Modelling Traceability in Recommender Systems

Abstract

Recommender Systems are valuable tools which suggest meaningful and useful items to users. In a previous research project, a real recommender system which offers personalized recommendations of items to health professionals and medical specialists in the context of Continuing Medical Education (CME) was designed and developed. Traceability helps recommender systems to generatejustifications about the criteria used for selecting the suggestions of items to the active user. This paper presents a novel approach for modelling traceability in recommender systems in the given context. The proposed approach shows how to use different levels of relationships between users to trace the origin of the recommendations. An important contribution of this research is to explain how to generalize the proposed model of traceability in recommender systems. In addition, an automated approach towards communicating the origin of the recommendations to the users is proposed.