基于中红外光谱结合化学计量学对不同产地山茱萸鉴定与分析  

Identification and Analysis of Cornus officinalis from Different Habitats Based on Mid Infra-Red Spectroscopy Combined with Chemometrics

作　　者：安淑静 王婷[1] 牛豆 韩彬凯 毕淮龙 康杰芳[1] AN Shujing;WANG Ting;NIU Dou;HAN Binkai;BI Huailong;KANG Jiefang(Key Laboratory of Medicinal Resource and Natural Pharmaceu ical Chemistry of Ministry of Education,National Engineering Laboratory for Resource Developing of Endangered Chinese Cr de Drug in Northwest of China,College of Life Sciences,Shaanxi Normal University,Xi'an 710119,China)

机构地区：[1]陕西师范大学教育部药用资源与天然药物化学重点实验室,西北濒危药材资源开发国家工程实验室,陕西西安710119

出　　处：《中医药学报》2021年第8期49-54,共6页Acta Chinese Medicine and Pharmacology

基　　金：国家重点研发计划(2017YFC1701302);陕西省重点研发计划(2020ZDLSF05-10);陕西省自然科学基金(2019JM-352)。

摘　　要：目的:为山茱萸产地鉴别和质量评价建立便捷、准确、稳定的方法,对区域化种植和生产实践提供基础理论依据。方法:采集7个省份29批山茱萸药材的中红外(MIR)光谱数据,结合线性回归分析(LR)、BP-神经网络(BP-ANN)和支持向量机(SVM)三种化学计量学方法对山茱萸的产地鉴别和活性成分含量测定进行研究。结果:在两个主产区陕西省和河南省的山茱萸药材的产地鉴别模型中,SVM的识别率最高,达到86.21%。进一步模型验证发现,SVM的识别率为92.31%和87.50%。在活性成分含量预测模型中,除原儿茶酸外,SVM模型的R2均大于LR和BP-ANN,且SVM建立的模型的MSE和MAD均大于LR和BP-ANN。进一步对模型的验证发现,通过高效液相法(HPLC)测定29批山茱萸样品的活性成分含量的实际值与模型预测值相比,除原儿茶酸外,SVM建立的模型的R2更接近1。结论:MIR结合SVM建立模型,识别率和拟合度高,可用于山茱萸的产地鉴别和活性成分的含量测定。Objective:To establish a convenient,accurate and stable method for producing area identification and quality evaluation of Cornus officinalis,thus to provide basic theoretical basis for regional planting and production practice.Methods:The mid infra-red spectroscopy(MIR)data of 29 batches of Cornus officinalis from 7 provinces were collected.The identification of producing areas and determination of active components of Cornus officinalis were studied by three kinds of chemometrics of linear regression analysis(LR),Artificial Neural Networks(BP-ANN)and support vector machine(SVM).Results:The results showed that the recognition rate of SVM was the highest,which reached 86.21%,among the two main producing areas of Cornus officinalis in Shaanxi Province and Henan Province.Further model validation showed that the recognition rates of SVM were 92.31%and 87.50%.Except for protocatechuic acid,the value of R2 of SVM model was greater than LR and BP-ANN,and MSE and MAD of SVM model were greater than LR and BP-ANN.Further validation of the model showed that the true value of the content of active components in 29 batches of Cornus officinalis samples determined by HPLC was higher than that predicted by the model;except for protocatechuic acid,the value of R2 of SVM model was closer to one.Conclusion:The model established by MIR combined with SVM has high recognition rate and fitting degree,which can be used to identify the origin of Cornus officinalis and determine the content of active components of the herb.

关 键 词：山茱萸 中红外光谱 化学计量学 产地鉴别 质量评价 

分 类 号：R284.1[医药卫生—中药学]