Identifying Influence of Nodes in Complex Networks with Coreness Centrality: Decreasing the Impact of Densely Local Connection

Ranking nodes by their spreading ability has been a fundamental issue related to many real applications, such as information and disease control. In the well-known coreness centrality measure, nodes' influence is ranked only by summing all neighbors' k-shell values while the effect of the topological connections among them on the nodes' spreading ability are ignored. In this work, we propose an improved coreness measure by decreasing the impact of densely local connections. Comparing the results from a series of susceptible-infected-recovered simulations on real networks, we show that our improved method can rank the nodes' spreading ability more accurately than other ranking measures such as k-shell, distance based method, mixed degree decomposition and coreness centrality method.