A Feasibility Study on Noninvasive Blood Glucose Estimation Using Machine Learning Analysis of Near-Infrared Spectroscopy Data.

For GreenDoc under ABV, 850 nm emerged as the primary predictor ( β = −618.4, SE = 72.2, t = −8.56, p = 3.84 × 10 −13 , 95% CI: −759.9 to −476.8), whereas 970 nm exhibited only a weak trend ( β = −252.7, SE = 144.3, t = −1.75, p = 0.083).

also in 1,404 other papers

Furthermore, since 770 nm exhibited low explanatory power across all devices and, in some cases, did not reach statistical significance, it is advisable to exclude this wavelength from analysis or assign it a lower weighting in model development. 3.2.

also in 111,027 other papers