Application of graph learning in travel recommendation system

Abstract

In the context of Vietnam's rapidly growing tourism industry, supporting tourists in choosing attractions that suit their personal interests has become very important. This paper studies the application of graph learning in recommender systems, thereby building a travel recommendation system, aiming to provide suggestions about attractions based on user ratings. The recommendation system is built on Graph Neural Network, specifically the LightGCN algorithm, an advanced machine learning technique that allows learning features and relationships from graph-structured data. We have tested the algorithm on three different datasets before applying it to the real system. The system is being deployed and tested on the Internet.