An Efficient Feature Selection Technique for User Authentication using Keystroke Dynamics

Securing the sensitive data and computer systems by allowing ease access to authenticated users an d withstanding the attacks of imposters is one of the major challenges in the field of computer security. ID and password are the most widely used method for authenticating the computer systems. But, this method has many loop holes such as password sharing, shoulder surfing, brute force attack, dictionary dttack, guessing, phishing and many more. Keystroke Dynamics is one of the famous and inexpensive behavioral biometric technologies, which identifies the authenticity of a user when the user is working via a keyboard. Keystroke features like dwell time, flight time, di-graph, tri-graph and virtual key force of every user are used in this paper. For the purpose of preprocessing Z-Score method is used. Ant Colony Optimization (ACO), Particle Swarm Optimization (PSO), Genetic Algorithm (GA) algorithm is used with Extreme Learning Machine (ELM) for feature subset selection. In order to classify the obtained results ELM algorithm is used. Comparison of ACO, PSO and GA with ELM respectively is done to find the best method for feature subset selection. From the results, it is revealed that ACO with ELM is best for feature subset selection.