A Scale-Consistent Approach for Recommender Systems

In this paper we propose and develop a relatively simple and efficient approach for estimating unknown elements of a user-rating matrix in the context of a recommender system (RS). The critical theoretical property of the method is its consistency with respect to arbitrary units implicitly adopted by different users to construct their quantitative ratings of products. It is argued that this property is needed for robust performance accuracy across a broad spectrum of RS application domains.