Node influence ranking in complex networks: A local structure entropy approach

Abstract

The mining of important nodes in complex networks is a topic of immense interest due to its wide applications across many disciplines. In this paper, a Local Structure Entropy (LSE) approach is proposed based on the Taslli entropy by removing nodes, and by considering the information of the first-order and second-order neighboring nodes, in order to explore the impact of removing nodes on the network structure. With this method, the degree and betweenness of the first-order and second-order adjacent nodes are combined by Taslli entropy, and the influential nodes are measured by the structural characteristics of the network after nodes removal. To verify the effectiveness of LSE, we compare our method with five existing methods and perform experiments on seven real-world networks. The experimental results indicate that the influential nodes identified by LSE are better than the existing methods in terms of the range of information dissemination and robustness. Moreover, it is negatively correlated with closeness centrality and the PageRank algorithm.