Recommender System Employing Personal-Value-Based User Model

This paper proposes a recommender system based on personal-value-based user model. Conventional methods such as collaborative and content-based approaches tend to be less accurate for new users and items due to the lack of a relation between items and user preferences. While existing recommender systems usually employ user preferences of items for making recommendations, proposed method focuses on users’ personal values, which mean value judgments regarding on which attributes users put a high priority. The proposed recommender system employing personal-value-based user model is thus expected to realize more precise recommendations in cold-start situations. As one of typical cold-start situations, a prototype system is developed for recommendation using external resources. Experimental results show that generated user models reflect each user’s value judgment on attributes. In addition, the results also show that recommendation employing the proposed user model realizes improvements of precision in cold-start situations.