A Real-Time Personalized Stress Detection System Using Keystroke Dynamics
Author(s)
Abstract
Workplace stress has become an increasing concern in modern digital environments where individuals spend long hours interacting with computers. Continuous stress can negatively affect productivity, decision-making ability, and overall well-being. Traditional stress detection approaches often rely on website sensor or questionnaire-based assessments, which may be intrusive, expensive, or unsuitable for continuous monitoring. To address these limitations, this research proposes a real-time stress detection system based on keystroke dynamics and machine learning techniques. The proposed system analyses typing behaviour such as key hold duration, typing speed, latency between keystrokes, and error frequency to identify patterns related to stress conditions. These behavioral features are processed and used to train machine learning models capable of distinguishing between stressed and non- stressed states. A real -time monitoring dashboard is also developed to visualize stress indicator and provide meaningful insights into user behavior. The system operates without requiring additional hardware, making it a non-intrusive and scalable solution for stress monitoring. Experimental evaluation indicates that typing-based behavioral analysis can effectively detected stress patterns and support early identification of mental strain in computer-based work environments.
Keywords