Abstract: Pattern recognition techniques were studied to classify the Pu’er tea on the basis of their geographical origin. In this experiment, 85 samples of Pu’er tea in three mainly origin regions (Xishuangbanna, Pu’er and Linchang city) were collected as experimental samples. 16 heavy rare-earth elements in Pu’er tea samples were determined by using inductively couple plasma-mass spectrometry (ICP-MS). Principal component analysis (PCA) and discriminant analysis of Wilk’s method were used as classification procedures. The recognition accuracy for the training samples (Pu’er shu tea and Pu’er sheng tea) on three origin regions were all 100%, cross-validating were 100%. The recognition accuracy for the training samples (mixture of Pu’er shu tea and Pu’er sheng tea) on Xishuangbannan, Pu’er and Linchang city set was 93.3%, 87.9%, 90.9 %, respectively; cross-validating were 90%, 87.9%, 90.9%, respectively. The recognition accuracy for the testing 18 samples were 84.6 %、81.4%、82.9% respectively. Those results revealed that it was possible to distinguish three mainly origin regions of Pu’er tea by the contents of heavy rare-earth elements.

Key words: heavy rare earth element, Pu'er tea, discrimination of region origin, ICP/MS