Dynamically identify important nodes in the hypergraph based on the ripple diffusion and ant colony collaboration model

Abstract

Identifying important nodes plays an indispensable role in analyzing and regulating networks, and hypergraphs, as a classic high-order network, can represent the complex connections between nodes more concisely and intuitively. However, most existing methods for identifying important nodes in a hypergraph architecture are static and have low accuracy. The only few dynamic methods are very complex and the results are highly random. In view of the above situation, this paper proposes a algorithm to dynamically identify important nodes in a hypergraph based on information dissemination dynamics (IDD). The algorithm mainly includes two models, namely the ripple diffusion model and the ant colony collaboration model. Through multiple experiments in data sets of different sizes, it was proved that the important nodes identified by IDD are generally stronger in all aspects than the important nodes identified by other comparison methods, and the degree of matching with the real important nodes set is also higher.