Item Recommendation using Tag Expansion and Temporal Information
Hyun-Woo Kim, Hyoung-Joo Kim · 2012
Most recommender systems have cold-start problem. The system generates poor recommendations to new users, because new users do not provide enough information to make user profile. In this paper, we propose a recommendation method which alleviates cold-start problem and predicts user's interests. In social tagging system, tagging information can be used in recommendation process. We investigate tag expansion to solve cold-start problem. A user's tag set constitutes the user profile and it is expanded by n-gram model in natural language processing. We also take an item's tag popularity and temporal information into account. The experimental results show that proposed approach recommends items precisely with tag expansion to cold-start users. The system provides better recommendations reflecting the user's interests.