Hybrid method for active authentication using keystroke dynamics

Abstract

Nowadays there is a growing interest to active authentication methods in security society. These methods are used for user identity validation with behavioral biometrics such as keystroke or mouse moving dynamics. In this article a new hybrid method for active authentication using keystroke dynamics is presented. This method is a combination of a new keystroke data representation model based on potential functions and machine learning algorithms based on decision trees. The proposed method is tested in static and dynamic authentication scenarios on the benchmark Si6 dataset and datasets collected by the authors. The experimental results confirm that the proposed hybrid method is applicable to reallife authentication systems.