基于主成分散点图的西南烟区烤烟质量综合评价  

Comprehensive Evaluation on Flue-cured Tobacco Quality in Southwest Tobacco-growing Region Based on Principal Component Scatter Diagram

作　　者：张玺[1] 谢晋[1] 梁增发 王寒[1] 王维[2] ZHANG Xi;XIE Jin;LIANG Zengfa;WANG Han;WANG Wei(China Tobacco Guangdong Industrial Co.,Ltd.,Guangzhou,Guangdong 510385;College of Agriculture,South China Agricultural University,Guangzhou,Guangdong 510642,China)

机构地区：[1]广东中烟工业有限责任公司,广东广州510385 [2]华南农业大学农学院,广东广州510642

出　　处：《贵州农业科学》2023年第2期91-98,共8页Guizhou Agricultural Sciences

基　　金：广东中烟工业有限责任公司项目“基于双喜品牌需求普洱基地高端烟叶原料定制化开发与应用”(H21335)。

摘　　要：【目的】构建多指标烤烟综合评价模型,并利用主成分散点图进行烤烟质量综合评价,为西南烟区烤烟质量的综合评价及烟叶工业应用的质量选择提供科学依据。【方法】试验烟叶选择云南普洱(C2F、C3F、B2F)、四川会理(C3F)、云南曲靖(B2F)和四川西昌(B2F),验证试验烟叶选择云南普洱(C2F、B2F)和云南曲靖(C2F、B2F)的原烟和复烤后片烟,以化学成分、感官质量及致香物质为评价指标,构建多指标烤烟综合评价模型,利用主成分散点图法进行烤烟质量综合评价。【结果】样品不同指标共提取4个主成分,其累计贡献率为81.848%,第1主成分为感官指标,贡献率为32.024%;第2主成分为化学指标,贡献率为24.308%;第3主成分为致香指标,贡献率为18.821%;第4主成分贡献率仅6.695%;4个主成分(y_(1)、y_(2)、y_(3)、y_(4))综合评价(得分)的数学模型分别为y_(1)=0.225x_(1)-0.059x_(2)-0.434x_(3)-0.413x_(4)+0.424x_(5)-0.517x_(6)+0.373x_(7)+0.036x_(8)-0.641x_(9)+0.873x_(10)+0.733x_(11)+0.781x_(12)+0.739x_(13)-0.420x_(14)+0.872x_(15),y_(2)=-0.031x_(1)+0.237x_(2)+0.535x_(3)+0.782x_(4)+0.179x_(5)-0.602x_(6)-0.846x_(7)-0.467x_(8)+0.674x_(9)+0.153x_(10)-0.319x_(11)+0.401x_(12)+0.478x_(13)-0.609x_(14)+0.151x_(15),y_(3)=0.640x_(1)+0.854x_(2)+0.668x_(3)+0.266x_(4)-0.304x_(5)+0.429x_(6)+0.211x_(7)+0.464x_(8)+0.231x_(9)+0.372x_(10)+0.233x_(11)+0.221x_(12)-0.169x_(13)+0.401x_(14)+0.383x_(15),y_(4)=-0.363x_(1)+0.337x_(2)+0.108x_(3)+0.232x_(4)+0.652x_(5)-0.124x_(6)+0.189x_(7)+0.073x_(8)-0.199x_(9)-0.116x_(10)-0.032x_(11)+0.011x_(12)+0.104x_(13)+0.364x_(14)-0.121x_(15),不同等级烟叶样品的综合得分依次为C2F>C3F>B1F>B2F;经聚类分析,不同来源各等级烟叶样品聚为4类,与主成分散点图法分析结果一致;验证试验C2F的综合评价得分均高于B1F。【结论】西南烟区不同等级烟叶样品的综合得分依次为C2F>C3F>B1F>B2F;主成分分析结合三维散点图法构建西南烟区烤烟质量综合评价模型的�【Objective】In order to provide the scientific basis for the comprehensive evaluation on flue-cured tobacco quality and the quality selection of industrial application of tobacco leaves in Southwest tobacco-growing area,the comprehensive evaluation model of multiple index is established and flue-cured quality is evaluated comprehensively by using principal component scatter diagram.【Method】Selecting Yunnan Pu’er(C2F,C3F and B2F),Sichuan Huili(C3F),Yunnan Qujing(B2F) and Sichuan Xichang(B2F) as tested tobacco,and selecting the raw material and cured tobacco strips of Yunnan Pu’er(C2F and B2F) and Yunnan Qujing(C2F and B2F) as verification test tobacco,the comprehensive evaluation model of multiple index is established based on the evaluation index of chemical component,sensory quality and aroma substance,and the flue-cured tobacco quality is evaluated comprehensively by using principal component scatter diagram.【Result】There are four principal components extracted from different indexes,and their cumulative contribution rate is 81.848%.The first principal component is sensory index with 32.024% of contribution rate;the second principal component is chemical index with 24.308% of contribution rate;the third principal component is aroma substance with 18.821% of contribution rate;the contribution rate of the fourth principal component is 6.695%.The mathematical model of comprehensive evaluation of four principal components(y_(1),y_(2),y_(3 )andy_(4)) is respectively y_(1)=0.225x_(1)-0.059x_(2)-0.434x_(3)-0.413x_(4)+0.424x_(5)-0.517x_(6)+0.373x_(7)+0.036x_(8)-0.641x_(9)+0.873x_(10)+0.733x_(11)+0.781x_(12)+0.739x_(13)-0.420x_(14)+0.872x_(15),y_(2)=-0.031x_(1)+0.237x_(2)+0.535x_(3)+0.782x_(4)+0.179x_(5)-0.602x_(6)-0.846x_(7)-0.467x_(8)+0.674x_(9)+0.153x_(10)-0.319x_(11)+0.401x_(12)+0.478x_(13)-0.609x_(14)+0.151x_(15),y_(3)=0.640x_(1)+0.854x_(2)+0.668x_(3)+0.266x_(4)-0.304x_(5)+0.429x_(6)+0.211x_(7)+0.464x_(8)+0.231x_(9)+0.372x_(10)+0.233x_(11)+0.221x_(12)-0.169x_(13)+0.401x_(14)+0.383x_(15),y_(4)=-0.363x

关 键 词：烤烟 化学成分 致香成分 感官质量 烟叶品质 主成分分析 主成分散点图 综合评价 

分 类 号：S572[农业科学—烟草工业]