| 摘要: |
| 采用高效液相色谱-气相色谱-质谱联用法（HPLC-GC-MS）测定中部和下部烟叶的巨豆三烯酮、β-紫罗兰酮、氧化紫罗兰酮、茄酮等11种致香成分，应用遗传算法（GA）对筛选出的8种致香成分建立中部和下部烟叶支持向量机（SVM）分类判别模型.结果表明，中部和下部烟叶的SVM分类判别模型的建模、留一法及预报准确率分别为95.45%，89.39%和81.25%.利用Fisher判别矢量方法考察了中部和下部烟叶的空间分布规律，分析出中部和下部烟叶致香成分中，巨豆三烯酮、β-紫罗兰酮、氧化紫罗兰酮差异显著. |
| 关键词: 烟叶部位 致香成分 遗传算法(GA) 支持向量机(SVM) |
| DOI：10.3969/J.ISSN.1000-5137.2019.04.013 |
| 分类号:S572;TP18 |
| 基金项目:国家自然科学基金青年基金（21706156） |
|
| Difference analysis of aroma components in tobacco leaves based on GA-SVM |
|
SHEN Yushu1, CAO Xiaowei2, YU Jie3, SHA Yunfei3, YUE Baohua1
|
|
1.School of Materials Science and Engineering, Shanghai University, Shanghai 200444, China;2.College of Chemistry and Materials Science, Shanghai Normal University, Shanghai 200234, China;3.Technology Center, Shanghai Tobacco Group Co., Ltd., Shanghai 200082, China
|
| Abstract: |
| Eleven different aromatic components including megastigmatrienone,beta-Ionone,Ionone oxide and solanone from middle and lower tobacco leaves were determined successfully via high performance liquid chromatography-gas chromatography-mass spectrometry (HPLC-GC-MS) system.By using genetic algorithm(GA),8 aromatic components were selected to build a support vector machine(SVM) classification model for discriminating middle and lower tobacco leaves.The results showed that the accuracies of modeling,leave-one-out,and prediction were 95.45%,89.39% and 81.25%,respectively.The spatial distribution of middle and lower tobacco leaves was investigated by Fisher discriminant vector method,which showed that megastigmatrienone,beta-Ionone,and Ionone oxide were evidently different in the middle and lower tobaccos leaves. |
| Key words: tobacco leaves stalk positions aromatic components genetic algorithm(GA) support vector machine(SVM) |
