A Re-Ranking Algorithm for Identifying Influential Nodes in Complex Networks

Influential nodes in complex networks play more significant role than other nodes in many applications such as viral marketing. Identification of the most influential nodes and ranking them based on their spreading ability is a hot topic in the field of complex networks. However, when selecting a set of influential nodes, we need to consider the mutual influence between the nodes, rather than simply selecting nodes with strong influence. And how to select the initial node set to maximize the spreading scale after the propagation process has become a challenging issue. In this paper, based on node ranking methods, a novel re-ranking algorithm Re-rank by Information Spreading Function (
$RINF$
) is proposed to identify a set of influential nodes in complex networks. For each step in the proposed algorithm, select the node with the highest score, and then update scores of nodes based on local paths of the selected node with the help of information spreading probability function
$INF$
. The effectiveness of proposed algorithm is evaluated by Susceptible-Infected-Recovered (SIR) model. Compared with original methods, the performance of re-ranked versions of corresponding methods has increased from 3.1% to 61.1%. Compared with the best result of all benchmark methods, the performance of PageRank_INF (PageRank after re-ranking) improved by 2.9%, 1.4%, 1.5%, 3.6%, 3.2%, and 3.4% on BA, Jazz, PGP, Sex, USAir and Router network, respectively. Experimental results show that the proposed algorithm can well identify influential nodes in complex networks.