Research on compaction behavior of traditional Chinese medicine compound extract powders based on unsupervised learning
10.16438/j.0513-4870.2024-0996
- VernacularTitle:基于无监督学习研究中药复方提取物粉体的压缩成型特性
- Author:
Ying FANG
1
;
Yan-long HONG
2
;
Xiao LIN
3
;
Lan SHEN
3
;
Li-jie ZHAO
1
Author Information
1. Innovation Research Institute of Traditional Chinese Medicine, Shanghai University of Traditional Chinese Medicine, Shanghai 201203, China; Engineering Research Center of Modern Preparation Technology of Traditional Chinese Medicine of Ministry of Education, Shanghai University of Traditional Chinese Medicine, Shanghai 201203, China
2. Engineering Research Center of Modern Preparation Technology of Traditional Chinese Medicine of Ministry of Education, Shanghai University of Traditional Chinese Medicine, Shanghai 201203, China; Shanghai Innovation Center of Traditional Chinese Medicine Health Service, Shanghai University of Traditional Chinese Medicine, Shanghai 201203, China
3. Engineering Research Center of Modern Preparation Technology of Traditional Chinese Medicine of Ministry of Education, Shanghai University of Traditional Chinese Medicine, Shanghai 201203, China; School of Pharmacy, Shanghai University of Traditional Chinese Medicine, Shanghai 201203, China
- Publication Type:Research Article
- Keywords:
tablet;
compression behavior;
principal component analysis;
cluster analysis;
factor analysis;
unsupervised learning
- From:
Acta Pharmaceutica Sinica
2025;60(2):506-513
- CountryChina
- Language:Chinese
-
Abstract:
Direct compression is an ideal method for tablet preparation, but it requires the powder's high functional properties. The functional properties of the powder during compression directly affect the quality of the tablet. 15 parameters such as Py, FES-8KN, FES-12KN, FES-16KN, CR-8KN, CR-12KN, and CR-16KN were used as the characteristic variables in this paper. Unsupervised learning methods like principal component analysis, cluster analysis, and factor analysis were applied to analyze and classify the compression behavior data of 36 traditional Chinese medicine powders. The results showed that both different dimensionality reduction classification methods could effectively differentiate the compression behavior characteristics of 36 traditional Chinese medicine compound powders. The hierarchical cluster analysis results showed a better agreement with the actual compression phenomena of the powders, where group 1 was high elasticity and low compressibility, group 2 was easily compressed and hard to break, group 3 was excellent compressibility and compactibility. This study is expected to provide references and ideas for predicting the behavior of traditional Chinese medicine powders and the screening of tablet formulations.
