Multi granularity based label propagation with active learning for semi-supervised classification
作者：
Highlights：
• Granular computing offering some guidelines for sound structured thinking.
• Detailed analysis of the impact of neighborhood size k.
• Learning labels by two label propagation processes with diverse neighborhood size k.
• Three-way decision and active learning applied for further annotating data.
• Better results compared with the random data labeling methods.
摘要
•Granular computing offering some guidelines for sound structured thinking.•Detailed analysis of the impact of neighborhood size k.•Learning labels by two label propagation processes with diverse neighborhood size k.•Three-way decision and active learning applied for further annotating data.•Better results compared with the random data labeling methods.
论文关键词：Semi-supervised learning,Granular computing,Multi granularity,Label propagation,Active learning,Three-way decision
论文评审过程：Received 12 June 2020, Revised 11 November 2020, Accepted 21 November 2021, Available online 18 December 2021, Version of Record 1 January 2022.
论文官网地址：https://doi.org/10.1016/j.eswa.2021.116276
