Security Aspects of Behavioral Biometrics for Strong User Authentication
Vladimir Jancok, Michal Ries · 2022
Efficient user identification and authentication are fundamental for securing access to systems processing sensitive data. This paper provides an analysis of current research in the field of user identification and identity verification with a focus on the behavioral biometrics supported by Machine Learning. It identifies the methods for user modeling with the potential of application in real-world scenarios such as strong authentication and fraud detection domain. This paper further elaborates on the current state-of-the-art approaches, feature extraction, and classification methods. We describe our experimental setup and provide an evaluation of our method in the selected deployment. We focus on user interactions in a controlled web environment. We performed classification experiments with the machine learning models on various datasets showing promising results in the robustness and proving relevance as a modern non-intrusive security measure.