Published January 2013 | Version v1
Journal article

Application of PCA-LDA method to determine the geographical origin of tea based on determination of stable isotopes and multi-elements

  • 1. Institute of Quality and Standard for Agro-products, Zhejiang Academy of Agricultural Sciences, Hangzhou (China)
  • 2. College of Pharmacy, South-central University for Nationalities, Wuhan (China)
  • 3. Tea Research Institute, Chinese Academy of Agriculture Sciences, Hangzhou (China)
  • 4. Research Center of Agricultural Quality Standards and Testing Techniques, Henan Academy of Agricultural Sciences, Zhengzhou (China)

Description

The ratio of stable isotope and concentration of multi-element in tea was determinated with isotope ratio mass spectrometry (IRMS) and inductively coupled plasma mass spectrometry (ICP-MS). Pattern recognition techniques with principal component analysis (PCA) and linear discriminant analysis (LDA) were used to classify the geographical origins of tea from Fujian, Shandong and Zhejiang province, and Yuyao, Jinhua and Xihu region of Zhejiang. The results showed the values of δ15N, δ13C, δD, δ18O and the ratios of 206Pb/207Pb, 208Pb/206Pb and 87Sr/86Sr in tea samples were different from different origins. There was also large variable for the concentrations of 27 mineral elements, such as Li, Be, Na and so on, with a specific character of origin. The method of PCA could be used to classify the geographical origin of tea from different origins but with a cross in the scatter plot. However, PCA combining with LDA could gave correct assignation percentages of 99% for the tea samples among Fujian, Shandong and Zhejiang provinces, and 87% for the tea samples among Yuyao, Jinhua and Xihu region of Zhejiang. These results revealed that it was possible and feasible to classify the geographical origin of tea by the method of PCA-LDA based on the determination of isotopes and multi-elements. (authors)

Additional details

Publishing Information

Journal Title
Journal of Nuclear Agricultural Sciences
Journal Volume
27
Journal Issue
1
Journal Page Range
p. 47-55
ISSN
1000-8551

Optional Information

Notes
4 figs., 4 tabs., 31 refs.