Weighted kshell degree neighborhood method: An approach independent of completeness of global network structure for identifying the influential spreaders

Identifying the most influential spreaders in a network is very important to maximize or control the spreading in many fields such as accelerating the information diffusion, increasing the publicity of a new product, controlling the rumor in the social network, decelerating virus spreading and so forth. The kshell method and the degree centrality are the two popular measures applied to capture the spreading ability of a node. Nevertheless, the kshell method usually performs better in a complete network, whereas the degree centrality is used to measure the local influence of a node when complete structure of the network is unavailable. In this paper, we propose a measure namely "weighted kshell degree neighborhood method" which is independent of the degree of completeness of network structure. The proposed method estimates the spreading capability of nodes using composition of both the node's kshell and degree with tunable parameters. The effectiveness of the proposed method is verified with six real networks in comparison to Susceptible-Infected-Recovered (SIR) spreading epidemic model as a reference. The experimental result shows that the proposed method effectively identify more influential spreaders than the kshell method and degree centrality, including the other methods such as neighborhood coreness centrality, mixed degree decomposition, weight neighborhood centrality, and weighted kshell decomposition. The proposed method is also cost effective in terms of computational time even for a larger network also.