A Graph Neural Network Approach to Cross‐Cultural Narrative Visualization for Educational Reform

ABSTRACT Thai literary classics constitute an important medium for the interaction of Southeast Asian multicultural traditions, containing rich narrative symbols, identity construction mechanisms, and cross‐text cultural influence patterns. However, existing approaches to cross‐cultural influence analysis largely rely on manual interpretation or coarse representations, which limits their ability to capture contextual dependencies and systematically reconstruct cultural symbol systems. To address these challenges, this study proposes an automated cross‐cultural influence identification and visualization framework that integrates systemic functional linguistics with graph‐based learning. The proposed framework models culturally embedded narrative roles through a structured narrative function representation and captures influence relationships across literary texts in an interpretable manner. Experiments conducted on a newly constructed multilingual Thai literary corpus demonstrate that the proposed approach consistently outperforms representative baseline methods in cultural narrative function recognition accuracy, structural coherence, and cross‐context generalization. In particular, the framework shows clear advantages in identifying low‐frequency cultural functions and implicit symbolic expressions. Overall, this work provides a scalable computational framework for cross‐cultural narrative structure modeling and visualization in Thai literature, contributing methodological support to digital humanities research and regional literary education.