中国农业科学 ›› 2011, Vol. 44 ›› Issue (22): 4653-4659.doi: 10.3864/j.issn.0578-1752.2011.22.012

• 贮藏·保鲜·加工 • 上一篇    下一篇

基于矿质元素的苦荞产地判别研究

张强, 李艳琴   

  1. 1.山西师范大学生命科学学院,山西临汾 041004
    2.山西大学生物技术研究所/化学生物学与分子工程教育部重点实验室,山西太原 030006
  • 收稿日期:2011-04-14 出版日期:2011-11-15 发布日期:2011-09-09
  • 通讯作者: 通信作者李艳琴,Tel:0351-7016679;E-mail:yanqin@sxu.edu.cn
  • 作者简介:张 强,Tel:0357-2051196;Fax:0357-2051197; E-mail:zqlf@163.com
  • 基金资助:

    国家自然科学基金项目(30771300)

The Origin Discrimination of Tartary Buckwheat Based on the Mineral Elements

 ZHANG  Qiang, LI  Yan-Qin   

  1. 1.山西师范大学生命科学学院,山西临汾 041004
    2.山西大学生物技术研究所/化学生物学与分子工程教育部重点实验室,山西太原 030006
  • Received:2011-04-14 Online:2011-11-15 Published:2011-09-09

摘要: 【目的】分析不同省份苦荞矿质元素的特点,筛选判别苦荞产地的有效指标,同时探索苦荞产地溯源和判别的方法。【方法】对苦荞主产区山西、甘肃、青海、四川和云南省的39个苦荞品种中的7种矿质元素Cu、Zn、Fe、Mn、Ca、P和Se的含量进行统计分析,在对7种矿质元素进行逐步筛选的基础上,应用非参数判别的K最近邻法进行判别分析。【结果】不同省份苦荞品种的矿质元素含量存在不同程度的差异,云南苦荞Cu、P含量最高;山西苦荞Se含量最高;青海苦荞Zn、Fe、Ca含量最高;四川苦荞Cu、Zn、Fe、Ca、P含量均最低,Mn和Se含量也较低;甘肃苦荞Mn含量最低。Se、Mn、Zn、Ca和P对苦荞分类有极显著影响,Fe和Cu对苦荞判别影响不显著;判别结果回判正确率和交互验证正确率均为97.4%。【结论】矿质元素含量和现代统计技术相结合的判别方法,用于苦荞产地溯源和判别是有效、可行的。

关键词: 苦荞, 矿质元素, 判别分析, K最近邻法, 数据挖掘

Abstract: 【Objective】The characteristics of mineral elements in tartary buckwheat from different provinces were analyzed in order to make choice of the effective index in tartary buckwheat origin discrimination, and to explore the method of tartary buckwheat origin traceability and discrimination.【Method】Seven mineral element contents (Cu, Zn, Fe, Mn, Ca, P and Se) were analyzed in thirty-nine samples from five provinces of China(i.e., Shanxi, Gansu,Qinghai,Sichuan and Yunnan). On the basis of stepwise selection,the K Nearest Neighbor (KNN) analysis of nonparametric discriminant was applied to data analysis.【Result】The results showed that the mineral element contents were different in tartary buckwheat varieties from different provinces. In conclusion, element contents of Cu and P were the highest in the Tartary buckwheat varieties from Yunnan while Se was the highest in those from Shanxi. However, the contents of Zn,Fe and Ca were the highest in those from Qinghai, while Cu, Zn, Fe, Ca and P were the least in those from Sichuan, and Mn was the least in those from Gansu. Then, mineral element character indexes (Zn, Mn, Ca, P and Se) affecting the classification greatly were selected with discrimination analysis. A percentage of 97.4% of correct classification was achieved by resubstitution and cross-validated.【Conclusion】The determination of mineral element content in combination with modern statistical techniques should be a useful and convenient tool for the origin discrimination and standardization of tartary buckwheat.

Key words: tartary buckwheat, mineral element, discriminant analysis, K Nearest Neighbor, data mining