CONTINOUS USER AUTHENTICATION USING BEHAVIORAL BIOMETRICS FOR ENHNACED WEB APPLICATION SECURITY

This paper presented a real-time continuous authentication framework designed to enhance system security through be-havioral monitoring and intelligent anomaly detection. The proposed system integrates behavioral feature extraction, session-based monitoring, Isolation Forest–based anomaly de-tection, and risk-based adaptive authentication within a unified modular architecture. By combining behavioral biometrics with machine learning techniques, the system achieves robust user verification and anomaly detection in dynamic usage environments. The integration of Isolation Forest for unsupervised anomaly detection enhances reliability in identifying deviations from normal user behavior, while the adaptive authentication mod-ule enables secure and efficient responses through risk-based OTP verification. Experimental evaluations demonstrate that the proposed framework operates efficiently in real time with minimal computational overhead, highlighting its practical feasibility for scalable and lightweight deployment. Furthermore, the inclusion of session-based data storage and periodic retraining enables the system to adapt to evolving user behavior over time, improving long-term accuracy and robustness. Overall, the proposed architecture establishes a scalable foundation for continuous authentication systems that combine cybersecurity, machine learning, and behavioral biometrics, contributing toward enhanced security and reduced reliance on traditional static authentication mechanisms in modern digital environments.

[1]     “AI-powered anomaly detection in zero trust environments: A compre-hensive review of methods and evaluation,”K. Sandhu, S. G. R. Bojja, S. Venkataramanan, V. K. R. Vangoor, and S. Thota, Nanotechnology Perceptions, vol. 20, no. S16, pp. 2634–2654, 2024. [2]     “A generic privacy-preserving protocol for keystroke dynamics-based continuous authentication”,A. F. Baig and S. Eskeland,in Proc. 19th Int. Conf. Security and Cryptography (SECRYPT), 2022, pp. 491–498, doi: 10.5220/0011141400003283. [3]     “Continuous authentication using mouse dynamics,”S. Mondal and P. Bours, in Proc. Int. Conf. on Biometrics and Kansei Engineering (ICBAKE), Gjøvik, Norway, pp. 1–8. [4]     “Behavioral biometrics: A powerful defense against social engineer-ing attacks,”K. V. K. Pasumarthi, Journal of Computer Science and Technology Studies, vol. 7, no. 8, pp. 1166–1173, Aug. 2025, doi: 10.32996/jcsts.2025.7.8.132. [5]     “AI-powered zero trust access evaluation using behavioral fingerprint-ing,”H. Yadav, G. Subramanian Ramachandran, and K. Sharma, Inter-national Journal of Computer Applications, vol. 187, no. 16, June 2025. [6]     “Privacy-preserving continuous authentication using behavioral biomet-rics,”, A. F. Baig, S. Eskeland, and B. Yang, International Journal of Information Security, vol. 22, no. 6, pp.1833–1847, Jul. 2023, doi: 10.1007/s10207-023-00721-y. [7]     “User authentication method based on keystroke dynamics and mouse dynamics with scene-irrelated features in hybrid scenes,” X. Wang, Y. Shi, K. Zheng, Y. Zhang, W. Hong, and S. Cao,Sensors, vol. 22, no. 17, p. 6627, Sep. 2022, doi:10.3390/s22176627. [8]     ”Behavioral biometrics for continuous authentication”,A. Gunuganti,Journal of Biosensors and Bioelectronics Research, vol. 1, no. 3, pp. 1–5, 2023, doi: 10.47363/JBBER/2023(1)122. [9]     “Continuous Authentication and Behavioral Biometrics: Enhancing Cybersecurity in the Digital Era,” S. Oduri, International Jour-nal of Innovative Research in Science, Engineering and Technol-ogy (IJIRSET), vol. 13, no. 7, July 2024, pp. 13632–13640, doi: 10.15680/IJIRSET.2024.1307140. [10]    “ Zero-Trust Network Access (ZTNA)”,V. Mavroudis,Alan Turing In-stitute, 2024.