An approach to optimizing abstaining area for small sample data classification

作者：

Highlights：

• Analysis of the optimal reject option for supervised classification.

• Problems of reject option based ROC curve for small-sample problems.

• New methods of reject option designed for small-sample problems.

• Improvement of performance of classifiers with reject option.

摘要

•Analysis of the optimal reject option for supervised classification.•Problems of reject option based ROC curve for small-sample problems.•New methods of reject option designed for small-sample problems.•Improvement of performance of classifiers with reject option.

论文关键词：Supervised learning,Reject option,Small-sample setting,Abstaining classifier,ROC curve estimation

论文评审过程：Received 3 November 2016, Revised 2 November 2017, Accepted 6 November 2017, Available online 9 November 2017, Version of Record 14 December 2017.

论文官网地址：https://doi.org/10.1016/j.eswa.2017.11.013