Proposal of personal-value-based item modeling and its application to explanation of recommendation

This paper proposes a personal-value based item modeling, which is used for explaining recommendation. In recent years, studies on improvements of user's satisfactions for recommender systems by showing process of recommendation have been popular in addition to precision of recommendation. The proposed method extracts personal values of reviewers of a movie as the influence of movie's attribute on its total evaluation. This paper also examines the applicability of the extracted item model to explain recommendation by collaborative filtering.