Large-scale Correlation of Accounts Across Social Networks

TitleLarge-scale Correlation of Accounts Across Social Networks
Publication TypeTechnical Report
Year of Publication2013
AuthorsGoga, O., Perito D., Lei H., Teixeira R., & Sommer R.
Other Numbers3419

Organizations are increasingly mining the personal data users generate as they carry out muchof their day?to?day activities online. A range of new business models specifically exploit whatusers publish on their social network profiles, including services performing background checksand analytics providers who, e.g., associate demographics with consumer behavior. In this workwe set out to understand the capabilities of machine learning techniques for linkingindependent accounts that users maintain on different social networks, based solely on theinformation people explicitly and publicly provide in their profiles. We perform a large scalestudy that assesses a range of correlation approaches for matching accounts between fivepopular social networks: Twitter, Facebook, Google+, Myspace, and Flickr. Our results show forinstance that by exploiting usernames, real names, locations, and photos, we can robustly


This work was partially supported by funding provided to ICSI through National Science Foundation grant CNS?1065240. Any opinions, findings, and conclusions or recommendations expressed in this material are those of theauthors or originators and do not necessarily reflect the views of the National Science Foundation.

Bibliographic Notes

ICSI Technical Report TR-13-002, Berkeley, California

Abbreviated Authors

O. Goga, D. Perito, H. Lei, R. Teixeira, and R. Sommer

ICSI Research Group

Networking and Security

ICSI Publication Type

Technical Report