动态分布式社团检测算法

Ynamic Distributed Community Detection Algorithm

针对手持设备网络,提出动态分布式社团检测算法。首先利用节点相遇历史数据计算累积相遇持续时间与相遇次数均值,并作为动态阈值确定节点朋友集合,然后构建本地关系图,最后对本地关系图进行多社团检测。仿真结果表明该算法能够反映节点间关系的变化,更适用于动态变化的移动容迟网络环境。

This paper presents Dynamic Distributed Community Detection Algorithm for Pocket Switched Networks. The average value of accumulated contact duration and the average value of contact times within a time window are exploited as two dynamic thresholds to detect the Familiar Set. The Familiar Set information is then exchanged and updated to build up local social graph which is exploited to detect different communities of each node. Simulation results show that this algorithm reflects variations in the relationship between nodes, which is suitable for dynamic network environment.