摘要
为了对江西烤烟进行综合评价,利用DPS数据处理系统软件对江西烤烟71个烟叶样本的香气成分进行主成分分析,按累积贡献率79.60%提取前7个主成分,计算综合得分,并将综合得分以卡方距离为衡量样本间品质差异的指标,运用离差平方法进行聚类分析,将71个样本分为4类。从第1类到第4类,评吸总分和香气指数B值逐渐降低。
In order to evaluate the quality of flue-tobacco grown in Jiangxi Province comprehensively, the content of aroma components in 71 flue-cured tobacco samples were analyzed by principal component analysis with DPS(Data Processing System) software. Seven principal components with accumulative contribution rate over 79.60% were obtained, and the scores of the seven principal components and comprehensive scores of the tobacco samples were also obtained. The Chi-square distance between the comprehensive scores was taken as a measurement of tobacco samples' differences, the system was clustered by square sum of deviation method, and the 71 samples were divided into 4 ranks. The comprehensive scores and aroma index B values gradually reduced from Rank 1 to Rank 4.
出处
《生物灾害科学》
2015年第3期239-245,共7页
Biological Disaster Science
基金
江西省烟草公司项目[赣烟(2011.01.001)]
池州市烟草公司项目(池烟2011.11.18)
关键词
烤烟
香味物质
主成分分析
聚类分析
flue-cured tobacco
aroma component
principal component analysis
cluster analysis