研究论文

基于致香成分的上部烟叶和中部烟叶分类判别

展开
  • 1. 上海烟草集团有限责任公司技术中心, 上海200082
    2. 河南师范大学环境学院, 河南新乡453007

收稿日期: 2017-12-08

  网络出版日期: 2019-05-05

基金资助

国家自然科学基金资助项目(21273145)

Classification of upper and middle leaves of tobaccos based on aromatic

Expand
  • 1. Technology Center, Shanghai Tobacco Group Co., Ltd., Shanghai, 200082, China
    2. School of Environment, Henan Normal University, Xinxiang 453007, Henan, China

Received date: 2017-12-08

  Online published: 2019-05-05

摘要

对上部烟叶和中部烟叶中致香成分的差异性进行了分析. 首先,利用液相色谱-气相色谱-质谱联用法测定烟叶样品内的中性致香成分,包括茄酮、二氢猕猴桃内酯、巨豆三烯酮等共11种; 其次,利用遗传算法筛选出影响上部烟叶和中部烟叶差异性的8种致香成分,进而建立了基于致香成分的上部烟叶和中部烟叶支持向量机分类判别模型,建模、留一法和预报和准确率分别为88.65%, 84.40%和82.86%;最后, 利用Fisher判别矢量方法考察上部烟叶和中部烟叶的空间分布.结果表明, 3-羟基-$\beta$-二氢大马酮、巨豆三烯酮和茄酮是影响上部和中部烟叶差异性的3种主要致香成分,可为烟叶质量管理中烟叶的部位特征和香气控制提供参考.

本文引用格式

董惠忠, 毕艳玖, 赵晓华, 葛炯, 沙云菲 . 基于致香成分的上部烟叶和中部烟叶分类判别[J]. 上海大学学报(自然科学版), 2019 , 25(2) : 309 -316 . DOI: 10.12066/j.issn.1007-2861.1985

Abstract

The difference of aromatic components in the upper part and middle part of tobacco leaves was analyzed. Eleven different aromatic components, including solanone, dihydroactinidiolide and megastigmatrienone, were separated successfully through the HPLC-GC-MS system. Then eight aromatic components were selected as the major variants that caused the difference in tobacco odor by genetic algorithm. After that, a model that distinguishes the upper part and middle part of tobacco leaves was establishedby support vector machine. The accuracy of the proposed model, Leave-one-out cross-validations and the accuracy for forecasting unknown samples reached 88.65%, 84.40% and 82.86%respectively. Finally, Fisher discriminant vector method was used to distinguish the distribution of aroma on the upper part and middle part of tobacco leaves. The result shows that 3-hydroxy-$\beta $- dihydrogen ketone, megastigmatrienone and solanone were three major aromatic components that contribute to the difference in the upper part and middle part of tobacco leaves. This study is meaningful in that it provides help for tobacco leaves quality control and aroma control in the tobacco industry.

参考文献

[1] 张鑫, 郭佳, 倪力军 , 等. 基于红外与近红外光谱的烟叶部位识别[J]. 光谱学与光谱分析, 2007,27(12):2437-2440.
[2] 李翠英, 贺立源, 马文杰 , 等. 采用轮廓特征的烟叶部位组分类研究[J]. 计算机工程与应用, 2009,45(26):236-239.
[3] 于春霞, 马翔, 张晔晖 , 等. 基于近红外光谱和SIMCA算法的烟叶部位相似分析[J]. 光谱学与光谱分析, 2011,31(4):924-927.
[4] 牛玉德, 高华锋, 薛林 , 等. 烤烟部位量化识别判定方法研究[J]. 湖北农业科学, 2016,55(18):4749-4752.
[5] 韩小渊, 范磊, 卢晓延 , 等. 主脉特征在烟叶部位识别中的应用[J]. 烟草科技, 2017,50(2):22-26.
[6] 杨艳芹, 储国海, 周国俊 , 等. 双内标气相色谱-质谱联用法测定烟草干馏香料致香成分含量[J]. 理化检验(化学分册), 2016,52(11):1272-1276.
[7] 尧珍玉, 曾池, 施鸣 , 等. 烟草中关键致香物质积累、降解及其对品质影响的研究进展[J]. 贵州农业科学, 2017,45(5):28-31.
[8] Holland J H. Adaptation in natural and artificial systems [M]. Ann Arbor: University of Michigan Press, 1975.
[9] 蔡峰, 刘太昂, 张朝平 , 等. 基于近红外数据预测烟草淀粉含量的GA-SVR模型[J]. 计算机与应用化学, 2014,31(8):969-971.
[10] Vapnik V N . The nature of statistical learning theory[M]. Berlin: Springer, 1995: 136-196.
[11] Vapnik V N. 统计学习理论的本质 [M]. 张学工,译. 北京: 清华大学出版社, 2000: 152-208.
[12] 董勇, 孙广玲, 刘志 . SVM+模型中可用信息用作特权信息[J]. 上海大学学报(自然科学版), 2017,23(4):524-533.
[13] Wilkins C L, Isenhour T L . Multiple discriminant function analysis of carbon-13 nuclear magnetic resonance spectra: functional group identification by pattern recognition[J]. Analytical Chemistry, 1975,47(11):1849-1851.
[14] Rasmussen G T, Ritter G L, Lowry S , et al. Fisher discriminant function for a multilevel mass spectral filter network[J]. Journal of Chemical Information and Modeling, 1979,19(4):255-259.
[15] 邓红梅, 黄伟 . 二阶精度混合Legendre-球面调和拟谱方法求解Fisher型方程[J]. 上海大学学报(自然科学版), 2015,21(3):331-335.
[16] 张静, 宋锐, 郁文贤 . 基于混淆矩阵和Fisher准则构造层次化分类器[J]. 软件化学, 2005,16(9):1560-1567.
文章导航

/