中文摘要：

采用气相色谱-质谱（GC-MS）和液相色谱（LC）方法,结合主成分分析（PCA）、高斯混合模型（GMM）对49个茶叶样本进行分类判别研究.通过PCA对茶叶的GC-MS信号进行特征提取,结合LC测得的茶多酚等10个变量,运用GMM对茶叶样本进行分类,训练集正确率为99.44%,预测集正确率为90.47%,结果表明该方法适用于茶叶的分类及品质评价.

英文摘要：

Gas chromatography-mass spectrometer（ GC- MS） and liquid chromatography（ LC）,combined with principal component analysis（ PCA） and Gaussian mixture model（ GMM）,were applied for classification of 49 tea samples. The PCA was firstly employed to reduce the dimensionality of GC- MS variables. The variables used in classification also included ten compositions determined by LC,such as tea polyphenols. Then the GMM was used to establish the classification models. The classification result showed that the accuracy rate of training set and prediction set was 99. 44% and 90. 47%,respectively.It could be concluded that GMM combined with chromatography for the classification of tea had a good performance.