基于支持向量机的烟叶不同部位6种化学成分差异性分析

Difference Analysis of Six Chemical Compositions in Tobacco Parts Based on Support Vector Machine

基于143个烟叶样品的6种化学成分(莨菪亭、芸香苷、苹果酸、柠檬酸、果糖、蔗糖)建立中部和上部烟叶的支持向量机分类判别模型,并利用该模型对独立的36个烟叶样品进行预报。结果表明,基于烟叶中6种化学成分建立的中部和上部烟叶的支持向量机分类模型的建模、留一法和预报准确率分别为89.51%、85.31%和83.33%。进一步考察了143个烟叶样品在主成分分析投影图中的空间分布及中部和上部烟叶的主要化学成分差异性,得出6种化学成分中蔗糖、莨菪亭和芸香苷是影响中部和上部烟叶差异的主要成分。研究结果表明,改变上部烟叶中的蔗糖、莨菪亭和芸香苷含量,可以扩大其在实际生产中的应用范围。

Support vector classification model has been constructed to discriminate the parts of tobacco from the upper section to the middle one, based on the dataset of 143 tobacco samples and six chemical compositions which are scopoletin, rutin, malic acid, citric acid, fructose, sucrose individually. The result shows that the accuracies of modeling, leave-one-out cross-validation and testing prediction are 89.51%, 85.31% and 83.33% respectively. Further, the spatial distributions in the principal component analysis based on the 143 samples and the main chemical difference analysis between the upper and middle sections have also been investigated, which leads the conclusion that the determined chemical compositions are sucrose, scopoletin and rutin. The research indicates that tuning these determined chemical components could expand the application of tobacco upper parts in practical production.