Ambient Stress Detection for Open Environment Using Hybrid Model

Stress is an omnipresent aspect of daily existence that confronts individuals on diverse occasions. Nevertheless, enduring prolonged or intense stress can jeopardize our well-being and impede the rhythm of our daily routines. Early detection of mental stress symptoms can prevent many health issues associated with stress. Our study aims to employ advanced machine learning and image-processing techniques to identify stress in open environments. Distinguishing itself from previous stress detection systems, our upgraded system is implemented through web applications. This system conducts periodic analyses of employees, assessing both physical and mental stress levels, and provides appropriate remedies. This includes the distribution of periodic survey forms. Utilizing the hybrid model (Random Forest classifier and Support vector machine) for stress detection, our primary focus is on creating a healthy and spontaneous working environment for all individuals in the working sector, aiming to optimize their productivity. The end users of our system are people and other working employees, who benefit from personalized stress analyses and remedies, as well as insights for creating a healthier and more productive work environment. During our study, we separately diagnosed the Support Vector model and Random Forest classifier, but our combined model achieved an accuracy of up to 98.12%.