sleepWell:Stress Level Prediction Through Sleep Data. Are You Stressed?

Stress, sadness, and anxiety are now major issues in our regular lifystyle. The primary goal of this study is to detect stress in individuals using ML(Machine Learning) techniques with the ultimate goal of enhancing their quality of day to day life. We use a publicly available multimodal dataset SayoPillow, to propose alternative machine learning models for the identification of stress on humans while sleeping. Many physiological factors can now be continuously monitored because of advancements in wearable technology. The importance of stress detection has grown in recent years, mostly because early stress detection can help people better manage their health and reduce the harmful effects of sustained stress exposure. This study helps to select best machine learning models to classify level of stress while someone is sleeping. In our study we have implemented several machine learning-based forecasting models the level of stress of individuals. In order to offer a comprehensive review of the approaches utilized in this domain, we first identify sleep patterns and the key components of sleep evaluation. We examine the main technical developments in the fields of sleep behavior analysis, data gathering, and monitoring, highlighting their advantages and disadvantages. Finally, we also talk about potential research difficulties and directions. The F1-score, recall, and accuracy were determined and compared using ML models. In our study the SVM and Gaussian Naive Bayes (GNB) model outperformed rival models with F1-scores of 1.0.