主成分回归分析法在永州烟叶评吸质量与外观质量的关系模型构建中的应用  

Building Relation Model of Smoking Quality and Appearance Quality from Yongzhou Tobacco Leaf by Principal Component Regression

作　　者：张延军[1] 李旭华[1] 林锐锋[1] 文俊[1] 叶为民[1] 

机构地区：[1]广东中烟工业有限责任公司,广州510145

出　　处：《中国农学通报》2012年第18期256-259,共4页Chinese Agricultural Science Bulletin

基　　金：广东中烟工业有限责任公司科研发展计划项目"满足‘双喜’配方需求的外观品质因素和内在质量的研究"(粤烟[2010]科字第006号)

摘　　要：为了研究烟叶评吸质量指标与外观质量指标的关系,利用统计分析软件SPSS的主成分分析和逐步回归分析的方法建立了湖南永州评吸质量与外观质量的关系模型。结果表明:(1)建立的回归方程:?=1.286+0.150Z1+0.058Z2,回归方程达到极显著水平,方程的决定系数R2=0.893,对烤烟评吸质量进行预测是可行的。(2)通过主成分分析可知,颜色、成熟度和油份是决定烤烟评吸质量的主要因素,叶片结构和身份也是决定评吸质量的重要因素。In order to study the relationship of smoking quality and appearance quality,established the relation model between smoking quality and appearance quality of Hunan Yongzhou area,carried out with the principal component analysis and stepwise regression analysis of SPSS software.The results showed：（1）The regression equation Y=1.286 ＋ 0.150Z 1 ＋ 0.058Z 2,which reached extremely significant level,and the coefficient determination（R2） was 0.893.The predication to smoking quality through this equation was feasible.（2）Through the principal component analysis can concluded the mostly factors of smoking quality was color,maturity and oil;the important factors was leaf structure and degree.

关 键 词：主成分分析 逐步回归 评吸质量 外观质量 

分 类 号：S572[农业科学—烟草工业]