Analysis of Multipath Effects in Global Trade Networks Based on Multimodal, High Dimensional, Heterogeneous Transformer Architecture: Deconstruction of Nonlinear Cascade and Dynamic Chain Reaction

This paper proposes a multi-path effect analysis model for major economies in the<br>global trade network based on the MH-DP Transformer architecture. The global trade<br>network comprises interconnected nodes, with interactions among major economies<br>(China, the US, Japan, and Europe) being crucial. A trade network model with four<br>regional nodes and their connections is established to simulate trade flow propagation<br>and capture the nonlinear feedback mechanism through an iterative process. The study utilized three datasets: the Global Trade Dataset from the World Bank, the OECD GlobalValue Chain Dataset, and the SCM World Supply Chain Risk Dataset. These datasets provide comprehensive information on trade volumes, industrial interdependencies, and supply chain risks, respectively. To ensure uniformity and compatibility, the data underwent preprocessing, including normalization, standardization, and handling missing values. The datasets were aligned temporally, geographically, and industrially.