User Authentication Based On Keystroke Dynamics

The most common way to enforce access control is user authentication based on username and password. This form of access control has many flaws which make it vulnerable to hacking. Biometric authentication such as the keystroke dynamics is used in which the keyboard is used in order to identify users. Then the classifier is tailored to each user to find out whether the given user is genuine or not. The contribution of this approach is twofold: first it reduces the possibility of over fitting second it allows scalability to a high volume of users. Here, measured mean, median values, and standard deviation of keystroke features such as latency, dwell time, digraph and their combination are used. The algorithms used for feature subset selection are Particle Swarm Optimization (PSO), Genetic Algorithm (GA), Ant Colony Optimization (ACO) and the proposed Renovated Artificial Bee Colony Optimization (RABCO) algorithm. Back Propagation Neural Network (BPNN) is used for classification. Keywords— Keystroke Dynamics, Feature Extraction, Feature subset selection, Artificial Bee Colony Optimization, Back Propagation Neural Network.