化学计量学结合气相色谱法对12种植物油判别模型的构建

测定了12种植物油的脂肪酸组成和含量，探讨了利用植物油脂肪酸的指标对不同种类的植物油进行分类和判别的可能性。采用气相色谱法对12种植物油脂肪酸的组成和含量进行测定，利用SPSS 22. 0统计软件进行主成分分析、系统聚类、 K-平均值聚类和判别分析。结果表明：12种植物油的主要共有组成为油酸（15.635%~66.569%）和亚油酸（10.521%~58.227%），其中棉籽油（16.285%）中单不饱和脂肪酸含量最低，橄榄油（67.628%）中单不饱和脂肪酸含量最高；橄榄油（11.284%）中多不饱和脂肪酸含量最低，核桃油（67.167%）中多不饱和脂肪酸含量最高；12种植物油中不饱和度最小为棉籽油3.235，最大为低芥酸菜籽油14.672。主成分分析降维得到5个主成分，利用主成分分析数据，依次进行聚类分析和判别分析，系统聚类和K-平均值聚类结果一致，可对12种植物油聚类区别，通过判别分析建立了3个典则判别函数，对不同植物油的分类和判别的效果良好。

Construction of a model for distinguishing 12 kinds of vegetable oils by Chemometrics and Gas chromatography

The fatty acid composition and content of 12 kinds of vegetable oils were measured, and the possibility of using the index of vegetable oil fatty acids to classify and distinguish different kinds of vegetable oils was explored.The composition and content of 12 kinds of vegetable oil fatty acids were determined by gas chromatography, and SPSS 22.0 statistical software was used for principal component analysis, hierarchical clustering, K-means clustering and discriminant analysis.The results showed that the main common components of 12 kinds of vegetable oils were oleic acid (15.635%~66.569%) and linoleic acid (10.521%~58.227%). Among them, cottonseed oil (16.285%) had the lowest monounsaturated fatty acid content and olive oil (67.628%) has the highest content; olive oil (11.284%) has the lowest polyunsaturated fatty acid content, and walnut oil (67.167%) has the highest content; among the 12 vegetable oils, the least unsaturation is cottonseed oil 3.235 and the largest is canola oil 14.672.Principal component analysis dimensionality reduction obtains 5 principal components. Using principal component analysis data, cluster analysis and discriminant analysis are performed in sequence. The results of systematic clustering and K-mean clustering are consistent, and 12 kinds of vegetable oils can be clustered and distinguished. Analysis and establishment of 3 canonical discriminant functions, the classification and discrimination of different vegetable oils are effective.