Web user profiling based on browsing behavior analysis

Abstract

Determining the source of criminal activity requires a reliable means to estimate a criminal’s identity. One way to do this is to use web browsing history to build a profile of an anonymous user. Since an individual’s web use is unique, matching the web use profile to known samples provides a means to identify an unknown user. This paper describes a model for web user profiling and identification. Two aspects of browsing behavior are examined to construct a user profile, the user’s page view number and page view time for each domain. Four weighting models, based on the term frequency and term frequency – inverse document frequency weighting schemes, are proposed and compared. Experiments involving 51 personal computers demonstrate that the profiling model is very effective at identifying web users.