PATTERN-RECOGNITION APPLIED TO GAS-CHROMATOGRAPHIC PROFILES OF VOLATILE COMPONENTS IN 3 TEA CATEGORIES

Volatile components in unfermented green tea, semi-fermented Oolong te a and fully fermented black tea were analyzed by gas chromatography (G C) and gas chromatography-mass spectrometry (CC-MS). For differentiati ng, three tea categories based on their volatile components, unsupervi sed and supervised pattern recognition techniques were applied to the resulting GC data. Three distinct clusters each corresponding to green tea, Oolong tea and black tea were observed in the dendrogram and the principal component (PC) score plot. However, a subcluster of Oolong tea was observed in the vicinity of black tea cluster in both the dend rogram and the PC plot. The first and second PC corresponded to the fe rmentation products and aroma components originally contained in tea l eaves, respectively. Both the partial least squares (PLS) analysis and linear discriminant analysis correctly differentiated tea samples int o the three categories. (E)-2-Hexenal, the major fermentation product from unsaturated fatty acids, was the most efficient for the discrimin ation. Although three teas are produced from the same plant species, p attern recognition clarified the existence of the apparent quality dif ference among their volatile component profiles.