基于主成分分析法评价不同产地香牙蕉的营养品质  

Evaluation of Nutrition and Quality of Banana from Different Producing Areas Based on Principal Component Analysis

作　　者：戎瑜 王明月[1] 吕岱竹[1] 宋佳 马晨[1] RONG Yu;WANG Mingyue;LYU Daizhu;SONG Jia;MA Chen(Analysis and Testing Center,Chinese Academy of Tropical Agriculture Sciences/Hainan Provincial Key Laboratory of Quality and Safety for Tropical Fruits and Vegetables,Haikou,Hainan 571101,China)

机构地区：[1]中国热带农业科学院分析测试中心/海南省热带果蔬产品质量安全重点实验室,海南海口571101

出　　处：《热带作物学报》2022年第5期1045-1054,共10页Chinese Journal of Tropical Crops

基　　金：海南省自然科学基金青年基金项目(No.321QN282);财政部和农业农村部国家现代农业产业技术体系项目(No.CARS-31);中国农业科学院科技创新工程重大科研项目。

摘　　要：通过主成分分析和聚类分析2种方法对采自5个产地的香牙蕉(广西、云南、福建、海南和广东)功能性营养成分与品质指标进行比较和分析,首先对5个产地香牙蕉样品的各营养和感官指标进行相关系数矩阵分析,得到5个产地香牙蕉样品分别在F1风味和矿物质元素指标(甜度、可溶性固形物、总酸、水分、钾、磷、镁、铜)、F2维生素类指标(维生素B6、烟酸、β-胡萝卜素和多糖)、F3外观指标(长度和围径)、F4感官指标(香味、可食率)、F5功能性指标(槲皮素、总抗性淀粉)五大主成分上的累计方差贡献率达78.89%,因此这5个因子能够用作分析香牙蕉品质的主要指标。通过三维散点图和计算综合得分,云南、广东、海南省3地香牙蕉样品总体得分居高,结果表明香牙蕉营养品质较高,而福建省香牙蕉营养品质较低。进一步结合聚类分析的层次聚类和K-均值聚类2种分析方法,与主成分分析的结果一致,云南省香牙蕉品质最高,其余产地品质分属一类与二类。这可能与各指标差异性分析中的云南省香牙蕉样品在β-胡萝卜素、香蕉多糖和矿物质元素等指标上显著高于其他4个产区(P<0.05)相关。通过以上不同方法综合分析,可以得出来自五大产区的香牙蕉样品综合营养品质水平,研究结果为我国香牙蕉的品种培育、产地选择提供理论依据及参考。In this study,principal component and cluster analysis methods were used to compare and analyze the functional nutritional components and quality indexes of banana samples collected from Guangxi,Yunnan,Fujian,Hainan and Guangdong in China.Firstly,five principal components were obtained according to the correlation coefficient matrix analysis conducted on nutritional and sensory indexes of banana samples.The five main component factors were detailed as follows:F1 as the index of flavor and mineral elements(sweetness,soluble solids,total acid,water,potassium,phosphorus,magnesium,copper),F2 as the vitamin index(vitamin B6,niacin,β-carotene,polysaccharides),F3 as the appearance index(length and diameter),F4 as the sensory index(flavor,edible rate),and F5 as the functional index(quercetin and total resistant starch),the cumulative variance contribution of the above five principal component factors reached 78.89%.Therefore,they could be used as the main indexes to analyze banana quality.Through the three-dimensional scatter plot and the calculation of the comprehensive score,the overall scores of Yunnan,Guangdong and Hainan were higher than that of Guangxi and Fujian,indicating that the nutrition quality of bananas in the three places was of superior grade,while the nutrition quality of bananas in Fujian was lower.The results of hierarchical clustering and K-means clustering were consistent with the results of principal component analysis.The quality of banana in Yunnan was the highest,and the quality of other producing areas belonged to class I and class II.This might be related to the fact that the contents ofβ-carotene,banana polysaccharide and mineral elements in banana samples of Yunnan were significantly higher than those in other four producing areas(P<0.05).Through the comprehensive analysis of the above different methods,the nutrition quality level of banana samples from the five major producing areas could be obtained,which would provide a theoretical basis and reference significance for the species cultiva

关 键 词：香牙蕉 营养品质 主成分分析 聚类分析 

分 类 号：S668.1[农业科学—果树学]