Identifying influential spreaders based on diffusion K-truss decomposition

Identifying the most influential spreaders is important in optimizing the network structure or disseminating information through networks. Recent study showed that the K-truss decomposition could filter out the nodes that performed a worse spreading behavior in the maximal K-shell subgraph. The spreaders belonging to the maximal K-truss subgraph show better performance compared to previously used importance criteria. However, the accuracy of the K-truss or the K-shell in determining node coreness is largely susceptible to core-like group. In this paper, we propose an improved diffusion K-truss decomposition method by considering both the diffusion and clustering of edges to eliminate the impact of core-like group on identifying influential nodes. To validate the effectiveness of the proposed method, we compare it with five typical methods by carrying out Monte-Carlo simulations over six real complex networks. Simulation results demonstrate that the proposed method can effectively disintegrate the core-like group and accurately identify the influential nodes.