动态网络中稳定社区发现

Detecting Stable Communities in Dynamic Networks

现实中的复杂网络通常是动态的.网络中的节点或联系随着时间的推移会发生变化,这种变化势必会造成网络中原本社区结构的改变.然而有些社区是稳定的,在短时间内它们不会发生剧烈变化.挖掘动态网络中的稳定社区有助于揭示动态网络的核心节点集,把握网络中的主要信息,预测动态网络在未来一段时间的动向.因此,挖掘动态网络中的稳定社区是有意义的.结合模式增长的理论与GN算法,提出一种动态网络中的稳定社区发现方法.该方法采用GN算法对动态网络在每个时间片上的静态结构进行社区划分,应用每个时间片上的社区划分结果及给定的稳定阈值挖掘频现节点集,挖掘过程揭示了稳定社区形成的层次结构及动态网络中的稳定节点与联系.

Complex netw orks are often dynamic in real life. The change of nodes and contacts can lead to change of community structures over time. How ever,some communities are stable,i. e. they do not change dramatically in a short time. M ining stable communities of dynamic netw orks can help revealing core nodes,grasping important information,and predicting trends of netw orks. A method combining the pattern grow th and GN algorithm for discovering stable community in dynamic netw orks is proposed. The proposed method discovers communities on each time slice in dynamic netw orks by using GN algorithm,and then revels the hierarchical structure of stable communities by extending sets of nodes.