期刊
MOLECULES
卷 27, 期 14, 页码 -出版社
MDPI
DOI: 10.3390/molecules27144640
关键词
Lonicerae japonicae Flos; Lonicerae Flos; ATR-FTIR; multivariate statistical analysis
资金
- Science and Technology Innovation Program of Hunan Province [2021RC4064]
- Science and Technology Project of Changsha [kq2004054]
- Natural Science Foundation of Hunan Province [2021JJ40551, 2022JJ50008]
- First-class Discipline Project on Chinese Pharmacology of Hunan University of Chinese Medicine [201803]
In this study, a model for discriminating LJF and LF was established and applied in the discrimination of their related prescriptions. The results showed good accuracy and applicability, suggesting that this method is a rapid and effective tool for the successful discrimination of LJF and LF and their related prescriptions.
LJF and LF are commonly used in Chinese patent drugs. In the Chinese Pharmacopoeia, LJF and LF once belonged to the same source. However, since 2005, the two species have been listed separately. Therefore, they are often misused, and medicinal materials are indiscriminately put in their related prescriptions in China. In this work, firstly, we established a model for discriminating LJF and LF using ATR-FTIR combined with multivariate statistical analysis. The spectra data were further preprocessed and combined with spectral filter transformations and normalization methods. These pretreated data were used to establish pattern recognition models with PLS-DA, RF, and SVM. Results demonstrated that the RF model was the optimal model, and the overall classification accuracy for LJF and LF samples reached 98.86%. Then, the established model was applied in the discrimination of their related prescriptions. Interestingly, the results show good accuracy and applicability. The RF model for discriminating the related prescriptions containing LJF or LF had an accuracy of 100%. Our results suggest that this method is a rapid and effective tool for the successful discrimination of LJF and LF and their related prescriptions.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据