User Authentication by Service Utilisation Profiling

In the age of the information society, computer security has become of vital importance in order to maintain the integrity of a system. Many organisations are enforcing new IT policies calling for better security in an attempt to safeguard against misuse of resources and data. These IT polices are beginning to utilise a range of technologies to provide front-line authentication of the user, including fingerprint scanning and iris recognition. These technologies however only really assist in providing improved point-of-entry security. This paper presents a novel authentication technique that is able to provide transparent and continuous authentication of the user based upon the way in which a user interacts with their computer – which applications a person uses and when, and which websites a user browses for instance. Through the application of neural networks, results are very encouraging with an overall Equal Error Rate (EER) of 7%