Feature selection based on robust fuzzy rough sets using kernel-based similarity and relative classification uncertainty measures

作者：

Highlights：

• Radial basis function kernel-based similarity measure is proposed to compute fuzzy relations with mixed types features.

• A robust fuzzy rough set model is proposed using relative classification uncertainty measures.

• Heuristic backward coupled with forward algorithms are proposed to avoid getting stuck into a local optimum situation.

• Extensive experiments demonstrate that the proposed feature selection model is significantly superior to other models.

摘要

•Radial basis function kernel-based similarity measure is proposed to compute fuzzy relations with mixed types features.•A robust fuzzy rough set model is proposed using relative classification uncertainty measures.•Heuristic backward coupled with forward algorithms are proposed to avoid getting stuck into a local optimum situation.•Extensive experiments demonstrate that the proposed feature selection model is significantly superior to other models.

论文关键词：Feature selection,Fuzzy rough set,Similarity measure,Robustness,Noise

论文评审过程：Received 9 February 2022, Revised 24 August 2022, Accepted 25 August 2022, Available online 31 August 2022, Version of Record 12 September 2022.

论文官网地址：https://doi.org/10.1016/j.knosys.2022.109795