User Behavior Tracking and Bot Detection Using Machine Learning

This paper presents a machine learning-based approach for user behavior tracking and anomaly detection, integrated with an automatic logout mechanism to prevent unauthorized access. By continuously monitoring user interactions—such as mouse movements, scrolling speed, keyboard typing patterns, and click behavior—the system builds a behavioral profile for each session. Anomalies are detected using a trained Random Forest classifier to differentiate between human and bot-like activity in real time. When suspicious behavior is identified, the system triggers an automatic logout and sends an alert via email to ensure proactive protection. The behavioral data is collected using JavaScript on the client-side and transmitted securely to the Flask backend, where it is preprocessed and analyzed. Instead of relying on a single indicator, several patterns such as timing, navigation flow, and interaction intensity are combined to describe how a genuine user typically behaves. The model was trained on a diverse set including real user sessions and artificially generated bot activity, allowing it to distinguish between the two with better reliability. The system can also support more personalized user experiences by adapting as new behavior data is collected. The solution offers a practical way to counter automated attacks and session hijacking by providing continuous, intelligent monitoring.