基于气质的主成分支持向量基对新疆薰衣草精油判别分析  

Discrimination of Varieties of Lavender Oil by GC-MS Based on Principal Component Analysis and Support Vector Machine

作　　者：廖享 高晶[1] 唐军[1] 符继红[1] 

机构地区：[1]新疆大学理化测试中心,乌鲁木齐830046

出　　处：《新疆农业科学》2014年第7期1367-1372,共6页Xinjiang Agricultural Sciences

基　　金：国家自然科学基金项目(21265021)

摘　　要：【目的】快速鉴别不同品种的薰衣草精油,为精油品质控制提供可靠的科学方法依据。【方法】通过气相色谱质谱(GC-MS)测定三个品种共66个薰衣草精油样品。应用峰面积归一法确定各成分的相对含量。对构建的特征信息数据进行主成分分析(PCA),选取7个主成分代替原始数据,再利用支持向量机技术进行分类和预测,对不同品种的薰衣草精油进行鉴别。【结果】通过48个样本建立支持向量机的分类模型,对18个样本进行预测,对训练集样本的训练正确率达到97.92%,对预测集样本的正确识别率达到94.44%。【结论】主成分分析结合支持向量机方法具有很好的分类和鉴别作用,可作为薰衣草精油品种区分的有效方法之一,为薰衣草精油的质量控制提供了一定的科学依据。[Objective] The aim of this study is to provide a fast and reliable scientific approach to identify different varieties of lavender essential oil.[Method] Gas chromatography-mass spectrometry （GC-MS） were used to detect 66 lavender essential oil samples,and the relative content of the components was determined by peak area normalization method.Principal component analysis （PCA） was employed to study the constructed information.7 principal components were selected to replace the original data to classify and predict the varieties of lavender oils by support vector machine （SVM）.[Result] The classification model of support vector machine was set up by means of 48 samples,and 18 samples were predicted.The result of the proposed model showed that the discriminate rate for 48 samples in training set was 97.92％,and the accuracy of the prediction of the test set can reach 94.44％ for 18 samples.[Conclusion] The results demonstrated that the support vector machine data based on principal component analysis method had a good classification and identification effect to discriminate the lavender oil＇s variety.The method can be used to provide a scientific basis for the quality control of lavender essential oil.

关 键 词：薰衣草精油 气相色谱质谱 主成分分析 支持向量机 

分 类 号：S126[农业科学—农业基础科学]