卷烟叶组化学检测数据的因子分析和逐步判别分析  

Factor Analysis and Stepwise Discriminant Analysis of Chemical Data of Blended Tobacco

作　　者：曾晓鹰[1] 李庆华[1,2] 王玉 王保兴 侯英 许国旺[2] 

机构地区：[1]红云红河烟草(集团)有限责任公司技术中心,昆明市北郊上庄650202 [2]中国科学院大连化学物理研究所,辽宁省大连市中山路457号116023 [3]云南瑞升烟草技术(集团)有限公司

出　　处：《烟草科技》2010年第7期43-47,共5页Tobacco Science & Technology

基　　金：国家烟草专卖局资助项目"利用化学指标体系表征卷烟产品品质特征的研究"(110200401050);云南中烟工业公司资助项目"应用信息化平台构建SPC技术卷烟工艺质量控制体系"(2006XX01)

摘　　要：为了寻找主要影响卷烟叶组质量的化学检测数据,对3个模块叶组的总糖、还原糖、总氮、总植物碱、含水率、氯离子、钾离子、pH值、总挥发碱、果糖、麦芽糖、苹果酸、异戊酸、苯甲醇、茄酮、β-大马酮分别进行了因子分析和逐步判别分析。结果表明:①三个模块可分别用7个、7个和8个独立的公因子来描述其质量特征。其中,前3个具有最大方差因子载荷的主因子集中体现了总糖、还原糖、总氮、苯甲醇、茄酮、β-大马酮等指标的信息;②总糖、氯离子、pH值、总挥发碱、果糖、苹果酸、异戊酸、茄酮和β-大马酮等9个指标进入判别函数,该判别函数具有显著意义,其自身验证和交互验证回判的准确率分别达96.7%和93.3%。In order to find out the main chemical data influencing the smoking quality of blended tobacco, factor analysis and stepwise discriminant analysis were carried out on the basis of the test data of 16 chemical components in 3 modular tobacco blends,including total water-soluble sugar,reducing sugar,total nitrogen, total alkaloid,moisture,Cl - ,K ＋ ,pH,total volatile bases,fructose,maltose,malic acid,isovaleric acid, benzyl alcohol,solanone and β-damascone. The results showed that the quality of the three modular blends could be characterized by 7,7,and 8 independent common factors,respectively,the first three principal factors,which held the maximum variance factor loading,mainly represented the information about total sugar,reducing sugar,total nitrogen,benzyl alcohol,solanone,and β-damascone; 2 ） nine chemical indexes,including total sugar,Cl - ,pH,total volatile bases,fructose,malic acid,isovaleric acid,solanoneand β-damascone,were included in the discriminant function of significance with internal validation accuracy of 96. 7% and cross validation accuracy of 93. 3% .

关 键 词：卷烟叶组 化学指标 因子分析 逐步判别分析 

分 类 号：TS411.1[农业科学—烟草工业]