Chapter 4: Predicting Inter-Destination Tourism Flow Using Graph-Based Deep Learning: A Comprehensive Framework for Enhanced Tourism Management and Strategic Planning

The primary objective of this chapter is to develop a graph-based deep learning model for predicting Inter-Destination Tourism Flow (ITF) using multiple datasets by integrating both explicit and implicit features of tourism attractions. The methodology involves a hybrid architecture combining deep learning with graph-based techniques, where explicit features like popularity and amenities are encoded alongside implicit features derived from interactions within a tourism network using Graph Neural Networks (GNNs). The model’s predictions are further explained using Shapley Additive exPlanations (SHAP) to provide actionable insights for tourism management. Key findings demonstrate that the proposed model enhances ITF prediction accuracy, offering a valuable tool for destination marketing and infrastructure planning. The creation of a benchmark ITF dataset also facilitates future research in tourism flow analysis. The study concludes that the model’s explain ability and predictive capabilities can significantly contribute to sustainable tourism management.