Community centrality for node's influential ranking in complex network

Some tiny party of influential nodes may highly affect spread of information in complex networks. For the case of very high time complexity in the shortest path computation of global centralities, making use of local community centrality to identify influential nodes is an open and possible problem. Compared to degree and local centralities, a five-heartbeat forward community centrality is proposed in this paper, in which a five-step induced sub-graph of certain node in the network will be achieved. Next, we induce the minimal spanning tree (MMT) of the sub-graph. Finally, we take the sum of all weights of the MMT as community centrality measurement that needs to be the influential ranking of the node. We use the susceptible, infected and recovered (SIR) model to evaluate the performance of this method on several public test network data and explore the forward steps of community centrality by experiments. Simulative results show that our method with five steps can identify the influential ranking of nodes in complex network as well.