Influencer Identification of Threshold Models in Hypergraphs

Abstract:

      This paper mainly studies the influence maximization problem of threshold models in hypergraphs, which aims to identify the most influential nodes in hypergraphs. Firstly, we introduce a novel information diffusion rule in hypergraphs based on Threshold Models and conduct the stability analysis. Then we extend the CI-TM algorithm, originally designed for complex networks, to hypergraphs, denoted as the H-CI-TM algorithm. Secondly, we use an iterative approach to get the globally optimal solutions. The analysis reveals that our algorithm ultimately identifies the most influential set of nodes. Based on the numerical simulations, H-CI-TM algorithm outperforms several competing algorithms in both synthetic and real-world hypergraphs. Essentially, when provided with the same number of initial seeds, our algorithm can achieve a larger activation size. Our method not only accurately assesses the influence of individual nodes but also identifies a set of nodes with greater impact. Furthermore, our results demonstrate good scalability when handling intricate relationships and large-scale hypergraphs. The outcomes of our research provide substantial support for the applications of the threshold models across diverse fields, including social network analysis and marketing strategies.