An overlap-sensitive margin classifier for imbalanced and overlapping data

作者：

Highlights：

• An overlap-sensitive margin classifier was proposed.

• The method addresses both class imbalanced and overlapping problems.

• The hyperplane and margin separate the data sets into hard- and soft-overlap region.

• Comparison studies show that the proposed method outperformed existing methods.

摘要

•An overlap-sensitive margin classifier was proposed.•The method addresses both class imbalanced and overlapping problems.•The hyperplane and margin separate the data sets into hard- and soft-overlap region.•Comparison studies show that the proposed method outperformed existing methods.

论文关键词：Classification,Imbalanced class,Overlapping class,Support vector machine

论文评审过程：Received 16 August 2017, Revised 6 January 2018, Accepted 7 January 2018, Available online 9 January 2018, Version of Record 17 January 2018.

论文官网地址：https://doi.org/10.1016/j.eswa.2018.01.008