Detecting of Robotic Imitation of Human on-the-Website Activity With Advanced Vector Analysis and Fractional Derivatives

Abstract

This paper introduces a novel approach for the detection of automated entities in online environments through the analysis of mouse dynamics. Leveraging fractional derivatives and vector cross products, our methodology scrutinizes the intricate patterns embedded in mouse movements. Fractional derivatives capture the non-integer order dynamics, while vector cross products reveal deviations from expected trajectories. The combination of these advanced mathematical tools offers a unique perspective on distinguishing between human and bot behaviors. We present experimental results showcasing the efficacy of our approach in various scenarios, demonstrating its potential in the realm of cybersecurity and online integrity. Our findings contribute to the evolving landscape of bot detection methodologies, emphasizing the importance of incorporating mathematical rigor in the analysis of digital behavior.

Keywords

• Bot detection
• human-bot interaction
• clusterization
• feature engineering
• feature interaction