动态社会网络的社团结构检测与分析

真实社会网络如邮件、科学合作、对等网络等均可以用图进行建模. 近年来, 基于图的社团挖掘吸引了人们越来越多的研究兴趣, 它不仅可以帮助识别网络的整体结构, 还可以发现社团演变的隐藏规律. 尽管使用静态图进行社团挖掘已经被广泛采用, 但基于动态图的研究还比较少. 通过使用时间序列, 对动态图上的社团挖掘包括社团检测与分析进行研究, 提出了一个新的动态社团结构检测模型, 并采用真实网络数据集进行了实验. 实验结果显示该模型在社团结构发现的有效性和效率性方面均有着良好的表现.

Community Detecting and Analyzing in Dynamic Social Networks

Real networks such as e-mail, co-author and peer-to-peer networks can be modeled as graphs. Community mining on graphs has attracted more and more attentions in recent years. It not only can help to identify the overall structures of networks, but also can help to discover the latent rules of community evolution. Community mining on dynamic graphs has not been studied thoroughly, although that on static graphs has been exploited extensively. Based on time-sequence, the community mining including community detection and analysis on dynamic graphs is researched in this paper. And a two-step model is presented to discover the dynamic community structure. The effectiveness and efficiency of the model are validated by experiments on real networks. Results show that the model has a good trade-off between the effectiveness and efficiency in discovering communities.