期刊
PHYSICA A-STATISTICAL MECHANICS AND ITS APPLICATIONS
卷 472, 期 -, 页码 145-155出版社
ELSEVIER
DOI: 10.1016/j.physa.2017.01.030
关键词
Community detection; Label propagation; Similarity; Stepping algorithm; Evaluation function
资金
- National Natural Science Foundation of China [61473131, 71571081]
- 973 Project of China [2013CB329506]
- Fundamental Research Funds for the Central Universities of China [2015TS032]
Community or module structure is one of the most common features in complex networks. The label propagation algorithm (LPA) is a near linear time algorithm that is able to detect community structure effectively. Nevertheless, when labeling a node, the LPA adopts the label belonging to the majority of its neighbors, which means that it treats all neighbors equally in spite of their different effects on the node. Another disadvantage of LPA is that the results it generates are not unique. In this paper, we propose a modified LPA called Stepping LPA-S, in which labels are propagated by similarity. Furthermore, our algorithm divides networks using a stepping framework, and uses an evaluation function proposed in this paper to select the final unique partition. We tested this algorithm on several artificial and real-world networks. The results show that Stepping LPA-S can, obtain accurate and meaningful community structure without priori information. (C) 2017 Elsevier B.V. All rights reserved.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据