Effective and efficient community search with size constraint on bipartite graphs

Community search on bipartite graphs has been extensively studied in suspicious-group detection and team formation. However, the existing studies focused on the cohesiveness of the community, but ignored the size constraint on bipartite graphs, potentially leading to large community sizes and high costs. In this study, a size-constrained (α,β)–community (SCC) containing a query vertex on a bipartite graph was investigated, where the upper layer size of the community cannot exceed threshold s and the lower layer size cannot exceed threshold t. For supporting SCC search in different situations, two search methods—peeling and expansion—are proposed by peeling from the (α,β)-core containing the query vertex and expanding from the query vertex respectively. An efficient lower bound based on degree gap is proposed by terminating unpromising search branches early to increase the efficiency of the community search. The experimental results indicated that the proposed methods can be used to find communities within the size thresholds, with the efficiency of the search increased based on the lower bound.