Wavenumbers Combination Optimization for NIRS Analysis of Glucose in Human Serum

Abstract

The wavenumbers combination selection of near infrared spectroscopy (NIRS) analysis was very important for improving model prediction effect, reducing model complexity and designing special NIRS instruments with high signal noise ratio. Based on the prediction effect of single wavenumber linear regression model, a wavenumbers combination selection method of NIRS analysis of glucose in human serum was developed. 25 wavenumbers with good prediction effect were selected. All wavenumber combinations of the twenty-five wavenumbers were used to establish multiple linear regression (MLR) models respectively. According to the prediction effect, the optimal MLR model was the eleven wavenumbers combination of 7340, 7328, 7311, 7253, 7251, 7234, 7228, 7220, 7218, 7207, 7203 (cm <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">-1</sup> ), the corresponding root mean squared error of predication (RMSEP) was 0.384 mmol/L. And the prediction effect was obvious better than one of partial least squares (PLS) model. These indicated that the wavenumbers combination selection method based on the prediction effect of single wavenumber linear regression model could be applied to the NIRS analysis and could provide valuable reference for designing minitype special NIRS instruments.