Related Experiment Video
Updated: Jan 11, 2026
An Application for Pairing with Wearable Devices to Monitor Personal Health Status
Published on: February 3, 2022
Predicting clusters of physical activity based on individual characteristics: an event-based ecological momentary
Maya Braun1,2, Geert Crombez1, Dries Debeer1
1Department of Experimental Clinical and Health Psychology, Ghent University, Ghent, Belgium.
None:
Objective: Planning can help bridge the physical activity intention behaviour gap, but creating plans has proven burdensome for individuals. Personalised plan recommendations can alleviate this burden and improve plan quality. This study aimed to identify clusters of physical activities and predict these clusters based on static and dynamic person variables to identify relevant variables for personalising recommendations.
None:
Design: In a 14-day ecological momentary (EMA) assessment study, 52 participants completed surveys on static (e.g. gender, SES) and dynamic (e.g. mood) personal characteristics at baseline, each morning, and each time they were active for at least 5 min. Clusters of physical activities were identified based on activity (e.g. intensity) and context (e.g. location) characteristics. Clusters, activity domain, and location were predicted using both a conditional random forest algorithm approach and a multilevel multinomial regression approach.
None:
Results: A five-cluster solution was identified, with clusters being 'active transport activities', 'work-related activities', 'household activities', 'organised sport activities', and 'in the city activities'. We predicted up to 65% of clusters, exceeding baseline comparison. Predictions were largely based on static characteristics, most notably sociodemographic information.
None:
Conclusions: Our findings imply that personalised recommendations do not require daily assessment-though there is a risk of stereotyping. Future work should integrate passively collected data and evaluate different way of creating recommendations.
