Reputation Network Analysis for Email Filtering

Abstract

In addition to traditional spam detection applications, new methods of filtering messages – including whitelist and social network based filters – are being investigated to further improve on mail sorting and classification. In this paper, we present an email scoring mechanism based on a social network augmented with reputation ratings. We present an algorithm for inferring reputation ratings between individuals and demonstrate through experiments that it is accurate based on current data. We then integrate those scores into a mail application, TrustMail, to show how they may be used in combination with other techniques for sorting and filtering mail.