A reliable ensemble based approach to semi-supervised learning
作者：
Highlights：
• We observe a shortage of easy-to-use wrapper methods for semi-supervised learning.
• The RESSEL method is introduced which combines ensemble- and semi-supervised learning.
• RESSEL is shown to improve upon supervised- and semi-supervised wrapper methods.
• An ablation study and parameter sensitivity study are conducted.
• The suitability of learners for self-training based on their ranking ability is investigated.
摘要
•We observe a shortage of easy-to-use wrapper methods for semi-supervised learning.•The RESSEL method is introduced which combines ensemble- and semi-supervised learning.•RESSEL is shown to improve upon supervised- and semi-supervised wrapper methods.•An ablation study and parameter sensitivity study are conducted.•The suitability of learners for self-training based on their ranking ability is investigated.
论文关键词：Ensemble learning,Out-of-bag error,Ranking,Self-training,Semi-supervised learning,Wrapper
论文评审过程：Received 15 July 2020, Revised 18 November 2020, Accepted 29 December 2020, Available online 14 January 2021, Version of Record 20 January 2021.
论文官网地址：https://doi.org/10.1016/j.knosys.2021.106738
