A Study on the Economic Effect Propagation Model of Regional Tourism Networks Based on Graph Attention Network

Addressing current challenges in regional tourism development research, such as unclear identification of tourism network nodes, weak representation of interregional economic linkages with technical infrastructure, and inaccurate simulation of economic effect propagation pathways, this study develops a regional tourism network economic effect propagation model based on Graph Attention Networks (GAT). First, drawing upon complex network theory, tourism economics, and spatial econometrics, this study clarifies the core essence of regional tourism network economic effects and defines key evaluation dimensions: node contribution, linkage strength, propagation efficiency, and economic growth drivers. Second, by collecting multidimensional data, including tourism resource endowment, transportation accessibility, economic development level, visitor flow data, and digital communication infrastructure, from China’s 31 provinces and municipalities from 2018 to 2023, we constructed a regional tourism network dataset, RTN-2024, encompassing 8 types of tourism resources and 5 levels of association strength. Third, we propose a three-stage research framework of “tourism network construction-economic effect identification-dissemination model optimization”. Utilizing GAT, we assign edge weights and extract node feature importance within the tourism network. This enables the construction of a dynamic economic effect dissemination model that accounts for spatial distance, policy interventions, and infrastructure connectivity, providing a computational reference for evaluating economic spillover effects in tourism networks supported by wireless communication systems, antenna-based access facilities, and regional electromagnetic information infrastructure.