CAGM: A communicability-based adaptive gravity model for influential nodes identification in complex networks

Abstract

Identifying influential nodes is a very hot and challenging issue in the field of complex system and network. A great deal of algorithms have been developed to address the influential nodes identification problem, but most of previous studies are compromising between result accuracy and time cost. In this work, we propose a communicability-based adaptive gravity model (CAGM) for influential nodes identification. The key idea of CAGM algorithm is that the importance of each node is evaluated by comprehensively considering the influence probability and influence intensity information of neighbor nodes located in its influence radius. More specifically, the communicability network matrix is introduced to depict the influence probability between each pair of nodes. By integrating k-shell, degree and distance information, the influence radius of each node can be determined uniquely so as to portray the inherent heterogeneity of nodes in complex networks. To verify the effectiveness and applicability of CAGM, several groups of simulated experiments on twelve real and artificial datasets are concluded. Experimental results show that CAGM performs better than eight popular algorithms in terms of top-10 nodes, discrimination ability, imprecision function and ranking accuracy.