Construction and Improvement of a Model for Quantifying Blood Glucose Concentration Using Mid‐Infrared Spectroscopy

ABSTRACT Measurements of blood glucose concentration use invasive methods such as venous blood sampling, and finger‐prick blood testing using self‐monitoring blood glucose meters with subcutaneous sensors. For daily use, the development of noninvasive blood glucose measurement methods is required. In this study, we constructed a model to estimate blood glucose concentrations noninvasively from mid‐infrared absorption spectra measured using photothermal deflectometry enhanced by total internal reflection. We improved the estimation accuracy of the model using Savitzky–Golay preprocessing and the Boruta variable selection method. In addition, the model was corrected using subject data from the first day of measurements to improve estimation accuracy.