Recommendations In The Virtual Societies: Some Considerations

Abstract

We aim to show how the classical tool of social recommendation, massively used in social interactions, has to be more deeply analyzed within the new digitally infrastructured societies. We state that within these highly dynamic contexts it is fundamental to restructure the concept of recommendation. We tested our solutions by means of a multi-agent social simulation, identifying when it is more convenient to combine inferential processes with recommendations, i.e. focusing on recommending categories of agents rather than specific individuals. We found that within open networks and in the presence of not so reliable recommenders, category’s recommendations are better. The results we obtained are in agreement with the literature and can be of important interest for the development of this sector.