GIR-based ensemble sampling approaches for imbalanced learning

作者：

Highlights：

• A novel metric measuring the class distribution imbalance is proposed.

• Theoretical properties of the proposed distribution imbalance metric are studied.

• Two adaptive sampling-based approaches are proposed for imbalance learning.

摘要

•A novel metric measuring the class distribution imbalance is proposed.•Theoretical properties of the proposed distribution imbalance metric are studied.•Two adaptive sampling-based approaches are proposed for imbalance learning.

论文关键词：Imbalanced learning,Generalized imbalance ratio,Undersampling and oversampling,Adaptive learning,Boosting and bagging

论文评审过程：Received 13 September 2016, Revised 1 May 2017, Accepted 11 June 2017, Available online 13 June 2017, Version of Record 21 June 2017.

论文官网地址：https://doi.org/10.1016/j.patcog.2017.06.019