About this distinction
The Euclid team won 1st place in the IFMBE Scientific Challenge 2022, ahead of more than 20 competing teams, using machine learning and oversampling techniques.
The challenge
Active and Healthy Ageing (AHA) apps only help older adults if people keep using them. The task was to learn the patterns of user activity in an AHA mobile application and identify early dropouts: given a window of consecutive scheduled measurements and static user features such as demographics, predict whether the user will keep up with the next three scheduled measurements.
Results
Euclid (Democritus University of Thrace & ATHENA Research Center) ranked first with a score of 0.8635, ahead of Ubon Ratchathani University (0.7970) and the University of Valladolid (0.7500). The winning work, “Predicting Early Dropouts of an Active and Healthy Ageing App”, was presented at the IUPESM World Congress on Medical Physics and Biomedical Engineering 2022 (Singapore, 12–17 June 2022), and led to follow-up research on federated learning for dropout prediction in healthy-ageing and elderly social care applications.