Short-Term Interval Prediction of Glucose with Probabilistic Models
Sprache des Titels:
Diabetes Technology & Therapeutics
Prediction of future glucose using continuous glucose monitoring (CGM) data is an active area of research and many predictors have been proposed. An inherent difficulty is the high variability associated with unknown or immeasurable influence factors. The approach pro-posed here utilizes Gaussian mixture models to predict a range of future glucose levels, tak-ing into consideration their specific probability distribution.