Research on Node Ordered Structure Identification of Temporal Multilayer Networks Based on Inter-layer Multi-order Structural Similarity
Accurate identification of key nodes in dynamic temporal networks is critical to network propagation control and governance. To address the issues that traditional temporal multilayer networks insufficiently consider node heterogeneity and inadequate inter-layer information fusion, this paper proposes the MSAM (Supra-Adjacency Matrix based on Multi-order Neighborhood Structure) model for node importance identification based on inter-layer multi-order structure similarity. The model characterizes intra-layer attributes via K-shell decomposition and the Pareto principle, measures inter-layer similarity by integrating multi-order neighbor and hierarchical fluctuation information, and realizes heterogeneous node coupling modeling. Experiments on real-world datasets demonstrate that MSAM achieves significantly better ranking performance than baseline models. It exhibits strong robustness and generalization ability under different propagation intensities, effectively improving the accuracy of node importance identification in dynamic networks.