The goal of our study is detecting daily level of psychological stress by using data obtained from wearable sensors (ex. smart watch/bracelet) and smartphones. For evaluating our apprach data from stress-level questionnaires will be obtained. The study can be viwed as 2 step process: data colecting and stresss modeling.
In the data collecting process wearable sensors and a mobile application will be used to collect the following data:
1. Data directly related to stress level provided by users who will have to:
1.1. fill in a self-assessment questionnaire,
1.2. record random audio message,
1.3. answer questions about caffeine/alcoholic drinks and
1.4. answer questions about the quality of sleep.
2. Data from all the available smartphone sensors.
3. Data about the phone usage
4. Basic static information about the user
5. Wearable sensor data: Heart rate, skin conductance, skin temperature…
In the stress modelling process machine learning methods will be used to produce models that will recognize the level of psychological stress that the subject is experiencing. The machine learning models will use only data from the wearable sensors and the smartphone sensors. The data obtained with the questionnaires will be used as a ground truth for evaluating our models.