User and group profiling based on user process usage

Abstract

User profiling based on process usage is on approach for adding an extra security layer to our computer systems. In addition it can be of great value for classification of a company/school network and the their user groups. Groups, or classes of users, in a company might belong to the same division or department that solve similar tasks. In a company, accountants probably use the same set of tools, as would a group of students in a graphic design class. Studying if these similarities in the process that they use, can say something about an individual or a group. It is valuable in the terms of analyzing individual user and group behavior. Recognizing individual users behavior, provides the possibility of an extra layer of security in the form of an authentication scheme. Recognizing group behavior might provide valuable insight when it comes to building a profile for a new user, and see why this user fits the group or not.

This thesis makes use of statistical approaches to discuss the possibility of using process profiling to classify users into groups.