Ecole d'ingénieur et centre de recherche en Sciences du numérique

On the reliability of profile matching across large online social networks

Goga, Oana; Loiseau, Patrick; Sommer, Robin; Teixeira, Renata; Gummadi, Krishna

KDD 2015, 21st ACM SIGKDD Conference on Knowledge Discovery and Data Mining, 10-13 August 2015, Sydney, Australia

Matching the profiles of a user across multiple online social networks brings opportunities for new services and applications as well as new insights on user online behavior, yet it raises serious privacy concerns. Prior literature has showed that it is possible to accurately match profiles, but their evaluation focused only on sampled datasets. In this paper, we study the extent to which we can reliably match profiles in practice, across real-world social networks, by exploiting public attributes, i.e., information users publicly provide about themselves. Today's social networks have hundreds of millions of users, which brings completely new challenges as a reliable matching scheme must identify the correct matching profile out of the millions of possible profiles. We first define a set of properties for profile attributes-Availability, Consistency, non-Impersonability, and Discriminability (ACID)-that are both necessary and sufficient to determine the reliability of a matching scheme. Using these properties, we propose a method to evaluate the accuracy of matching schemes in real practical cases. Our results show that the accuracy in practice is significantly lower than the one reported in prior literature. When considering entire social networks, there is a non-negligible number of profiles that belong to different users but have similar attributes, which leads to many false matches. Our paper sheds light on the limits of matching profiles in the real world and illustrates the correct methodology to evaluate matching schemes in realistic scenarios.

Document Doi Arxiv Hal Bibtex

Titre:On the reliability of profile matching across large online social networks
Type:Conférence
Langue:English
Ville:Sydney
Pays:AUSTRALIE
Date:
Département:Data Science
Eurecom ref:4622
Copyright: © ACM, 2015. This is the author's version of the work. It is posted here by permission of ACM for your personal use. Not for redistribution. The definitive version was published in KDD 2015, 21st ACM SIGKDD Conference on Knowledge Discovery and Data Mining, 10-13 August 2015, Sydney, Australia http://dx.doi.org/10.1145/2783258.2788601
Bibtex: @inproceedings{EURECOM+4622, doi = {http://dx.doi.org/10.1145/2783258.2788601}, year = {2015}, title = {{O}n the reliability of profile matching across large online social networks}, author = {{G}oga, {O}ana and {L}oiseau, {P}atrick and {S}ommer, {R}obin and {T}eixeira, {R}enata and {G}ummadi, {K}rishna }, booktitle = {{KDD} 2015, 21st {ACM} {SIGKDD} {C}onference on {K}nowledge {D}iscovery and {D}ata {M}ining, 10-13 {A}ugust 2015, {S}ydney, {A}ustralia}, address = {{S}ydney, {AUSTRALIE}}, month = {08}, url = {http://www.eurecom.fr/publication/4622} }
Voir aussi: