摘要
该文考察了紫外可见漫反射光谱结合化学计量学对当归及其掺伪品进行鉴别的可行性,并对40个当归纯品和75个掺伪品进行鉴别。首先采用Kennard-Stone(KS)分组法将数据集分成86个样品的训练集和29个样品的预测集。考察了训练集原始光谱以及连续小波变换及其结合SG平滑、数据标准化方法预处理后的光谱及主成分分析得分图,确定最佳预处理方法,最后采用偏最小二乘-判别分析建立鉴别模型。结果表明,对原始紫外可见光谱进行主成分分析,无法将当归和掺伪品分开。而经连续小波变换-SG平滑-数据标准化预处理后,不仅可区分当归纯品和其掺伪品,还可区分不同含量的掺伪品。经最佳预处理后仅使用一个因子建立偏最小二乘-判别分析模型,对预测集中当归及其掺伪品的鉴别正确率可达100%。因此,紫外可见光谱结合连续小波变换-SG平滑-数据标准化-偏最小二乘-判别分析模型,可以实现当归及其掺伪品的快速准确鉴别。
The feasibility of UV-visible(UV-Vis)diffuse reflectance spectroscopy combined with chemometrics for the identification of Angelicae Sinensis Radix(ASR)and its adulterants was investigated in this research.A total of 40 pure and 75 adulterated ASR samples were prepared.The Kennard-Stone(KS)grouping method was firstly used to divide the dataset into a training set with 86 samples and a prediction set with 29 samples.Then the raw spectra and the preprocessed spectra by continuous wavelet transform(CWT),CWT-SG smoothing,and CWT-SG smoothing-autoscaling are analyzed.Spectra and their corresponding principal component analysis(PCA)plots were used to determine the optimal preprocessed method.Finally,partial least squares-discriminant analysis(PLSDA)was used to build discrimination model.The results showed that PCA of the raw UV-Vis spectra could not separate ASR from their adulterated products.After the spectra preprocessed by CWT-SG smoothing-autoscaling,not only the pure ASR and its adulterants can be completely separated in the principal component space,but also the adulterants with different contents can be further identified.The optimal preprocessing method combined with PLS-DA can achieve 100%prediction accuracy for the identification of ASR with its adulterants in prediction set with one factor.Therefore,UV-Vis combined with CWT-SG smoothing-autoscaling PLS-DA is a rapid and accurate method for discriminating ASR and its adulterants.
作者
Mpango Prisca
陆占魁
李子涵
卞希慧
Prisca Mpango;LU Zhan-kui;LI Zi-han;BIAN Xi-hui(School of Environment Science and Engineering,Tiangong University,Tianjin 300387,China;USF(Tianjin)Water Technologies and Engineering Co.Ltd.,Tianjin 116023,China;NMPA Key Laboratory for Technology Research and Evaluation of Drug Products,Shandong University,Jinan 250012,China)
出处
《分析测试学报》
CAS
CSCD
北大核心
2024年第11期1858-1862,共5页
Journal of Instrumental Analysis
基金
“药物制剂技术研究与评价”国家药品监督管理局重点实验室开放课题资助(2023TREDP01)。
关键词
当归
掺伪品
可视化识别
紫外可见漫反射光谱
主成分分析
化学模式识别
预处理
Angelicae Sinensis Radix
adulterants
visual identification
ultraviolet-visible diffuse reflectance spectroscopy
principal component analysis
chemical pattern recognition
preprocessing
