A simple graph-based semi-supervised learning approach for imbalanced classification

作者：

Highlights：

• A new graph-based semi-supervised learning method is proposed.

• The graph structure and the class imbalance are taken into account in our method.

• There are interesting connections between our method and different areas.

• Experimental results demonstrate our method can achieve promising performance on several benchmark datasets.

摘要

•A new graph-based semi-supervised learning method is proposed.•The graph structure and the class imbalance are taken into account in our method.•There are interesting connections between our method and different areas.•Experimental results demonstrate our method can achieve promising performance on several benchmark datasets.

论文关键词：Graph-based semi-supervised learning,Class imbalance,Markov stability,Group inverse

论文评审过程：Received 24 June 2020, Revised 24 February 2021, Accepted 30 April 2021, Available online 17 May 2021, Version of Record 30 May 2021.

论文官网地址：https://doi.org/10.1016/j.patcog.2021.108026