms and class label to select features.
• Similarity between features will be considered in computation of feature relevance.
摘要
•We proposed an unsupervised method to remove redundant and irrelevant features.•The algorithm needs no learning algorithms and class label to select features.•Similarity between features will be considered in computation of feature relevance.
论文关键词：Feature selection,Unsupervised methods,Filter approaches,Ant colony optimization,Classification accuracy
论文评审过程：Received 19 August 2015, Revised 30 December 2015, Accepted 13 January 2016, Available online 22 January 2016, Version of Record 8 February 2016.
论文官网地址：https://doi.org/10.1016/j.eswa.2016.01.021
