【摘要】 现有社区搜索方法用于从无向图中挖掘满足内聚性和影响力要求的社区，没有考虑有向图中边的方向对社区的影响，导致有向图上社区挖掘的结果出现影响力和内聚性不足的问题。基于此，提出有向图上的影响力社区搜索问题，并设计相应的在线搜索算法；为进一步提升社区挖掘的效率，提出有向图上基于索引的影响力搜索方法及其优化策略。此外，提出一种基于并行思想的索引构建方法，加速索引的构建过程。最后，基于8个真实数据集进行验证，实验结果验证了所提算法的有效性和高效性。更多还原

【Abstract】 Existing community search methods are used to explore communities in undirected graphs that meet cohesion and influence requirements, without considering the impact of edge direction in directed graphs. This oversight leads to insufficient influence and cohesion in the results of community detection on directed graphs. The problem of influence community search on directed graphs was proposed, and a corresponding online search algorithm was designed. To further enhance the efficiency of community mining, an index-based influence search method on directed graphs and its optimization strategies were proposed. In addition, a parallel-based index construction method was proposed to accelerate the index building process. Finally, based on eight real-world datasets, validation is conducted, and the experimental results confirm the effectiveness and efficiency of the proposed algorithm.更多还原