论文标题
基于新的集中度指标和新的模块化功能，对复杂网络上的社区检测
Community detection on complex networks based on a new centrality indicator and a new modularity function
论文作者
论文摘要
社区检测是网络研究中的一项重要且具有挑战性的任务。如今，人们大大关注当地的社区发现方法。其中，具有贪婪算法的社区检测通常始于识别称为网络中央节点的局部基本节点。社区通过优化模块化函数从这些中心节点扩展。在本文中，我们提出了一个新的中央节点指标和一个新的模块化函数。我们称之为局部中心性指标（LCI）的中央节点指标与众所周知的全球最大程度指标和局部最大程度指标一样有效。在某些特殊的网络结构上，LCI的性能更好。另一方面，我们的模块化函数F2克服了以前文献中提出的模块化函数的某些缺点，例如分辨率限制问题。结合贪婪的算法，LCI和F2，使我们能够确定现实世界网络和模拟基准网络的正确社区结构。基于归一化信息（NMI）的评估表明，基于LCI和F2的贪婪算法的社区检测方法的性能优于许多其他方法。因此，我们在本文中提出的方法可能值得注意。
Community detection is a significant and challenging task in network research. Nowadays, plenty of attention has been focused on local methods of community detection. Among them, community detection with a greedy algorithm typically starts from the identification of local essential nodes called central nodes of the network; communities expand later from these central nodes by optimizing a modularity function. In this paper, we propose a new central node indicator and a new modularity function. Our central node indicator, which we call local centrality indicator (LCI), is as efficient as the well-known global maximal degree indicator and local maximal degree indicator; on certain special network structure, LCI performs even better. On the other hand, our modularity function F2 overcomes certain disadvantages,such as the resolution limit problem,of the modularity functions raised in previous literature. Combined with a greedy algorithm, LCI and F2 enable us to identify the right community structures for both the real world networks and the simulated benchmark network. Evaluation based on the normalized mutual information (NMI) suggests that our community detection method with a greedy algorithm based on LCI and F2 performs superior to many other methods. Therefore, the method we proposed in this paper is potentially noteworthy.
