Machine Learning Based Assessment of Mental Stress using Wearable Sensor Data

Mental stress has become a growing concern due to its adverse impact on both physical and mental health. The ability to detect and assess stress levels in real-time can aid in early intervention and stress management. In this study, we propose a machine learning-based approach for the assessment of mental stress using wearable sensor data. Data from wearable sensors, which capture physiological signals such as heart rate, Electrodermal Activity, and TEMP, are utilized to predict stress levels. We implement various machine learning algorithms including Decision Trees (DT), Random Forest (RF), CatBoost, Logistic Regression, and XGBoost,LSTM,DNN,FNN  to classify stress and non-stress states. The performance of these algorithms is evaluated and compared using key metrics such as accuracy, precision, recall, and F1-score. The results demonstrate the effectiveness of machine learning techniques in stress detection, highlighting the potential of wearable sensor technology in continuous, non-invasive stress monitoring. This research provides a foundation for the development of intelligent systems aimed at improving mental health by facilitating timely stress management interventions.

Keywords: Mental stress detection, wearable sensors, physiological signals, machine learning, stress classification, non-invasive monitoring.