Sistema de recomendación de contenidos audiovisuales basado en tecnologías semánticas para grupos de usuarios de TV digital

Abstract

Nowadays, every human activity that is supported by a technological platform generates an overwhelming amount of information which turns simple tasks into complex ones. In the domain of entertainment for example, Recommender Systems are used to reduce this information overload by assisting users in selecting contents according to their preferences. For this work, we consider the domain of groups of users of Terrestrial Digital Television, since watching a movie or program is normally done with family and friends. We propose a system that includes two services: i) an authentication service that makes use of facial recognition techniques to identify users in a group; and ii) a recommendation service, which uses the preferences of the group members (called homogeneous group when all members share most of is preferences and heterogeneous when they do not) in order to infer knowledge from information stored in an ontological profile, built up from users television program evaluation history. The system is evaluated by different size groups created from a five member’s typical family, and considering every possible combination of its members. We also include different age ranges and the presence of children.