Year: 2025
Type: article
Abstract: Personalized recommendation systems are widely used across domains but often fail to reflect region-specific characteristics, especially those tied to user stay patterns.Although accommodation reviews contain valuable insights into tourist behavior and sentiment, they remain underutilized in localized recommendation research.This study considers the HypAR (hypergraphs with attention on reviews for explainable recommendation) (Jendal et al., 2024) model, which incorporates accommodation review data to capture stay characteristics.HypAR integrates aspect, opinion, and sentiment information from textual reviews and employs a hypergraph attention mechanism to more precisely model the relationships between users and items.HypAR outperforms traditional methods such as MF, NMF, BPR, BiVAECF, LightGCN, DeepCoNN, and NARRE.By incorporating explainable AI, it enhances interpretability and captures complex dynamics, enabling the development of next-generation, region-aware recommendation systems.
Authors: Se-Yeon Jeong, Sooyoung Cheon
Institutions: Korea University, Korea University
Language: English
Subfield: Information Systems
Field: Computer Science
Domain: Physical Sciences
Open Access status: bronze
Funder: Korea University
OpenAlex ID: W7117961529
FWCI:
Cites:
Cited by:
Related to: