基于标签传递的有向图重叠社团发现算法研究

社团发现作为网络科学中一个重要的基础问题受到了广泛的关注和重视.针对社团结构的研究为我们提供了从中尺度上分析和理解网络的途径,具有重要的理论和实际意义.已有的研究大多关注无向图和非重叠社团的发现.本文基于标签传递和用户排序的思想设计了一个有向图上的重叠社团发现算法,实际数据上的实验表明了算法在发现用户多重社团属性和确定社团规模方面的有效性.

An overlapping community detection algorithm based on label propagation

With the increasing interest of the complex network research, community structure attracts more and more attention as a foundation problem of complex network.Research on community structure offers a mid-scale view to analyze and understand networks with great theoretical and practical significance.Most existing work focus on non-overlapping community detection in undirected networks and.In the present work, we propose an overlapping community detection algorithm in direct network based on label propagation and node ranking.Experiments on real-world network show that our algorithm can effectively identify multi-labels of users, which is also helpful in determining the community size.