User Preference Through Learning User Profile for Ubiquitous Recommendation Systems

摘要

As ubiquitous commerce is coming, the ubiquitous recommendation systems utilize collaborative filtering to help users with fast searches for the best suitable items by analyzing the similar preference. However, collaborative filtering may not provide high quality recommendation because it does not consider user's preference on the attribute, the first rater problem, and the sparsity problem. This paper proposes the user preference through learning user profile for ubiquitous recommendation systems to solve the current problems. In addition, to determine the similarity between the users belonging to particular categories and new users, we assign different statistical values to the preference through learning user profile. We evaluated the proposed method on the EachMovie dataset and it was found to significantly outperform the previously proposed method.