No Free Lunch in imbalanced learning

作者：

Highlights：

• No Free Lunch theorems are verified in imbalanced learning tasks.

• Without a priori assumptions, resampling strategies have identical impact.

• Larger data set pools demonstrate predictive performance convergence.

• Reservations regarding performance of estimation techniques confirmed.

摘要

•No Free Lunch theorems are verified in imbalanced learning tasks.•Without a priori assumptions, resampling strategies have identical impact.•Larger data set pools demonstrate predictive performance convergence.•Reservations regarding performance of estimation techniques confirmed.

论文关键词：Supervised learning,Imbalanced domain learning,No Free Lunch

论文评审过程：Received 29 October 2020, Revised 14 May 2021, Accepted 9 June 2021, Available online 12 June 2021, Version of Record 17 June 2021.

论文官网地址：https://doi.org/10.1016/j.knosys.2021.107222