Result Paper on Mouse Dynamics Based Bot Detection System

Abstract

With the increasing number of automated attacks on web applications, traditional security mechanisms are becoming insufficient. Bots can bypass CAPTCHA and mimic human-like interactions, making detection challenging. This project proposes a Mouse Dynamics Based Bot Detection System to distinguish human users from automated scripts. The system captures real-time mouse movement data such as position, velocity, and acceleration. Behavioral features are extracted and analyzed using machine learning techniques. The system evaluates patterns like movement randomness, speed variation, and interaction duration. A web-based interface integrates data collection and real-time prediction. The system provides immediate classification results with confidence scores. It enhances security by identifying abnormal interaction patterns. The approach is lightweight and does not interrupt user experience. Overall, it improves authentication systems by adding behavioral biometrics.