不同产地甘草的近红外指纹图谱模式识别鉴别方法

Rapid Identification of Different Origin of Licorice by Near Infrared Spectroscopy and Pattern Recognition Method

目的:对四种产地(新疆、内蒙古(家种)、内蒙古(野生)、甘肃)的甘草进行鉴别。方法:利用近红外漫反射(NIRDRS)技术结合欧氏距离判别分析方法(Euclidean distance discriminant analysis)和偏最小二乘判别分析方法(Partial least-squares discriminant analysis,PLSDA)对四种产地(新疆、内蒙古(家种)、内蒙古(野生)、甘肃)的甘草进行鉴别。结果:PLSDA模型对不同产地甘草的训练集和预测集分类正确识别率均为100.0%。结论:近红外指纹图谱技术结合偏最小二乘判别分析方法为中药材产地的鉴别提供了一种简便、快速、准确的鉴定方法。

Objective：To indentify Xinjiang,Inner Mongolia（seed）,four（wild）in Inner Mongolia,gansu origin of radix glycyrrhizae.Methods：The similarity of four different origin（Sinking,Nei monggol（cultivatted）,Nei monggol（agrestal）and Gansu）of licorice were rapidly calculated by near-infrared diffuse reflectance spectroscopy（NIRDRS）combined with Euclidean distance discriminant analysis and Partial least-squares discriminant analysis（PLSDA）.Meanwhile,four different origin of licorice were rapidly identified by NIRDRS coupled with PLSDA.Results：Perfect classification ratio with 100% was obtained for the NIR fingerprint information based on PLSDA model,which indicated that the model had good classification performance.Conclusion：An efficient,nondestructive and accurate detection method was established for identification of origin of Chinese herbs by NIRDRS combined with PLSDA.