Stage: Retrospective sensor-data research
A Korean study published on August 17, 2026 explores machine learning for classifying voiding patterns from smart-diaper sensor signals. Researchers analysed data collected in 2022 from 97 older residents at three long-term care facilities.
The method uses learning from unlabelled time-series data to reduce dependence on manually labelled examples. Evaluation separated participants across cross-validation groups. This retrospective analysis does not establish diagnostic performance in a new patient population. Authors include researchers affiliated with HYGERA Networks.