Abstract: In recent years, the prevalence of stress has become a significant public health concern, influencing both physical and mental well-being. This study examines the possibilities of applying techniques for ML to detect human stress based on sleeping habits. By leveraging data on sleep patterns, such as duration, quality, interruptions, and variability, we aim to develop a forecasting model that can precisely determine stress levels. We collected sleep data from a diverse group of participants using wearable devices and self-reported surveys over several weeks. A number of ML techniques, such as SVM, RF, and NNs, to create predictive models. The models' performance was assessed utilizing measures such as F1-score, recall, accuracy, and precision. Our findings demonstrate that Random Forests and Neural Networks outperform other algorithms in detecting stress from sleep data.

Keywords: SVM- Support Vector Machine, RF-Random Forest, NN-Neural Network, ML-Machine Learning

| DOI: 10.17148/IARJSET.2024.11766