Machine Learning Approaches to Automatic Stress Detection: A Review

People experience mental stress on a daily basis from a variety of different reasons, including environmental reasons (traffic, noise, or bad weather), social reasons (family issues, friends, and financial problems), or from events such as wedding planning or giving a presentation in front of large audience. A manageable amount of stress is healthy and can motivate a person; however, a large amount of continuous stress or a strong response to stress can be harmful. For this reason, the detection of mental stress, as well as its prediction, has become a significant area of research. In this paper, we review and summarize various approaches found in the literature for stress detection using machine learning and suggest directions for future research and interventions.