An accurate way to cross reference users across Social Networks

Abstract

Social Networks have permitted people have their own virtual identities which they use to interact with other online users. It is also completely possible and not uncommon for a user to have more than one online profile or even a completely different anonymous online identity. Sometimes it is needed to unmask the anonymity of certain profiles, or to identify two difference profiles as belonging to the same user. Entity Resolution (ER) is the task of matching two different online profiles potentially from different social networks. Solving ER has a number of benefits, amongst them are enhanced terrorist screening, improved marketing strategies, elimination of duplicate profiles, identification of fake profiles, etc. Our solution compares profiles based on various different attributes such as their usernames, real names, locations, languages and other similar attributes. We developed various string similarity algorithms and used other algorithms from another project to compare profiles, without using training. The system we developed was tested in a needle-in-a-haystack scenario where the system was tasked with matching two profiles that were in a pool of extremely similar profiles.