中国烟草学报2017，Vol.23Issue(1)：35-42,8.DOI:10.16472/j.chinatobacco.2016.248
基于烟丝挥发/半挥发性组分构建卷烟果香风格特征预测模型
Establishment of quantitative prediction model for fruit note of cigarette based on volatile/semi-volatile components in tobacco
摘要
Abstract
73 volatile/semi-volatile components in 80 cigarette brands sold in the market were detected by GC-MS / MS method to establish a prediction model for fruity style characteristic of cigarette by factor analysis and step-wise linear regression method.Results showed that ① The cumulative variance interpretation rate of seven common factors to the original indicators was 87.77％.② Carbonyl compound containing N or O atoms,cyclic ketones and long-chain carboxylic acids,alcohol,benzyl alcohol compound,D/L-menthol had greater contribution to fruity characters.③ Regression model had good result for smoking fruity index of the training set,with determination coefficient R at 0.939.The model also had great predictive ability for smoking fruity index of the testing set,with mean absolute error at 0.1876 only.With accurate regression and prediction results,the model could be used as an objective method to predict sensory quality in cigarette.关键词
果香风格特征/GC-MS/MS/因子分析/逐步线性回归Key words
chemical composition/fruity style characteristics/GC-MS/MS/factor analysis/step-wise linear regression引用本文复制引用
李超,王明锋,张志敏,张承明,王文元,李娥贤,秦云华..基于烟丝挥发/半挥发性组分构建卷烟果香风格特征预测模型[J].中国烟草学报,2017,23(1):35-42,8.基金项目
云南中烟工业有限责任公司2015年度科技项目“适用于卷烟产品质量管控的分析检测平台搭建及应用”(2015JC07) （2015JC07）
云南中烟工业有限责任公司2013年度科技项目“红塔集团质量监督检测实验室建设”(2013JC10) （2013JC10）
中国烟草总公司2015年度科技项目高静压(HHP)技术在卷烟香精香料中的应用研究(110201402036) （HHP）
