农业工程学报Issue(13)：276-285,10.DOI:10.3969/j.issn.1002-6819.2014.13.034
基于主成分与聚类分析的苹果加工品质评价
Evaluation of apple quality based on principal component and hierarchical cluster analysis
摘要
Abstract
The purpose of this study was to investigate the variations in physical and chemical characteristics of apple fruit from 30 varieties grown in the same place using pattern recognition tools. Twenty quality parameters of apple samples (e.g. weight,volume, density, color, hardness, sugar-acid ratio, Vitamin C, etc.) were analyzed. Interrelationships between the parameters and the apple variety were investigated by descriptive statistics, principal component analysis (PCA) and hierarchical cluster analysis (HCA). PCA is a mathematical tool which performs a reduction in data dimensionality and allows the visualisation of underlying structure in experimental data and relationships between data and samples.In hierarchical cluster analysis, samples are grouped on the basis of similarities, without taking into account the information about the class membership. The results obtained following HCA are shown as a dendrogram in which five well-defined clusters are visible. Samples will be grouped in clusters in terms of their nearness or similarity. Cluster analysis uses less information (distances only) than PCA. It is interesting to observe what kind of classification can be made on the basis of distances only. The results showed that density, fruit shape index and water content of 30 apple varieties were not significantly different. The remaining seventeen measurements were investigated by principal component analysis. The first six components represented 83.56% of the total variability on the base of the total variance explained and screen plot of principal component analysis. The first principal component was related to titratable acidity, sugar-acid ratio and solid-acid ratio attributes, which were the taste quality factor. The second principal component was related to L,a, andb attributes, which were the color factor. Following that were sweetness factor, texture factor, quality factor and size factor. The sample score plots visually displayed the relationship between measurements and apple varieties. The varieties of Liaofu, Lixiang, Golden Delicious and Rome Beauty were located on the first quadrant of the plot for PC1 vs. PC2. They had better quality of among the mid-early maturity apples with high content of sugar-acid and solid-acid ratios which were good characteristics of taste even though their skin color was green. The varieties of Hanfu, Rainier and Fuji were located on the fourth quadrant of the same plot. They were common late maturity apples with red color and good quality for eating. Priam, Baixing and Pvma were located on the second quadrant of the plot, which may be suitable for processing not for eating, and they had a green skin color and sour taste. The varieties located on third quadrant were red but had low sugar-acid and solid-acid ratios, which can be used as breeding materials for red color and good quality of the early maturity apple. HCA classified 30 varieties into five main groups on the basis of the measured parameters, which was consistent with the results of PCA score plots. These results can be used preliminarily to determine whether the 30 apple varieties were appropriate for eating or for processing or for use as a breeding stock for future apple variety development.关键词
主成分分析/聚类分析/水果/苹果/理化品质Key words
principal component analysis/cluster analysis/fruits/apples/physical chemical characterization分类
轻工纺织引用本文复制引用
公丽艳,孟宪军,刘乃侨,毕金峰..基于主成分与聚类分析的苹果加工品质评价[J].农业工程学报,2014,(13):276-285,10.基金项目
国家公益性行业（农业）科研专项（200903043-01-03）；沈阳市科技计划项目资助 （）
