一种动态网络社区划分算法
An Algorithm of Identifying Communities in Dynamic Networks
-
摘要: 基于MDL (minimum description length)原则提出一种动态社区划分算法DCI (dynamic community identifi-cation).通过定义相邻时刻静态社区的演化关系,使得DCI算法不仅能发现具有不同生命周期的动态社区,并且所得结果能准确描述动态社区的演化过程.实验结果表明,与已有算法相比,DCI算法可以更加准确地划分动态社区,并在可接受的时间内完成较大规模动态网络社区划分.Abstract: This paper proposed a MDL(Minimum Description Length) principle based dynamic community partition algorithm that is named DCI (Dynamic Community Identification).This paper defines evolving relationships between static communities at adjacent time points,so that DCI can not only identify the dynamic communities with different life cycles but also describe their evolving processes accurately.Experiments demonstrate that the DCI algorithm can identify dynamic communities more accurately than the existing algorithms and it's applicable to processing large scale dynamic networks in acceptable time.