香料的热脱附-气质联用指纹图谱偏最小二乘判别分析与差异表征  

Classify and quantity control casing flavor by PLS-DA combined with thermal desorption spectroscopy-gas chromatograph mass spectrometer

作　　者：陈康宁[1,2] 杨燕 李伟 王家俊 毛多斌[2] 侯英 

机构地区：[1]红云红河烟草(集团)有限公司,云南昆明650202 [2]郑州轻工业学院,河南郑州450002 [3]云南同创检测技术股份有限公司,云南昆明650106 [4]云南中烟工业有限责任公司技术中心,云南昆明650231

出　　处：《云南大学学报（自然科学版）》2015年第1期117-123,共7页Journal of Yunnan University(Natural Sciences Edition)

基　　金：红云红河烟草(集团)有限责任公司科技项目(HYHH2012HX03)

摘　　要：采集72个来自2类不同风格特征香料的热脱附-气质联用指纹图谱,结合偏最小二乘判别分析(PLS-DA),建立2类不同风格特征香料的类模型.使用26个验证样品对类模型进行验证(在P>0.05的条件下),结果表明,验证样品均获得了正确的分类,类模型判别效果良好;同时,通过因子载荷矩阵分析,计算各化学组分对2类香料分类的贡献率,表征2类香料化学组分的差异,结果令人满意.TDS-GC-MS chromatographic fingerprints of 72 flavors,which were belong to two different char- acteristic types ,were collected and integrated with analysis method of PLS-DA, to establish class model for two different characteristic flavors.When P〉0.05,26 verified samples were used to verify class model, and the result showed that the class model performed well in discrimination of verified samples＇ classifying; meanwhile, to calcu- late the contribution rate of each chemical compound to classify of the two types of flavors by analysis the factors of contribute matrix, to characterize the differences of chemical compounds in two types of flavors,led to a satisfy- ing results.

关 键 词：热脱附-气质联用指纹图谱 PLS-DA法 烟用香料 

分 类 号：O657.63[理学—分析化学]