Centrality Analysis in the Multi-layer Temporal Trade Network of the Belt and Road Initiative Based on Motif

In the Belt and Road trade network, a country’s centrality is reflected not only in first-order trade interactions but also in higher-order trade interactions. Higher-order trade interactions refer to trade exchanges involving three or more countries. This study uses motifs to characterize higher-order interactions. Firstly, based on the trade volume data of countries along the Belt and Road from 2005 to 2022, this study applies information entropy to simplify the network and constructs multilayer temporal import and export trade networks. Then, using the eigenvector centrality method in complex networks combined with motifs, the study theoretically derives supercentrality matrices and marginal centrality expressions. Finally, experiments using these centrality expressions reveal that motif types significantly influence the centrality scores and rankings of certain countries, highlighting the necessity of considering higherorder interactions in centrality studies. Moreover, motif weights, inter-layer coupling strength, and hidden transition probability have limited impact on centrality rankings, demonstrating the robustness and stability of the proposed motif-based marginal centrality measurement method.