Improving hypoglycemia prediction by benchmarking heart rate data quality using impute paradigm
Publication date
2026-08
Document type
Konferenzposter
Organisational unit
Conference
48th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC 2026) ; Toronto, Canada ; July 26–30, 2026
Publisher
ResearchGate
Part of the university bibliography
✅
Language
English
Description
This poster was presented at the IEEE Engineering in Medicine and Biology Society (EMBC 2026). The study investigates how improving the quality and completeness of heart-rate (HR) data can enhance short-term hypoglycemia prediction. We propose a modular imputation framework that systematically evaluates statistical, polynomial, and machine-learning approaches for imputing missing HR data across varying gap durations. Applied to the SubDiadata II dataset, the framework improves HR data completeness and, in turn, enhances hypoglycemia prediction performance using XGBoost and LightGBM.
Version
Published version
Access right on openHSU
Metadata only access