Machine Learning and Browser Fingerprinting Based Approach for Web Bot Detection

Web bots are becoming increasingly sophisticated, posing a serious threat to internet security and integrity. In response to this growing issue, this study proposes a novel bot detection method based on browser fingerprinting and machine learning. The proposed method is a model that was trained on browser fingerprints collected from web traffic using a custom script with a browser fingerprinting library. The traffic was classified as legitimate or malicious based on firewall rules and honeypots. This model seeks to enhance bot detection accuracy while decreasing false positives, filling a gap in the existing research on browser fingerprint-based bot detection. The experimental results demonstrate that detailed browser fingerprint data can serve as an effective indicator for distinguishing between human and automated traffic, providing a foundation for integrating browser fingerprinting into wider security measures. The study also analyzes the specific attributes on browser fingerprints that contribute most significantly to differentiating between legitimate and bot traffic, while acknowledging the limitations in the proposed method.