核桃感官和营养品质的主成分及聚类分析

Principal Component Analysis and Cluster Analysis of Sensory and Nutritional Quality of Walnut

为从感官指标和理化指标获得核桃的品质信息,按照GB/T 20398—2006和GB/T5009—2003中的方法测定核桃资源的外观品质和种仁中蛋白质、脂肪及N、P、K、S、Ca、Mg、Fe、Mn、Zn、Cu、Se的含量,应用主成分和聚类分析法对核桃品质进行分析。主成分分析得出一个5因子模型,解释了试验数据的80.83%;第1、2主因子的方差累积贡献率达46.21%,对应的指标为出仁率、壳厚、单果质量、横径、矿质元素Ca、Cu,显示出对核桃品质评价非常质量要;聚类分析将13个核桃样品分成3组,在一定程度上体现了核桃样品的亲缘关系和地域分布特征。

In order to more scientifically evaluate the nutrition quality and classify the walnut resources based on the sensory, physical and chemical indicators, the appearance quality and the contents of protein, fat and eleven mineral elements （N, P, K, S, Ca, Mg, Fe, Mn, Zn, Cu and Se） in the walnut seed were determined according to GB/T 20398--2006 and GB/T 5009--2003, and nutritional quality was analyzed by principal component analysis and cluster analysis. Five principal components accounted for 80.83% of total variances. The cumulative contribution rate of the first factor and the second components to total variances was 46.21%, suggesting physical indicators, shelling percentage, shell thickness, fruit weight, fruit diameter, and mineral elements cadmium and copper to be characteristic indicators in evaluating walnut, and cadmium and copper to be characteristic elements. Cluster analysis showed that thirteen walnut samples could be classified into three groups, thus providing an approach for the identification of genetic kinship and geographical distribution to some extent.