Trust-aware Recommendations

When forming a work team, the choice of its members is a key step. We believe that a recommender system could help to select the members of a team by taking into account individual competences criteria, as well as the relationships between people. One important aspect of this relationship is the trust that members have with each other. However, this concept of trust remains difficult to be used because it is subject to many definitions and is complex to grasp. We try in this research to shed some light on the notion of trust. In particular, we study if there is a correspondence between a previously declared trust (off-line) and an actual trust put into practice during an experiment. To use the trust to recommend profiles of competences that stem from a wider and more reliable social network, we also investigate if the use of transitivity and reciprocity for trust makes sense.