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 par...