Search papers, labs, and topics across Lattice.
This paper introduces ViHoRec, a meticulously constructed Vietnamese hotel recommendation dataset comprising 18,267 interactions among 6,832 users and 560 hotels, addressing the challenges of cross-platform name reconciliation, quality auditing, and privacy preservation. The dataset includes a reproducible construction pipeline and a cold-start benchmark that highlights the performance drop of learned models for users with limited interaction history. Key findings reveal that traditional UserKNN outperforms learned models in cold-start scenarios, emphasizing the dataset's utility for low-resource recommendation systems.
UserKNN outshines advanced models in cold-start scenarios, revealing the limitations of learned approaches in low-resource settings.
Recommender-system research for Vietnamese remains limited by the absence of a public, well-documented hotel interaction resource. Building such a resource is challenging for three reasons: cross-platform hotel names must be reconciled before interactions are comparable; quality must be audited with reproducible metrics rather than ad hoc cleaning; and public release must preserve privacy while remaining benchmarkable under realistic cold-start conditions. We introduce ViHoRec, a quality-controlled Vietnamese hotel recommendation dataset of 18{,}267 interactions between 6{,}832 users and 560 hotels, crawled from Booking.com, Traveloka, and Ivivu. Our contributions are: (i) a reproducible construction pipeline with cross-platform entity resolution and quantitative quality control; (ii) a privacy-preserving release with HMAC pseudonyms; and (iii) a public cold-start benchmark with temporal leave-last-one-out split, data-centric ablations, and dependency-free baselines. On the public split, learned models degrade sharply for users with short histories (BPR-MF Recall@10: 0.065 vs. 0.120), while UserKNN remains strongest overall, establishing ViHoRec as a sparse, cold-start-dominated testbed for low-resource recommendation. All data are publicly available at https://github.com/MinhNguyenDS/ViHoRec.
