Enhancing User Authentication through Keystroke Dynamics Analysis using Isolation Forest algorithm

Authentication in modern era, has evolved significantly to address the increasing complexity and security challenges of our digital world. Traditional methods of authentication, such as passwords and PINS have proven to be vulnerable to various forms of attacks. As a result there is a growing need for innovative and robust authentication techniques. Keystroke dynamics involves the analysis of user typing patterns, including the time interval between keystrokes and the key pressed duration, to establish a unique biometric profile for each individual. Unlike traditional authentication methods keystroke dynamics offer several advantages, such as continuous authentication without the need for additional hardware and the ability to adapt to a user's changing typing behavior. This model will additionally analyze users typing patterns have the capability to identify individuals using the system by considering their typing speed. The first premise involves developing a system that can capture the keystroke dynamics of a user during a training phase and create a unique profile for that user. The second premise involves implementing Mahalanobis distance for Dynamic keystroke Authentication with various machine learning algorithms. This model can maximize the strengths of both methods to improve the accuracy and robustness of the authentication system. Authentication using keystroke dynamics is secured and convenient method and can be used wherever there is requirement of authentication such as baking, e-commerce and secured access control system. From the experiment performed below we can attained that the model functions exceptionally well using Isolation Forest with an accuracy of 90%.