Proposal of user modeling method and recommender system based on personal values

This paper proposes a user modeling method and a recommender system based on personal values. Existing recommender systems usually employ user preference of items to make recommendations. On the other hand, marketing fields have been taking notice of personal values because of its significant relation to potential preferences of users. Although personal values are expected to bring a new framework for modeling user's potential preference, modeling method of user's personal values aiming to recommender systems has not yet been established. The proposed method focuses on users' value judgments, which mean what sort of attribute users put a high priority. By analyzing the relationship between ratings for an item and attributes, the influence of each attribute on user's value judgment is extracted as a user model. This paper also proposes a configuration of recommender system using user model based on personal values. Recommender systems based on personal values are expected to provide users with diversified recommended items from a non-traditional viewpoint.