Identifying influential spreaders based on improving communication transmission model and network structure

The identification of key nodes has garnered considerable attention across various research domains, including commercial marketing, infectious disease prevention and control, road network optimization, intelligent electricity management, and Chinese medicine formulation design. To evaluate critical nodes within complex networks, numerous algorithms have been proposed. However, these methods exhibit certain limitations such as one-sided considerations and single evaluation indexes. Consequently, the applicability of these key node identification approaches is limited to specific scenarios. In this paper, we propose a novel method for identifying key nodes based on principles from communication transmission theory and the law of gravity. Our approach fully explores the network structure and node relationships, integrating the Shannon channel capacity model and the gravity model. This method not only considers a node's own location and importance, but also the proximity of its neighbors to the target node within two hops. Through extensive experimentation, we demonstrate that our algorithm outperforms representative methods proposed in recent years in terms of consistency with the SIR model, similarity of Top-N nodes, and node importance differentiation. Real-world experiments conducted on COVID-19 networks further validate the effectiveness of our approach.