基于最小熵聚类的社团检测算法

提出一种基于最小熵的社团结构检测算法。首先用模糊关系表示交互网络,一种基于熵的测度来确定模糊关系的隶属度,且熵越小,节点越相似。然后提出一种新的模糊关系合成规则,通过应用该规则,模糊关系被转换为最小关系。最后,通过熵的值把这个最小模糊关系划分成一个个社团。在人工网络与真实网络中的测试结果表明,该算法可以有效地识别社团结构。

Community Detection Method Based on Minimum Entropy

This paper proposes an efficient method based on minimum entropy for community detection in complex networks.First,an interaction network is denoted by a fuzzy relation,and the entropy is proposed to determine the membership grade of the relation,and the lower the entropy,the more similar the vertices.Then,the fuzzy relation is transformed into a minimal fuzzy relation by a novel composition rule of fuzzy relation that is presented here.Finally,this minimal fuzzy relation is partitioned into clusters based on the value of entropy.The results both in artificial networks and real-world networks show that our method is efficient in community detection.