Machine learning-based glucose prediction with use of continuous glucose and physical activity monitoring data: The Maastricht Study.
William P T M van DoornYuri D ForemanNicolaas C SchaperHans H C M SavelbergAnnemarie KosterCarla J H van der KallenAnke WesseliusMiranda T SchramRonald M A HenryPieter C DagnelieBastiaan E de GalanOtto BekersCoen D A StehouwerSteven J R MeexMartijn C G J BrouwersPublished in: PloS one (2021)
Machine learning-based models are able to accurately and safely predict glucose values at 15- and 60-minute intervals based on CGM data only. Future research should further optimize the models for implementation in closed-loop insulin delivery systems.