Using qualitative reasoning to model users profiles: an approach to movie recommendation

Abstract

This paper presents a methodology for a collaborative recommender system (RS). The methodology is based on the compatibility of groups of users defining their profiles via a qualitative order-ofmagnitude model. The distributive lattice structure of the space of qualitative descriptions is considered in defining the distance between existing users and the RSs new users. An application to movie recommendations is presented to show and compare the efficiency of the proposed methodology.