ViHoRec: A Quality-Controlled Vietnamese Hotel Recommendation Dataset and Cold-Start Benchmark

Researchers introduce ViHoRec, a public dataset of 18,267 interactions between 6,832 users and 560 hotels for Vietnamese hotel recommendation. The resource addresses challenges in cross-platform entity resolution and privacy-preserving release using HMAC pseudonyms.

• Contains data crawled from Booking.com, Traveloka, and Ivivu with a reproducible construction pipeline.
• Includes a public cold-start benchmark featuring temporal leave-last-one-out splits and dependency-free baselines.
• Shows that learned models degrade sharply for users with short histories (BPR-MF Recall@10: 0.065 vs. 0.120).
• Establishes UserKNN as the strongest overall model on this sparse, cold-start-dominated testbed.

The dataset serves as a benchmark for low-resource recommendation systems under realistic cold-start conditions.