Related Experiment Video
Updated: Apr 14, 2026
06:50
O-cresol Concentration Online Measurement Based On Near Infrared Spectroscopy Via Partial Least Square Regression
Published on: November 8, 2019
7.1K
A online NIR sensor for the pilot-scale extraction process in Fructus aurantii coupled with single and ensemble
Xiaoning Pan1, Yang Li2, Zhisheng Wu3
1College of Chinese Medicine, Beijing University of Chinese Medicine, South of Wangjing Middle Ring Road, Chaoyang District, Beijing 100102, China. panxiaoning1112@163.com.
Sensors (Basel, Switzerland)
|April 16, 2015
Summary
Bagging-partial least squares (PLS) improved online near-infrared (NIR) monitoring of Fructus aurantii extraction compared to PLS alone. This ensemble method offers a robust strategy for monitoring traditional Chinese medicine (CHM) processes.
Area of Science:
- Analytical Chemistry
- Chemometrics
- Process Analytical Technology (PAT)
Background:
- Online near-infrared (NIR) spectroscopy is crucial for real-time monitoring of pharmaceutical extraction processes.
- Traditional Chinese Medicine (TCM) extraction requires accurate quantification of active compounds for quality control.
- Partial Least Squares (PLS) regression is a common chemometric method for spectral data analysis.
Purpose of the Study:
- To evaluate the performance of partial least squares (PLS) and ensemble bagging-PLS models for online NIR monitoring of Fructus aurantii extraction.
- To compare different preprocessing and variable selection methods, including Synergy Interval PLS (SiPLS) and Moving Window PLS (MWPLS).
- To propose an effective ensemble method for online NIR monitoring of TCM extraction processes.
Main Methods:
- Online NIR spectroscopy coupled with High-Performance Liquid Chromatography (HPLC) as a reference method.
- Development and comparison of single PLS models and ensemble bagging-PLS models.
- Application of preprocessing techniques and variable selection methods like SiPLS and MWPLS.
Main Results:
- Bagging-PLS models demonstrated superior performance with a lower Root Mean Square Error of Prediction (RMSEP) compared to standalone PLS models.
- Ensemble methods, specifically bagging-PLS, enhanced the accuracy and reliability of quantitative analysis for naringin, hesperidin, and neohesperidin.
- SiPLS combined with bagging-PLS showed potential for further optimization.
Conclusions:
- Ensemble methods, particularly bagging-PLS, are highly effective for online NIR monitoring of pilot-scale Fructus aurantii extraction.
- The proposed online NIR monitoring strategy using ensemble methods is suitable for quality control of TCM extraction processes.
- This approach provides a reliable and efficient alternative to traditional offline analytical methods.
