基于种子边的重叠社团发现算法

Overlapping community discovery algorithm based on edge of seed

社团结构是复杂网络的基本属性之一,重叠社团结构在现实世界中广泛存在,然而目前存在的发现重叠社团的算法非常有限。针对这一现状提出了基于种子边的重叠社团发现算法,该算法以边作为研究对象,主要思路是利用权重系数给出边的重要性排序,选出其中的种子边作为初始社团并对其进行扩展。最后,在Enron数据集上对算法进行了验证,并与LFM算法和改进加权的G-N算法进行了比较,证明了算法在有向赋权网路中发现重叠社团的有效性。

Community structure is one of the basic properties of complex networks, overlapping community structure in real world is widely existence. However, the existing overlapping community algorithms are very limited. For this situation, this paper proposed an overlapping community discovery algorithm based on the edge of the seed. The algorithm regarded edges as the re- search object, and used the weight coefficient to give the sorting of importance of edge, then it chose the seed edge as the initial community and extended. Finally,it made the experiment on the Enron data set and compared with the LFM algorithm and the improved weighted G-N algorithm. The results prove the effectiveness of this algorithm to find overlapping community in the di- rected and weighed network.