Intrusion detection using keystroke dynamics & fuzzy logic membership functions

If the password is compromised, either due it being weak or someone getting to know it through other means, the system cannot detect it. To overcome this problem, we propose a system whereby the system can detect whether the current user is the authorized user, a substitute user or an intruder pretending to be a valid user. Therefore the system checks the identity of the user by their behaviour pattern using keystrokes dynamics to authenticate user. A number of samples of login and password attempts of each user is gathered and stored in a database. From the samples collected, keystroke patterns are derived called feature sets and signatures are formed for each user using Fuzzy Logic algorithms. Once signatures are formed, users are authenticated by comparing their typing pattern to the respective signatures formed. We study the performance of such a system based on features like False Acceptance Rate (FAR) and False Rejection Rate (FRR), thus evaluating the efficiency of the system.