Classification for high-dimension low-sample size data

作者：

Highlights：

• The cause of data-piling is derived on High Dimensional Low Sample Size (HDLSS) data sets.

• A novel classification criterion on HDLSS, tolerance similarity is proposed.

• Leveraging on this criterion, a novel linear binary classifier (NPDMD) is designed.

• NPDMD is suitable for different real-world applications.

摘要

•The cause of data-piling is derived on High Dimensional Low Sample Size (HDLSS) data sets.•A novel classification criterion on HDLSS, tolerance similarity is proposed.•Leveraging on this criterion, a novel linear binary classifier (NPDMD) is designed.•NPDMD is suitable for different real-world applications.

论文关键词：Binary linear classifier,Quadratic programming,Data piling,Covariance matrix

论文评审过程：Received 29 November 2021, Revised 4 May 2022, Accepted 2 June 2022, Available online 6 June 2022, Version of Record 7 June 2022.

论文官网地址：https://doi.org/10.1016/j.patcog.2022.108828