Identifying Influential Nodes with a Community Structure Measure

Locating significant spreaders in social networks is an important issue for optimizing network structures so as to enhance the robustness of a system. While traditional methods are helpful in identifying influential nodes, they have some obvious limitations in overlooking the properties of social networks. For example, local metrics usually only account for neighbor size without reflecting the interactions between the nodes, though they are simple to use. Therefore, the topological connections among neighbors are neglected. On the other hand, global metrics are difficult to be applied in large social networks because of the high computational complexity. In order to take advantage of network properties, this paper introduces a new measure to identify influential nodes based on the combination of the influences of the community and the neighbors. SIR (Susceptible-Infected-Recovered) model is simulated to evaluate the nodes' spreading capabilities. In our experiment, we have demonstrated that the proposed measure can identify nodes with strong spreading influence, by comparing to the widely-used baseline measures.