FTIR/ATR Spectroscopy Wavelength Optimization in Analysis of Total Cholesterol in Human Serum

The wavelengths selection of Fourier transform infrared (FTIR) spectroscopy with attenuated total reflection (ATR) techniques was very important for improving model prediction effect, and reducing model complexity. Based on the difference of mean spectrum, a wavelengths combination selection method of FTIR/ATR analysis of total cholesterol in human serum was developed. 10 wavelengths with peaks were selected. All wavelength combinations of the 10 wavelengths were used to establish multiple linear regression (MLR) models respectively. According to the prediction effect, the optimal MLR model was the 5 wavelengths combination of 2926,1398,1302,1171, and 1086 (cm^-1), the corresponding root mean squared error of predication (RMSEP) and RP were 0.886 mmol/L and 0.705 respectively. And the prediction effect was close to one of partial least squares (PLS) model. These indicated that the wavelengths combination selection method could be applied to the FTIR/ATR analysis in human´s serum.