摘要
为科学地简化柑橘果实品质的评价指标,采用主成分分析、系统聚类及主成分聚类法,以湖北秭归地区生长的6个夏橙品种为材料,对测得的果实单果质量、出汁率、可溶性固形物含量、可滴定酸含量、抗坏血酸含量、果皮着色强度、果肉着色强度等18项常规指标进行了分析。主成分分析结果表明,这18项指标反映的夏橙品质可用4个主成分来表示(累计贡献率达100%),其中,前2个主成分囊括了所检测的18项品质指标,累计贡献率高达83.47%。利用上述4个主成分的值对6个夏橙品种进行聚类,在最大距离为1.0时,奥林达夏橙、无核夏橙、红夏橙和蜜奈夏橙归为一类;伏令夏橙、康贝尔夏橙分别单列为一类。聚为一类的夏橙品种果实品质相近,与实际口感和测得的色泽评价数据相吻合。根据这4个主成分对18项品质指标的特征向量进行主成分聚类,可将这些指标划分为6类,分别从这6类指标中挑选出果形指数、可溶性固形物、囊瓣数、抗坏血酸、果肉色调角及总可溶性糖等6个有代表性的指标,用这6个指标对夏橙品质聚类,在最大距离为1.0处,聚类结果与用4个主成分聚类结果相同,初步表明用以描述夏橙果实品质的18个指标可能简化为6个指标。
In order to simplify the indices for evaluating fruit quality without decreasing the information volume contained in original indices, multivariate statistics methods of principal component analysis, systematic cluster analysis and principal component cluster analysis were used for analyzing 18 fruit quality indices, including fruit mass ,juicy percentage, total soluble solid content, titratable acid content, vitamin C content, color intensity of rind, color intensity of pulp etc. for 6 late orange cuhivars grown in Zigui area, Hubei province. Results of analysis indicated that the cumulative contribution of 4 principal components reached 100% ,with the first 2 reached 83.47% and involved all 18 indices. Systematic clustering analysis using the 4 components as variants, when the maximum distance was around 1.0,01inda Orange, Seedless Late Orange, Rhode Red Orange and Midnight Orange were classified into one group while Valencia Orange and Campbell Orange as a separate group. Cultivars in one group had similar fruit quality in flavor and color. While principal component cluster analysis of the 4 component variants showed that 6 indices for a comprehensive system were obtained, and the 18 quality indices could be simplified as fruit shape index,total soluble solids content,segment number,vitamin C content,color hue angle of pulp and total soluble sugar content. However,using the 6 indices as variants in systematic clustering analysis on 6 late orange cuhivars, when the largest distance was 1.0, the clustering results resembled to the one by using the 4 component as variants with all 18 indices included ,which suggested that the 18 indices might be simplified and represented by the 6 ones.
出处
《果树学报》
CAS
CSCD
北大核心
2008年第5期640-645,共6页
Journal of Fruit Science
基金
教育部创新团队发展计划(IRT0548)
关键词
夏橙
果实品质
多元统计法
评价指标
Late orange
Fruit quality
Multivariate statistics
Quality evaluation indices
