NIRS结合PLS快速分析银黄颗粒中有效成分含量  

Fast determination of active components in Yinhuang granules by NIRS combined with PLS

作　　者：白雁[1] 王星[1] 龚海燕[1] 芦猛[1] 史会齐[1] 

机构地区：[1]河南中医学院,河南郑州450008

出　　处：《计算机与应用化学》2010年第12期1703-1706,共4页Computers and Applied Chemistry

基　　金：河南省重大公益科研项目(081100912500);河南省杰出人才项目(084200510017)

摘　　要：研究旨在探讨利用银黄颗粒样品的近红外漫反射光谱(NIRS)信息,建立黄芩苷和绿原酸含量的校正模型,为银黄颗粒质量的快速评价提供1种新方法。以HPLC分析值为参照,采用近红外漫反射光谱技术采集100批银黄颗粒样品的近红外漫反射光谱,结合偏最小二乘法(PLS)建立了黄芩苷和绿原酸含量的校正模型。黄芩苷和绿原酸含量的校正模型相关系数(R2)分别为0.998和0.995,校正均方差(RMSEC)为0.578和0.123,内部交叉验证均方差(RMSECV)为2.356和0.412;经外部验证,预测相关系数(r)分别为0.995和0.984,预测均方差为(RMSEP)0.597和0.166。结果表明,该方法准确、简便、无污染,可实现大批量银黄颗粒样品的快速分析。The research aimed to establish the calibration models of baicalin and chlorogenic by near-infrared reflectance spectroscopy（NIRS）,and provide a new method of rapid assessment for yinhuang granules fastly.Near-infrared reflectance spectra of 100 samples were collected,and the calibration models of baicalin and chlorogenic acid were established by partial least squares（PLS） with HPLC analysis values as reference.The correlation coefficients（R2） of the calibration models were 0.998 and 0.995,the root-mean-square error of calibration（RMSEC） were 0.578 and 0.123,the root-mean-square error of cross-validation（RMSECV） were 2.356 and 0.412,the correlation coefficients of prediction（r） were 0.995 and 0.984,the root-mean-square error of prediction（RMSEP） were 0.597 and 0.166.The results indicated that the method is accurate,simple and non-polluted,and could be applied for the fast analyzation of large quantities of numbers of yinhuang granules samples.

关 键 词：近红外漫反射光谱 偏最小二乘法 银黄颗粒 

分 类 号：TQ015.9[化学工程] TP391.9[自动化与计算机技术—计算机应用技术]