欢迎访问《中国粮油学报》编辑部网站！

张强,李艳琴.基于氨基酸特征的荞麦Fisher判别[J].中国粮油学报,2015,30(2):26-32

基于氨基酸特征的荞麦Fisher判别

The Fisher Discriminant of Buckwheat Based on the Amino Acids

投稿时间：2013-11-05 修订日期：2014-01-15

DOI：

英文关键词:common buckwheat, tartary buckwheat, amino acid, Fisher discriminant analysis, data mining

基金项目:国家自然科学基金项目（30771300）,山西省科技攻关项目（20120311006-1）

作者	单位	邮编
张强^*		041004
李艳琴	山西大学生物技术研究所

摘要点击次数: 1160

全文下载次数: 3

中文摘要:

根据甜荞和苦荞氨基酸的特点,筛选有效判别指标,建立判别准则。分别以氨基酸质量分数和相对含量为指标,采用逐步判别和Fisher判别法建立判别模型,并进行实验验证。采用质量分数法建立的判别模型,以组氨酸、谷氨酸、丝氨酸、甘氨酸、苏氨酸和酪氨酸为变量,回判正确率为94.71%,交互验证正确率为93.27%；采用相对含量法建立的判别模型,以组氨酸、谷氨酸、甘氨酸、丝氨酸、苏氨酸和丙氨酸的相对含量为变量,回判和交互验证正确率均为95.19%；采用12份荞麦材料对模型进行实验验证正确率为100%。组氨酸、谷氨酸、甘氨酸、丝氨酸、苏氨酸的质量分数和相对含量在荞麦判别中均发挥重要作用,氨基酸指标与判别分析技术相结合,可做为荞麦种类鉴别和质量控制的一种有效、可行的方法。

英文摘要:

According to the characteristics of common buckwheat and tartary buckwheat amino acid, this paper filtered the effective discrimination index and established a discriminant model. Using the content and the relative content of amino acids respectively for indices, the fisher discriminant model was established through stepwise method, and was verified by experiments. In the discriminant model, which was established through the content method with histidine, glutamic, serine, glycine, threonine and tyrosine for variables，the accuracy of resubstitution is 94.71% and the accuracy of cross-validated is 93.27%. In the discriminant model, which was established through the relative content method with the relative content of histidine, glutamic, serine, glycine, threonine and alanine for variables, the accuracy of resubstitution is 95.19% and the accuracy of cross-validated is also 95.19%. The accuracy of the model adopting 12 buckwheat samples was 100%. The content and the relative content of histidine, glutamic, serine, glycine and threonine play an important role in the discrimination of buckwheat. The combination of amino acid indices and modern statistical techniques should be a useful and convenient method for the discriminant and quality control of buckwheat.

关闭