Social Network Analysis and Mining

, Volume 3, Issue 3, pp 475–495

User similarities on social networks

  • Cuneyt Gurcan Akcora
  • Barbara Carminati
  • Elena Ferrari
Original Article

DOI: 10.1007/s13278-012-0090-8

Cite this article as:
Akcora, C.G., Carminati, B. & Ferrari, E. Soc. Netw. Anal. Min. (2013) 3: 475. doi:10.1007/s13278-012-0090-8

Abstract

A key problem in online social networks is the identification of user characteristics and the analysis of how these are reflected in the graph structure evolution. The basis to tackle this issue is user similarity measures. In this paper, we propose a novel user similarity measure for online social networks, which combines both network and profile similarity. Since user profile data could be missing proposed measure is complemented by a technique to infer missing items from profile of the user’s contacts. The second main contribution of this paper is an extensive performance evaluation of the proposed measures with respect to some of the most relevant measures already proposed in the literature. The performance evaluation study has been conducted on a variety of data sets (i.e., Facebook, Youtube, Epinions and DBLP data sets) to see how different scenarios and graph characteristics affect the measures’ performance.

Copyright information

© Springer-Verlag Wien 2012

Authors and Affiliations

  • Cuneyt Gurcan Akcora
    • 1
  • Barbara Carminati
    • 1
  • Elena Ferrari
    • 1
  1. 1.DISTA, Università degli Studi dell’InsubriaVareseItaly

Personalised recommendations