Hybrid Personalization for Recommendations

Abstract

In this paper we present the concept of hybrid personalization, the combination of multiple atomic personalization mechanisms. The idea of hybrid personalization is related to hybrid recommender systems, but works on a conceptual level—it is decoupled from the actual adaptation in the user interface. This has as an advantage that one can optimize the adaptation ‘behind the screens’ or—conversely—attach a new visualization mechanism to the personalization technique. We show the practical benefits of this layered, hybrid adaptation mechanisms by means of a case study on personalized curriculum planning where it is recommended which course could or should be followed at which state in the learning process.