Answers are grounded in this paper and cite their sources.
As the global tourism industry expands, travellers are increasingly overwhelmed by the vast amount of information available on social media. Identifying travel destinations that align with individual behavioural patterns and psychological preferences requires cognitive effort and significant time, necessitating sophisticated recommendation systems. This study proposes a novel graph-based recommendation framework using data from Instagram, which is a key platform for visual and social travel cues. We introduce a Semantic-Enriched Contrastive Learning (SCL) method designed to better capture nuanced user preferences by identifying semantically related ‘neighbours’ within user-item networks. By integrating three distinct forms of contrastive learning into a graph neural network-based collaborative filtering approach, our method effectively addresses the data sparsity and noise inherent in social media interactions. Comparative analyses demonstrate that SCL outperforms existing benchmarks in providing tailored recommendations, thereby enhancing the user decision-making process. Beyond technical performance, we discuss the implications of our findings for digital marketing strategies and the design of user-centric information systems in the tourism domain.
No takes yet. Share an insight, caveat, or question.
Cho et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: