An Improved Gated Graph Neural Network for Sports Tourism Recommendation: User Embedded Representations and Attention Mechanisms

Abstract

In recent years, as the tourism industry rapidly develops, personalized recommendation systems have become an important tool for improving user experience and satisfaction. However, traditional methods face problems such as insufficient recommendation accuracy and low computational efficiency when dealing with large-scale data and complex user needs. Therefore, a sports tourism recommendation model (TRM) based on improved gated graph neural network and user embedded representation is proposed. The model integrates user behavior data and attraction features, and adds attention mechanism and gated unit to improve recommendation accuracy. The dataset used in the study is publicly available on TripAdvisor, which includes detailed user reviews, ratings, and historical behavior data of tourists. The baseline models used for comparison are graph neural networks and attention-based graph neural networks. The performance of this model is evaluated based on several metrics, including F1 score, accuracy, AUC, and inference speed. The research results indicate that the proposed model achieves the highest F1 score of 0.95 after approximately 150 iterations and an accuracy of 0.98 after approximately 100 iterations. Moreover, the model performs outstandingly in terms of recommendation accuracy, relevance, and computational efficiency, with an AUC value of 0.97, inference speed of 0.02 seconds, and computation time of 45 seconds. The findings denote that the proposed model effectively improves the personalization and computational efficiency of tourism recommendations, and can provide users with more accurate recommendations of tourist attractions.