Key Node Identification Based on Hypergraph Threshold Model and PRK-Shell Algorithm

The identification of key nodes is vital in the research of many complex networks such as protein networks. Currently, the main research focuses on ordinary graphs that cannot describe high-order relationships. As the most classic method to identify key nodes in hypergraphs, K-Shell finds it difficult to fully mine the local features of nodes when dealing with multiple nodes in the same shell. Thus, this paper proposes PRK-Shell algorithm, which introduces PageRank based on the K-Shell to rank the importance of nodes in the same shell, so as to improve the accuracy of key node identification. For the verification effectiveness, the information propagation process of nodes was simulated in the hypergraph through percolation model and threshold model, with experimental simulations conducted on the generated hypergraph and the real hypergraph. The experimental results show that the PRK-Shell algorithm has significant improvement in the ability to identify key nodes.