Geographical origin identification of teas using UV-VIS spectroscopy

Abstract

In this work we proposed a method to verify the differentiating characteristics of simple tea infusions prepared in boiling water alone, which represents the final product as ingested by the consumers. For this purpose, total of 125 tea samples from different geographical provines of Vietnam have been analyzed in UV-Vis spectroscopy associated with multivariate statistical methods. Principal Component Analysis-Discriminant Analysis (PCA-DA), Partial Least Squares Discriminant Analysis (PLS-DA) and Artificial Neural Network (ANN) were compared to construct the identification model. The experimental results showed that the performance of ANN model was better than PCA-DA and PLS-DA model. The optimal ANN model was achieved when neuron numbers were 200, identification rate being 99% in the training set and 84% predition set. The proposed methodology provides a simpler, faster and more affordable classification of simple tea infusions, and can be used as an alternative approach to traditional tea quality evaluation.