Identification of green tea varieties and fast quantification of total polyphenols by near-infrared spectroscopy and ultraviolet-visible spectroscopy with chemometric algorithms
In this study, an approach based on near-infrared spectroscopy (NIRS), ultraviolet-visible spectroscopy (UV-Vis) and chemometric algorithms was developed for discrimination among five varieties of green tea, and further estimation of the total polyphenol content (TPC) in these tea varieties. Principal component analysis (PCA) and the random forest (RF) pattern recognition technique were used to classify these samples. Based on the joint information from the NIR and UV-Vis spectra, a successful classification model was established with RF. The classification accuracy was 96%. Furthermore, a partial least-squares regression (PLSR) model based on the NIR spectra and TPC values measured by the UV-Vis reference method was constructed for rapid analysis of the TPC in these tea samples. The values of RMSECV, RMSEC, and RMSEP were 0.3578, 0.1775 and 0.2693, respectively. The correction coefficients for the calibration and prediction set were 0.9966 and 0.9864, respectively. These results demonstrated that the proposed method can be efficiently utilized for fast, accurate, economic analysis of green tea.