结合拓扑势与信任度调整的重叠社区发现算法

现实世界中的复杂系统可建模为复杂网络，探究复杂网络中的社区发现算法对于分析复杂网络的拓扑结构和层次结构具有重要作用。早期研究通常将网络中的节点局限在一个社区中，但随着研究的深入发现社区结构呈现重叠特性。针对现有重叠社区发现算法存在划分社区结构不稳定、忽略节点交互和属性等问题，提出一种基于网络拓扑势与信任度调整的重叠社区发现算法。融合节点的属性和结构特征计算节点的拓扑势，依据节点的拓扑势选取核心节点。从核心节点出发构建初始社区群，计算各个社区间的调整信任度，实现社区的合并与再调整，从而识别重叠社区。在多个人工模拟网络和真实网络数据集上的实验结果表明，与基于贪婪派系扩张、种子扩张等的重叠社区发现算法相比，该算法将扩展模块度最高提升至0.719，能有效识别社区结构及重叠节点，提升重叠社区检测性能。

Overlapping Community Discovery Algorithm Combining Topology Potential and Trust Adjustment

All types of complex systems in the real world can be modeled as complex networks.The study of community discovery algorithms in complex networks plays an important role in analyzing the topology and hierarchy of complex networks.Early studies limited the nodes in the network to one community.With more extensive research, it was determined that community structures also display overlapping characteristics.This paper focuses on the defects of existing algorithms, which consist of an unstable structure of divided communities and ignore the node interaction and attributes.In this study, the proposed overlapping community discovery algorithm combines topology potential and trust adjustment.The topological potential is first computed using the attribute value and structural characteristics, and the core node is selected based on the topological potential.Next, starting from the core node, initial communities centered on the core node are formed.The communities are subsequently merged and readjusted by calculating the adjustment trust between each community to enable the detection of overlapping communities.The experimental results on several artificial simulated networks and real network datasets show that compared to overlapping community discovery algorithms based on Greedy Factional Expansion(GFE), Two Expansions of Seeds(TES), etc., the proposed algorithm increases the expansion module degree by up to 0.719, which allows effective identification of community structure and overlapping nodes, improving overlapping community detection performance.