Representative null space LDA for discriminative dimensionality reduction

作者：

Highlights：

• The research reveals the main problem of the classic null space LDA method: the intrinsic overfitting problem.

• A new approach, representative null space LDA (RNLDA), is proposed to solve the overfitting problem.

• Practical and efficient RNLDA algorithms and an automatic parameter setting algorithm are proposed.

摘要

•The research reveals the main problem of the classic null space LDA method: the intrinsic overfitting problem.•A new approach, representative null space LDA (RNLDA), is proposed to solve the overfitting problem.•Practical and efficient RNLDA algorithms and an automatic parameter setting algorithm are proposed.

论文关键词：Linear discriminant analysis,Dimensionality reduction,Feature selection,Null space,Overfitting,Singularity problem

论文评审过程：Received 8 January 2020, Revised 22 August 2020, Accepted 18 September 2020, Available online 22 September 2020, Version of Record 5 October 2020.

论文官网地址：https://doi.org/10.1016/j.patcog.2020.107664