Dual-granularity weighted ensemble clustering

作者：

Highlights：

• The cluster global reliability is estimated by an uncertainty measurement method in the rough set at the cluster level.

• The sample local similarity measurement method is proposed at the sample level.

• The size of the base clustering member has little effect to the quality of the final fusion result.

• The quality of the final fusion result maintains a relatively stable level when the diversity of the base clustering member changes.

摘要

•The cluster global reliability is estimated by an uncertainty measurement method in the rough set at the cluster level.•The sample local similarity measurement method is proposed at the sample level.•The size of the base clustering member has little effect to the quality of the final fusion result.•The quality of the final fusion result maintains a relatively stable level when the diversity of the base clustering member changes.

论文关键词：Ensemble clustering,Rough set theory,Rough fuzzy degree,The measurement of uncertainty,Similarity matrix

论文评审过程：Received 22 September 2020, Revised 28 March 2021, Accepted 3 May 2021, Available online 6 May 2021, Version of Record 11 May 2021.

论文官网地址：https://doi.org/10.1016/j.knosys.2021.107124