适用于3类茶的定性分类及主要内含成分定量分析的近红外预测模型的建立
Establishment the prediction model for qualitative classification and quantitative analysis of major components for three kinds of tea by near infrared spectroscopy

以白茶、乌龙茶和红茶为研究对象，力求同时实现对3类茶的定性鉴别和主要内含物含量的预测。采用因子化法建立定性预测模型，该模型能够正确识别全部独立验证集样品；采用偏最小二乘法结合主成分分析建立茶叶中含水量、茶多酚和咖啡碱的定量校正模型，并用相关系数、交互验证均方差和预测均方差对模型进行评价。验证集中含水量、茶多酚和咖啡碱3个内含物成分定量预测模型的相关系数分别为0.991 3、0.905 7和0.974 3。结果表明，预测模型能实现同时对3类茶叶的定性分类和主要内含物含量预测的目的，也达到了降低近红外预测模型成本的目的。
In order to identify and predict major components for white tea,oolong tea and black tea in the same time,the 3 different tea products were collected,and the qualitative and quantitative models were established based on NIRs (near infrared spectroscopy).The qualitative classification model,which was established by factor-method,could identify full of the independent validation set samples.At the same time,partial least squares (PLS) algorithm and principal component analysis (PCA) were conducted on the calibration of regression model.The performance of the final model was evaluated according to root mean square error of cross-validation (RMSECV),root mean square error of prediction (RMSEP) and correlation coefficient (R).The correlations coefficients (R) in the prediction set were achieved as follows:0.991 3 for moisture model,0.905 7 for tea polyphenols model,and 0.974 3 for caffeine model.It was demonstrated that it was feasible to establish a “saving” qualitative and quantitative prediction model for 3 different kinds of tea products with the premise of accuracy.