The classification of tea according to region of origin using pattern recognition techniques and trace metal data

Abstract

Trace metals in tea originating from various Asian and African countries were determined by using inductively coupled plasma-atomic emission spectrometry and inductively coupled plasma-mass spectrometry. Pattern recognition techniques were then used to classify the tea according to its geographical origin. Principal component analysis (PCA) and cluster analysis (CA), as exploratory techniques, and linear discriminant analysis (LDA) and soft independent modelling of class analogy (SIMCA), were used as classification procedures. In total, 17 elements (Al, Ba, Ca, Cd, Co, Cr, Cu, Cs, Mg, Mn, Ni, Pb, Rb, Sr, Ti, V, Zn) were determined in a range of 85 tea samples (36 samples from Asian countries, 18 samples from African countries, 24 commercial blends and seven samples of unknown origin). Natural groupings of the samples (Asian and African teas) were observed using PCA and CA (squared Euclidean distance between objects and Ward's method as clustering procedure). The application of LDA gave correct assignation percentages of 100.0% and 94.4% for the African and Asian teas, respectively, at a significance level of 5%. SIMCA offered percentages of 100.0% and 91.7% for African and Asian groups, respectively, at the same significance level. LDA, also at a significance level of 5%, allowed a 100% of correct case identification for the three classes China, India and Sri Lanka. However, a satisfactory classification using SIMCA was only obtained for the Chinese teas (100% of cases correctly classified), while teas from India and Sri Lanka appear to form the same class.