Identifying influential nodes in complex networks based on spreading probability

The identification of node importance is a challenging topic in network science, and plays a critical role in understanding the structure and function of networks. Various centrality methods have been proposed to define the influence of nodes. However, most existing works do not directly use the node propagation capacity for measuring the importance of nodes. Moreover, those methods do not have a high enough ability to distinguish nodes with minor differences, and are not applicable to a wide range of network types. To address the issues, we first define a method to calculate the propagation capability of nodes and divide the nodes in the network into an infected source and the uninfected nodes. The propagation capability of a source node is calculated from the probability that uninfected nodes are infected by the source, either directly or indirectly. Based on measuring the propagation ability of each node in the network, we propose a novel centrality method based on node spreading probability (SPC). Empirical analysis is performed by Susceptible–Infected–Recovered (SIR) model and static attacking simulation. We use six classical networks, and five typical methods to validate SPC. The results demonstrate that our method balances the measurement of node importance in the network connectivity and propagation structure with superior ability to discriminate nodes.