Optimizing feature selection methods by removing irrelevant features using sparse least squares
作者：
Highlights：
• We offer a sparse method (SLS) based on least squares to reduce dimensionality.
• Irrelevant features can be detected and removed by SLS.
• Perturbations of an irrelevant feature do not change the least squares solution.
• SLS can be augmented to any feature selection algorithm.
• SLS optimizes the performance of feature selection algorithms.
摘要
•We offer a sparse method (SLS) based on least squares to reduce dimensionality.•Irrelevant features can be detected and removed by SLS.•Perturbations of an irrelevant feature do not change the least squares solution.•SLS can be augmented to any feature selection algorithm.•SLS optimizes the performance of feature selection algorithms.
论文关键词：Feature selection,Least squares,Singular value decomposition,Irrelevant features,Rank-1 update,Supervised learning
论文评审过程：Received 2 April 2020, Revised 18 February 2022, Accepted 16 March 2022, Available online 4 April 2022, Version of Record 7 April 2022.
论文官网地址：https://doi.org/10.1016/j.eswa.2022.116928
