LBLP: link-clustering-based approach for overlapping community detection
LBLP: link-clustering-based approach for overlapping community detection
复制标题
DOI：
10.1109/tst.2013.6574677
复制
发表时间：
2013-08
影响因子：
6.6
通讯作者：
Le Yu;Bin Wu;Bai Wang
中科院分区：
文献类型：
--
作者：
Le Yu;Bin Wu;Bai Wang
近年来，复杂网络引起了广泛的研究关注。社区检测是复杂网络领域的一个重要问题，在信息传播、链接预测、推荐和营销等多种应用中都有重要作用。在本研究中，我们着重于使用链接分区来发现重叠的社区结构。提出了一种基于潜狄利克雷分配（Latent Dirichlet Allocation， LDA）的链路划分（Link Partition， LBLP）方法，该方法可以找到重叠范围可调的社区。该方法采用LDA模型检测链路分区，可以计算出每个链路的社区归属因子。在此基础上，可以有效地找到具有桥接链路的链路分区。我们通过使用现实世界和合成网络验证了所提出解决方案的有效性。实验结果表明，该方法可以找到有意义且相关的链路社区结构。
Recently, complex networks have attracted considerable research attention. Community detection is an important problem in the field of complex networks and is useful in a variety of applications such as information propagation, link prediction, recommendation, and marketing. In this study, we focus on discovering overlapping community structures by using link partitions. We propose a Latent Dirichlet Allocation (LDA)-Based Link Partition (LBLP) method, which can find communities with an adjustable range of overlapping. This method employs the LDA model to detect link partitions, which can calculate the community belonging factor for each link. On the basis of this factor, link partitions with bridge links can be found efficiently. We validate the effectiveness of the proposed solution by using both real-world and synthesized networks. The experimental results demonstrate that the approach can find a meaningful and relevant link community structure.