| 基于高光谱技术的五味清浊制剂快速无损检测方法研究 |
| 投稿时间：2024-02-27 点此下载全文
|
| 引用本文：戴胜云,吴东雪,黄瑞,刘杰,乔菲,魏锋,连超杰,郑健.基于高光谱技术的五味清浊制剂快速无损检测方法研究[J].中国现代中药,2024,26(10):1790-1798 |
| DOI：10.13313/j.issn.1673-4890.20240227002 |
| 摘要点击次数: 42 |
| 全文下载次数: 0 |
| 作者中文名 | 作者英文名 | 单位中文名 | 单位英文名 | E-Mail |
| 戴胜云 |
DAI Sheng-yun |
中国食品药品检定研究院，北京 102629 |
National Institutes for Food and Drug Control, Beijing 102629, China |
|
| 吴东雪 |
WU Dong-xue |
中国食品药品检定研究院，北京 102629 |
National Institutes for Food and Drug Control, Beijing 102629, China |
|
| 黄瑞 |
HUANG Rui |
中国食品药品检定研究院，北京 102629 |
National Institutes for Food and Drug Control, Beijing 102629, China |
|
| 刘杰 |
LIU Jie |
中国食品药品检定研究院，北京 102629 |
National Institutes for Food and Drug Control, Beijing 102629, China |
|
| 乔菲 |
QIAO Fei |
中国食品药品检定研究院，北京 102629 |
National Institutes for Food and Drug Control, Beijing 102629, China |
|
| 魏锋 |
WEI Feng |
中国食品药品检定研究院，北京 102629 |
National Institutes for Food and Drug Control, Beijing 102629, China |
|
| 连超杰* |
LIAN Chao-jie |
中国食品药品检定研究院，北京 102629 |
National Institutes for Food and Drug Control, Beijing 102629, China |
|
| 郑健* |
ZHENG Jian |
中国食品药品检定研究院，北京 102629 |
National Institutes for Food and Drug Control, Beijing 102629, China |
|
|
| 基金项目:国家重点研发计划项目（2023YFC3504100）；药品监管科学体系建设项目（RS2024Z006） |
|
| 中文摘要:目的 采用高光谱技术结合化学计量学方法对蒙古族药五味清浊制剂中胡椒碱、桂皮醛和羟基红花黄色素A进行含量测定，实现快速、无损、全面的五味清浊制剂质量评估。方法 选取2023年度国家药品抽检计划抽检的五味清浊制剂样品33批次（五味清浊散11批次、五味清浊丸22批次），采集其高光谱数据；对比多元散射校正、基线校正、标准正态变换、光谱转化、矢量归一化、光谱降噪、卷积平滑（9）结合一阶导数、卷积平滑（11）结合一阶导数、卷积平滑（9）结合二阶导数和卷积平滑（11）结合二阶导数10种光谱预处理方法，蒙特卡罗无信息变量消除法、竞争性自适应重加权采样法（CARS）2种变量筛选方法，偏最小二乘法、最小二乘法-支持向量机（LS-SVM）2种建模方法用于胡椒碱、桂皮醛和羟基红花黄色素A含量与高光谱数据定量校正模型时的性能。结果 采用CARS建立的胡椒碱和桂皮醛的LS-SVM模型预测能力全局最优，模型的相对预测偏差（RPD）分别为9.2、6.0，验证集相关系数（rpre）分别为0.993 5、0.985 2，说明模型验证集与测定值具有良好的非线性关系，模型预测效果良好。采用羟基红花黄色素A原始光谱建立的LS-SVM模型性能全局最优，RPD和rpre分别为3.7、0.976 2。结论 采用高光谱技术结合化学计量学方法可以快速测定五味清浊制剂中胡椒碱、桂皮醛和羟基红花黄色素A含量，方法操作简便，可为五味清浊制剂的质量控制提供参考。 |
| 中文关键词:蒙古族药 五味清浊制剂 高光谱 变量筛选 蒙特卡罗无信息变量消除法 竞争性自适应重加权采样法 偏最小二乘法 最小二乘法-支持向量机 |
| |
| Rapid and Non-destructive Determination of Wuwei Qingzhuo Preparation Based on Hyperspectral Technology |
|
|
| Abstract:Objective To determine quantitatively the content of piperine, Cinnamic aldehyde, and hydroxysafflor yellow A in Wuwei Qingzhuo Preparation of Mongolian medicine by hyperspectral technology combined with chemometrics for rapid, non-destructive, and comprehensive quality assessment of Wuwei Qingzhuo Preparation.Methods Thirty-three batches of Wuwei Qingzhuo Preparation samples, including 11 batches of Wuwei Qingzhuo San and 22 batches of Wuwei Qingzhuo pills, were selected from the National Drug Sampling and Inspection Plan in 2023 to collect hyperspectral data of the samples. The hyperspectral quantitative calibration model was compared with the performance of 10 spectral pre-processing methods, including multiple scattering correction (MSC), baseline correction, normal variable transformation (SNV), spectrum transformation (ST), vector normalization, noise filtering, Savitzky-Golay (SG) 9-1st smoothing, SG9-2nd smoothing, SG11-1st smoothing and SG11-2nd smoothing, 2 variable selection methods, including competitive adaptive reweighted sampling (CARS) and Monte Carlo uninformative variable elimination (MC-UVE), as well as 2 modeling methods, including partial least squares (PLS) and least square support vector machine (LS-SVM) for the determination of piperine, Cinnamic aldehyde and hydroxysafflor yellow pigment A content.Results The LS-SVM models of piperine and Cinnamic aldehyde constructed by the CARS method had the best predictive ability, and the relative predictive deviations (RPD) of the models were 9.2 and 6.0, respectively; the correlation coefficients of the validation sets (rpre) were 0.993 5 and 0.985 2, indicating that the validation sets of the models had a good nonlinear relationship with the measured values and that the models were ideal for prediction. The LS-SVM model established by the raw spectra of hydroxysafflor yellow pigment A was the most effective, with the RPD and rpre of 3.7 and 0.976 2, respectively.Conclusion With hyperspectral technology combined with chemometrics, the content of piperine, Cinnamic aldehyde, and hydroxysafflor yellow pigment A in Wuwei Qingzhuo Preparation can be rapidly determined. The method was easy to operate and was useful for the quality control of Wuwei Qingzhuo Preparation. |
| keywords:Mongolian medicine Wuwei Qingzhuo Preparation hyperspectral variable selection MC-UVE CARS PLS LS-SVM |
|
查看全文
查看/发表评论 下载PDF阅读器 |
|
|
|
