Tracing the global origins of black tea using rapid XRF techniques coupled with advanced machine learning

Having robust traceability of black tea is important to help prevent tea fraud. Developing rapid, accurate, environmentally friendly, and user-friendly methods to distinguish the geographical origins of black tea, is of great significance for safeguarding geographic indication (GI) products. In this study, the elemental contents of 791 authentic black tea samples from ten major tea-producing regions worldwide were quantified using X-ray fluorescence spectroscopy. The concentration of 15 elements in tea products was measured, and the characteristic elemental profiles for the ten GI regions were established. In addition, two unsupervised analysis techniques were used to visualize high-dimensional data, and six supervised models were employed to discriminate the ten GI regions. The results show that the machine learning models, including random forest, support vector machine, k-nearest neighbours, linear discriminate analysis, and the deep learning multilayer perceptron (MLP) model, demonstrated superior predictive capabilities compared to the traditional partial least squares discriminant analysis model, with the F1 score of identifying Assam tea improved from 66.1 % to a range of 87-97.7 %. The MLP model achieved the highest performance, with a 97.7 % overall F1 score in predicting the geographical origins of 532 authentic samples across ten GI regions. This research lays the foundation for establishing a comprehensive global black tea traceability system which has major implications for preventing tea fraud worldwide.