Constructing CP-Nets from Users Past Selection

Abstract

Although recommender systems have been significantly developed for providing customized services to users in various domains, they still have some limitations regarding the extraction of users’ conditional preferences from their past selections when they are in a dynamic context. We propose a framework to automatically extract and learn users’ conditional and qualitative preferences in a gamified system taking into consideration the players’ past behaviour, without asking any information from the players. To do that, we construct CP-nets modeling users preferences via a procedure that employs multiple Information Criterion score functions within an heuristic algorithm to learn a Bayesian network. The approach has been validated experimentally in the challenge recommendation domain in an urban mobility gamified system.