Related Experiment Video
Updated: Jul 12, 2026
HPLC Coupled with Chemical Fingerprinting for Multi-Pattern Recognition for Identifying the Authenticity of Clematidis Armandii Caulis
Published on: November 11, 2022
Variable selection for discriminating herbal medicines with chromatographic fingerprints
Fan Gong1, Bo-Tang Wang, Yi-Zeng Liang
1Research Center of Modernization of Chinese Herbal Medicines, Institute of Chemometrics & Intelligent Analytical Instruments, College of Chemistry and Chemical Engineering, Central South University, Changsha 410083, China. gongfan_csu@yahoo.com
Chemometric methods effectively reduce chromatographic fingerprint data for herbal medicine discrimination. Forward selection and key set factor analysis identified representative variables, improving pattern recognition.
Area of Science:
- Analytical Chemistry
- Chemometrics
- Pharmacognosy
Background:
- Chromatographic fingerprints of herbal medicines contain extensive data, with many variables lacking discriminatory information.
- Pattern recognition requires efficient variable selection to identify key features for accurate herbal medicine classification.
- Principal Component Analysis (PCA) is a common technique, but optimal variable reduction methods need evaluation.
Purpose of the Study:
- To investigate and compare chemometric approaches for selecting representative variables from chromatographic fingerprints of herbal medicines.
- To assess criteria for determining the optimal number of retained variables.
- To evaluate the effectiveness of reduced datasets for herbal sample discrimination.
Main Methods:
- Chemometric methods including forward selection and key set factor analysis were employed.
- Principal Component Analysis (PCA) was used for dimensionality reduction.
- Methods were assessed using Procrustes analysis, weighted similarity measures, and tri-variate PCA score plots.
- Bayes discrimination analysis was applied to the reduced datasets.
Main Results:
- Forward selection and key set factor analysis proved preferable for identifying representative variables.
- Procrustes analysis and weighted similarity measures were not indicative for extracting representative variables.
- High data matching between original and reduced datasets did not guarantee high prediction accuracy.
- Visual examination of PCA scores revealed limitations in separating all samples due to fingerprint complexity.
Conclusions:
- Chemometric variable selection is crucial for effective herbal medicine discrimination using chromatographic fingerprints.
- Specific methods like forward selection and key set factor analysis are recommended.
- Careful validation is needed, as data reduction does not automatically ensure improved predictive performance.
