Efficiency Issues and Improving Implementation of Keystroke Biometric Systems

Abstract

Keystroke dynamics is a heavy field for researches; a lot of solutions have been proposed in this domain using different implementations usually based on Euclidean distance for measuring similarity between features vectors. However, the Euclidean distance method has a higher error equal rate compared with other classification methods, which makes the method less effective. Therefore, in the article, the authors propose their version of keystroke dynamics implementation based on K-NN, F-NN, and Manhattan distance as classifiers to improve the authentication efficiency. The flight times and dwell time between keys are used in this study.