Authors: Supraja Gurajala, Joshua White, Brian Hudson and
Jeanna Matthews.
In online social networks (OSNs), the audience size commanded by an organization or an individual is a critical measure of that entity's popularity and this measure has important economic and/or political implications. Organizations can use information about their audience, such as age, location, etc to tailor their products or their message appropriately. But such tailoring can be biased by the presence of fake profiles on these networks. In this study, analysis of 62 million publicly-available Twitter user profiles was conducted and a strategy to retroactively identify automatically-generated fake profiles was established. Using a pattern matching algorithm on screen-names with an analysis of Tweet update times, a highly-reliable sub-set of fake user accounts were identified. Analysis of profile creation times and URLs of these fake accounts revealed distinct behavior of the fake users relative to a ground truth data set. The combination of this scheme with established social graph analysis will allow for time-efficient detection of fake profiles in OSNs.
