基于因子分析和Bayers判别的烤烟香型分类模型构建与验证

Construction and Verification of Classification Model for Flavor of Flue-cured Tobacco Based on Factor Analysis and Bayes Discriminant

为研究烤烟化学组成与其香型间的关系,通过抽样法收集了2011—2013年国内15省71市(县)500个烟叶样品。参照行业及文献相关标准测定影响其品质的114种化学指标,对各指标采用MFA(因子分析)降维处理,因子得分构建Bayes香型定量判别模型并验证。结果表明,原始指标可提出22个公因子,其对原变量的总方差解释率为80.459%;巨豆三烯酮(A、C)、His、假木贼碱、总细胞壁物质等是烟叶中普遍存在且能较好代表其品质特征的物质;定量判别模型能依据不饱合醛酮、氨基酸、碱、细胞壁物质等类物质的含量对烟叶样品香型进行较好的预测,回判及预测正确率≥83.3%。该判别模型使用简便、迅速,能简化烟叶香型的判别流程,快速和客观的评价烟叶品质。

In order to study the relationship between the chemical composition of flue-cured tobacco and its flavor, 500 tobaccosamples from 71 cities/counties from of domestic provinces were collected from 2011 to 2013. Based on industrial standards andmethods from the litemtoes standards we determined 114 chemical indicators which have been shown to have a significant impact onthe quality of tobacco. The dimensions of each index were reduced using MFA （factor analysis）, and the quantitative-Flavor factorscores were used in constructing and validating a Bayes discriminant model. The results showed that the original indicators can bemade 22 common factors, which can explain 80.459 percent total variance of the original variables. Megastigmatrienone （A,C）,His,anabasine, total cell wall material and other substances are widespread in tobacco leaves and can better represent their qualitycharacteristics. Flavor of tobacco could be predicted based on the quantitative discriminant model, which is constructed by unsaturatedaldehydes and ketones, acids, bases, and other substances in the cell wall material content, with the correct rate 83.3%. The model iseasy to use and could be important in simplifying the process of tobacco flavor discrimination.