An architecture and functional description to integrate social behaviour knowledge into group recommender systems

摘要

In this paper we consider the research challenges of generating a set of recommendations that will satisfy a group of users with potentially competing interests. We review different ways of combining the preferences of different users and propose an approach that takes into account the social behaviour within a group. Our method, named delegation-based prediction method, includes an analysis of the group characteristics, such as size, structure, personality of its members in conflict situations, and trust between group members. A key element in this paper is the use of social information available in the Web to make enhanced recommendations to groups. We propose a generic architecture named ARISE (Architecture for Recommendations Including Social Elements) and describe, as a case study, our Facebook application HappyMovie: a group recommender system that is designed to provide assistance to a group of friends that might be selecting which movie to watch on a cinema outing. We evaluate the performance (compared with the real group decision) of different recommenders that use increasing levels of social behaviour knowledge.