Hierarchical Overlapping Community Detection for Weighted Networks

Real-world networks often contain community structures, where nodes form tightly interconnected clusters. Recent research indicates hierarchical organization, where vertices split into groups that further subdivide across multiple scales. However, individuals in social networks typically belong to multiple communities due to their various affiliations, such as family, friends, and colleagues. These overlaps will emerge in the community structure of online social networks and other complex networks like in biology, where nodes have diverse functions. In this work, we propose an algorithm for hierarchical overlapping community detection in weighted networks. The overlap between clusters is realized via maximal cliques that are used as base elements for hierarchical agglomerative clustering on the graph (GHAC). The closed trail distance and the size of the maximal clique in overlap are used for the dissimilarity between clusters in agglomerative steps of the GHAC. The closed trail distance is designed for weighted networks. Experiments on synthetic networks and different evaluations of the results of experiments show that the proposed algorithm is comparable with other widely used algorithms for overlapping community detection and is efficient for detecting hierarchy structure in weighted networks.