Towards investigating influencers in complex social networks using electric potential concept from a centrality perspective

Abstract

Influencers in complex social networks are crucial for shaping the dynamics of information dissemination, viral marketing strategies, and controlling public opinion. This raises an important question: which influencers wield the most power within these networks, and how can we effectively identify and strategically utilize a key group of disseminators to maximize their influence across various domains? To address this, various heuristic strategies have been developed, targeting specific scenarios. However, these strategies often fall short by focusing solely on nodal topology and ignoring the intricate interconnections among nodes. This oversight can lead to imprecise evaluations and reduce the effectiveness of information spread. To overcome these limitations, a comprehensive approach that considers both the local and global topological dynamics of networks is essential. This paper introduces a novel centrality measure, electric potential centrality (EPC), which incorporates both the degree of nodes and the distances between them. Inspired by the concept of electric potential, EPC is designed to identify key influencers by extensively analyzing topological features, including local connections and the overall network structure. The effectiveness of EPC is rigorously evaluated through experiments on both synthesized and real-world networks, which are undirected and unweighted. Our findings reveal that EPC significantly surpasses existing methodologies, exhibiting the strongest correlation with baseline metrics within the susceptible–infected–recovered model.