On initial population generation in feature subset selection

作者：

Highlights：

• Feature subset selection (FSS) is an important tool for dimension reduction.

• Evolutionary algorithms are utilized for FSS.

• Result of evolutionary algorithms may be affected by initial population selection.

• We test a total of 5 different initial population generation methods in this study.

• Experimental results indicate that Information Gain method shows promising results.

摘要

•Feature subset selection (FSS) is an important tool for dimension reduction.•Evolutionary algorithms are utilized for FSS.•Result of evolutionary algorithms may be affected by initial population selection.•We test a total of 5 different initial population generation methods in this study.•Experimental results indicate that Information Gain method shows promising results.

论文关键词：Feature subset selection,Initial population,Multiobjective optimization

论文评审过程：Received 5 June 2018, Revised 8 June 2019, Accepted 26 June 2019, Available online 27 June 2019, Version of Record 4 July 2019.

论文官网地址：https://doi.org/10.1016/j.eswa.2019.06.063