Abstract:With the development of wearable technologies, a new kind of healthcare data has become valuable as medical information. These data provide meaningful information regarding an individual's physiological and psychological states, such as activity level, mood, stress, and cognitive health. These biomarkers are named digital since they are collected from digital devices integrated with various sensors. In this study, we explore digital biomarkers related to stress modality by examining data collected from mobile phones and smartwatches. We utilize machine learning techniques on the Tesserae dataset, precisely Random Forest, to extract stress biomarkers. Using feature selection techniques, we utilize weather, activity, heart rate (HR), stress, sleep, and location (work-home) measurements from wearables to determine the most important stress-related biomarkers. We believe we contribute to interpreting stress biomarkers with a high range of features from different devices. In addition, we classify the $5$ different stress levels with the most important features, and our results show that we can achieve $85\%$ overall class accuracy by adjusting class imbalance and adding extra features related to personality characteristics. We perform similar and even better results in recognizing stress states with digital biomarkers in a daily-life scenario targeting a higher number of classes compared to the related studies.

What problem does this paper attempt to address?

### Problems the paper attempts to solve This paper aims to extract digital biomarkers from multi - modal wearable device data for non - intrusive stress - state screening. Specifically, researchers use data from smartphones and smartwatches and machine - learning techniques (especially the random forest algorithm) to identify stress - related biomarkers. The following are the main objectives of this paper: 1. **Extract stress - related digital biomarkers**: - Researchers use multi - modal wearable device data (such as weather, activity, heart rate, stress, sleep, and location, etc.) to determine the biomarkers most related to stress levels. - Through feature - selection techniques, find out which modalities (such as sleep, activity level, etc.) are most important when classifying stress levels. 2. **Solve the class - imbalance problem**: - Since the number of samples at different stress levels is unbalanced, researchers use SMOTE (Synthetic Minority Over - sampling Technique) to solve this problem, thereby improving the classification performance of the model. 3. **Improve the accuracy of stress - level classification**: - By adjusting the class imbalance and adding additional parameters related to personality characteristics, researchers achieve an overall classification accuracy of 85%. - Compared with related research, they perform better in identifying stress states in daily life, especially when dealing with more categories of stress levels. ### Main contributions - **Extract new stress - related biomarkers**: Researchers extract new stress - related biomarkers from multi - modal data and verify the effectiveness of these biomarkers. - **Solve the class - imbalance problem**: By using the SMOTE technique, the model's ability to recognize stress levels that occur at low frequencies is improved. - **High - accuracy classification**: In classification tasks that include multiple stress levels, a relatively high overall classification accuracy is achieved. ### Conclusion This research provides a new perspective for understanding digital biomarkers of stress states, especially in the analysis of multi - modal data. By using machine - learning techniques, researchers successfully extract the biomarkers most related to stress levels and solve the class - imbalance problem, thereby improving the accuracy of stress - state classification. This provides an important reference for future non - intrusive stress monitoring in daily life.

Extracting Digital Biomarkers for Unobtrusive Stress State Screening from Multimodal Wearable Data

Stress Detection through Wrist-based Electrodermal Activity Monitoring and Machine Learning

Personalized Stress Monitoring using Wearable Sensors in Everyday Settings

Evaluating the Potential of Wearable Technology in Early Stress Detection: A Multimodal Approach

Classification of Stress via Ambulatory ECG and GSR Data

A Method for Stress Detection Using Empatica E4 Bracelet and Machine-Learning Techniques

Daily Stress Recognition from Mobile Phone Data, Weather Conditions and Individual Traits

StresSense: Real-Time detection of stress-displaying behaviors

Assessment of Mental, Emotional and Physical Stress through Analysis of Physiological Signals Using Smartphones

Multitask Learning for Mental Health: Depression, Anxiety, Stress (DAS) Using Wearables

Wearable sensor-based evaluation of psychosocial stress in patients with metabolic syndrome

Automatic Stress Detection in Working Environments from Smartphones' Accelerometer Data: A First Step

Evaluating Mental Stress Among College Students Using Heart Rate and Hand Acceleration Data Collected from Wearable Sensors

A machine-learning Approach for Stress Detection Using Wearable Sensors in Free-living Environments

Context-Aware Stress Monitoring using Wearable and Mobile Technologies in Everyday Settings

Stress Level Assessment by a Multi-Parametric Wearable Platform: Relevance of Different Physiological Signals

Assessing Mental Stress Based on Smartphone Sensing Data - An Empirical Study.

Emotion and Stress Recognition Utilizing Galvanic Skin Response and Wearable Technology: A Real-time Approach for Mental Health Care

Stress Monitoring Using Machine Learning, IoT and Wearable Sensors

Detection and Characterization of Physical Activity and Psychological Stress from Wristband Data

Mmstress: Distilling Human Stress from Daily Activities Via Contact-less Millimeter-wave Sensing