Mitigating Cybersecurity Risk of Threat Actors Through Dark Web Browser Fingerprinting

Cybersecurity is crucial in mitigating illegal access and hacking of online resources and network systems, and with the protection of network traffic vital by various detection and prevention techniques against cyber-attacks. Internet communication that occurs on the Dark Web in particular, contains sensitive data that should remain inaccessible to unauthorized users. However, threat actors are empowered by anonymous browsing on the Dark Web. This paper investigates browser fingerprint attacks on the Dark Web, focusing on identifying attack patterns. A detection model based on traditional machine learning techniques is developed using a large sample of fingerprint data. Feature selection techniques are applied to create optimized training and test datasets for classification. The model is validated using cross-validation methods to minimize overfitting and underfitting. Empirical results demonstrate that the proposed model effectively identifies malicious browsers on the Dark Web.