Abstract:
Stress is natural, especially during this unprecedented COVID-19 crisis that has brought various
emotions and challenges. However, stress experienced over an extended period can lead to
serious health problems. It is therefore important to timely detect and overcome it. Studies conducted in the past have shown the signi_cance of Electrodermal Activity (EDA) and Heart Rate
Variability (HRV) in stress detection. We present a digital solution that involves both stress
prediction and mitigation with the help of an Android Application. We use incremental learning
to personalize the machine learning model that predicts stress arousal using HRV indices and
EDA measurements collected continuously via a wearable device. If the user is found stressed
at any moment, the application provides personalized recommendations for a stress-relieving
activity. To provide personalized recommendations, we use a clustering algorithm preceded by
Thompson Sampling to determine user activity preferences in the cold start phase.
