Trust-Driven Hybrid Recommender System for Personalized Tourism in Sparse Data Environments

The proliferation of online tourism data necessitates advanced recommender systems (RS) to mitigate information overload. Traditional collaborative filtering (CF) methods suffer from data sparsity and oversimplified single-criteria ratings. This paper proposes a Trust-Aware Hybrid Recommendation Framework (TAHRF) algorithm that integrates trust networks, multi-criteria ratings, and dynamic preference adaptation. The algorithm combines user- and item-based trust metrics with real-time feedback to improve recommendation accuracy and coverage. Experiments on TripAdvisor datasets demonstrate a 16–40% reduction in mean absolute error (MAE) compared to benchmarks. A user study (N=130) validates the system’s usability, with 89% coverage under 99.8% sparsity. The work advances tourism RS by addressing cold-start challenges and enhancing transparency through map-based visualization.