Population-guided large margin classifier for high-dimension low-sample-size problems
作者：
Highlights：
• Unlike standard methods, our method (PGLMC) operates well on high dimensional low sample size (HDLSS) data sets.
• PGLMC is not sensitive to the intercept term and imbalanced data.
• Several theoretical properties of PGLMC are proven.
• Its dual formulation is relatively easy to implement (through quadratic programming).
• It is robust to the model specification, making it useful for various real applications.
摘要
•Unlike standard methods, our method (PGLMC) operates well on high dimensional low sample size (HDLSS) data sets.•PGLMC is not sensitive to the intercept term and imbalanced data.•Several theoretical properties of PGLMC are proven.•Its dual formulation is relatively easy to implement (through quadratic programming).•It is robust to the model specification, making it useful for various real applications.
论文关键词：Binary linear classifier,Data piling,High-dimension lowsample-size,Hyperplane,Large margin classification,Local structure information
论文评审过程：Received 18 January 2019, Revised 27 August 2019, Accepted 30 August 2019, Available online 31 August 2019, Version of Record 8 September 2019.
论文官网地址：https://doi.org/10.1016/j.patcog.2019.107030
