| In this study, the feasible use of near-infrared (NIR) spectroscopy for determination of β-phenylethanol content in Chinese rice wine was investigated. A total of 112 rice wine samples of various brands and wine ages were collected. Regression analysis model was developed in range of 800 -2 500 nm. The content of β-phenylethanol was determinated by gas chromatography (GC) as reference value. The result showed that the optimal model was obtained using partial least squares regression (PLSR) combined with full spectral wavelength range and standard normal variate (SNV) pretreatment. The calibration correlation coefficient (rc) was 0.949, RMSEC (value of the root -mean -square error of calibration) to 6.66 mg/L, RMSEP (value of the root-mean-square error of prediction) to 10.4 mg/L, RPD (valve of the residual predictive deviation) to 1.83. The result demonstrated that NIR might provide an alternative rapid technique for the screening and detection of β-phenylethanol content in Chinese rice wine. However, more samples should be incorporated in further in order to build a more accurate and robust model. |
