Abstract:Cluster ensemble can combine the outcomes of several clusterings to a single clustering that agrees as much as possible with the input clusterings and improves the stability of the clustering algorithm.However,little attention has been paid to the approaches to deal with consolidating the outcomes of hard clustering,which is a result of two-way clustering,into a soft clustering,which is a result of three-way clustering.Since the three-way decision theory is widely studied recently with its advantages of human cognition and decision-making,the main objective of this paper is to propose a transfer approach based on the framework of cluster ensemble.Here,an interval set is used here to represent a cluster,which is described by three regions according to the lower and upper bound of the cluster.First,the proposed approach gets different hard partitions through ensemble members which are the existing clustering algorithms.Then,it decides some objects definitely to the corresponding clusters by matching the clusters′ tag and intersection operations.Finally,it decides the rest objects by three-way decisions based on voting.The preliminary experimental results show that the proposed approach is effective.
胡凌超,于洪. 一种基于投票的三支决策聚类集成方法[J]. 小型微型计算机系统, 2016, 37(8): 1741-1745.
HU Ling-chao,YU Hong. Voting Cluster Ensemble Approach Based on Three-way Decisions. Journal of Chinese Computer Systems, 2016, 37(8): 1741-1745.