Stress Annotations from Older Adults - Exploring the Foundations for Mobile ML-Based Health Assistance.

The number of new mobile and wearable technologies with built-in sensors for quantifying every aspect of our lives is increasing. Consequently, new data sources and opportunities arise for the development of machine learning (ML) models and their applications. In this paper, we report on a four weeks field study with 16 older adults, aged between 66 and 81 years (50% female), who were asked to provide stress-related experience samples in different modalities, including paper-based diaries and data collected with the help of a wearable (i.e., a Microsoft Band 2). We provide insights into participants' stress annotation behavior, report on a detailed analysis of the recorded data and the resulting implications regarding the annotation of stressful situations by older adults, discuss how mobile annotation technology can benefit from the synergies with traditional methods and argue why we believe that appropriate annotation techniques are the basis to benefit individually from future powerful machine learning models.