Analysis of the dissemination pathways for the influence of intangible cultural tourism empowered by spatio-temporal graph neural networks

The dissemination of intangible cultural heritage (ICH) through tourism is a dynamic process shaped by spatial, temporal, and social factors. Understanding the mechanisms behind this dissemination is crucial for effective cultural tourism planning and heritage preservation. Traditional statistical models fail to capture the complex, nonlinear relationships and interactions that drive cultural influence, hindering policymakers’ ability to implement targeted strategies. This research addresses these challenges by introducing an Attention-Enhanced Spatio-Temporal Graph Neural Network (AST-GNN) framework to model and analyze the pathways of cultural tourism dissemination. The AST mechanism prioritizes critical regions and time windows, enhancing the model’s ability to focus on influential cultural hubs and temporal patterns. The GNN captures spatial dependencies and dynamic interactions across regions, enabling accurate modeling of the complex relationships in cultural tourism dissemination. Using the Cultural Tourism Influence Dissemination Data, the model performs pre-processing through data cleaning, normalization, and Principal Component Analysis (PCA) to enhance feature extraction. The AST-GNN is designed to capture both local spatial dependencies and long-range temporal dynamics, with attention mechanisms to prioritize critical nodes and time periods of influence. Pathway detection and network analysis identify three major dissemination modes: (1) hub-driven expansion from major cultural centers, (2) bidirectional reinforcement loops between adjacent regions, and (3) event-triggered surges generating temporary but intense diffusion effects. The experimental results show that AST-GNN outperforms traditional models, achieving an accuracy of 0.98, F1-score of 0.95, MAE of 0.241, and R² of 0.785. These findings provide actionable insights for cultural tourism marketing, resource allocation, and heritage preservation, contributing to data-driven and sustainable tourism planning.