Improved VoteRank Algorithm to Identify Crucial Spreaders in Social Networks

Abstract

In the field of complex networks, Identifying crucial spreaders with high propagation ability is an important aspect of research, especially in the background of the global spread of COVID-19. In view of this, a large number of ranking algorithms and their improved versions have been proposed to evaluate the importance of nodes in the network, such as degree centrality, betweenness centrality, and k-core centrality. However, most of these methods neglect to consider the average shortest path between important nodes in the process of node importance evaluation, which will be difficult to ensure that the initial crucial spreaders have a large influence on the network. Recently, the VoteRank algorithm proposed a new idea for identifying widely distributed key spreaders based on the voting mechanism, but there are some aspects of this algorithm that require improvement. In this paper, we propose a VoteRank improved by degree centrality, k-core, and h-index (DKHVoteRank) for identifying critical spreaders in the complex networks. We introduce additional metrics to optimize the voting mechanism of the VoteRank to ensure that our algorithm can identify a widely distributed spreaders with high importance in the network. We conducted simulation experiments based on the Susceptible-Infected-Recovered (SIR) model on 12 different complex network datasets, and the results show that our proposed algorithm performs significantly better than other benchmark algorithms in terms of propagation capability, propagation scale, and applicability of the algorithm.