 |
    |
Recommendations Given from Socially-Connected People
Daniel González, Regina Motz, and Libertad Tansini
Facultad de Ingeniería, Universidad de la República, Montevideo, Uruguaydanielgonzalezbernal@gmail.com
rmotz@fing.edu.uy
libertad@fing.edu.uy
Abstract. This paper presents how relationships among members of a social network can be used to explicitly specify the relevant features of a friendsourcing recommendation algorithm. One important contribution is to show how to conceptualize previous evaluations of items made by socially-connected users and the different features involved in this kind of algorithms, in a set of criteria for similarity between users in a social network. The paper presents how these specified criteria are used by the proposed friendsourcing recommendation algorithm and shows how the recommendation algorithm is integrated into a real recommender system to be used in a healthcare social network for the medical service of a university. Moreover, the work shows preliminary results which indicate that the information contained in social networks, processed with the proposed algorithm, is relevant for the generation of personalized recommendations.
Keywords: Friendsourcing, Recommendation Algorithm, Recommender Systems
LNCS 8186, p. 649 ff.
lncs@springer.com
© Springer-Verlag Berlin Heidelberg 2013