Optimization of the Baicalin Extraction Process and Study of Its Antioxidant Activity Using the Box–Behnken Response Surface Methodology and Genetic Neural Networks
Separations · 30.10.2026 · 13 (11) · 319
DOI: 10.3390/separations13110319 ↗Kurz gefasst
Fachartikel in der Zeitschrift Separations, online erschienen am 30.10.2026. Verfasst von Zebin Yu und 7 weiteren Autor:innen. Die Arbeit nennt 22 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Research Project of Zhejiang Chinese Medical University, Zhejiang Medicine and Health Science and Technology.
Abstract (Originalsprache)
Purpose
The extraction process for baicalin was optimized using a Box–Behnken response surface design combined with a GA-BP neural network, achieving efficient and energy-saving extraction of baicalin.
Method
To achieve the optimal combination of extraction conditions (X) and baicalin content (Y), the amount of baicalin extracted was used as the evaluation criterion. Using the Box–Behnken response surface methodology, combined with a three-dimensional response surface to analyze the interactions among factors, we performed quadratic polynomial regression modeling. Based on the response surface, a GA-BP neural network was constructed to establish a quantitative relationship between process conditions and baicalin content. The GA-BP neural network was then used to identify the optimal process parameters, which were compared with the results obtained from the response surface method to ultimately determine the optimal process. The antioxidant activity of baicalin was detected using DPPH.
Result
The optimal process parameters are as follows: number of extraction was three, 60 mL of solvent, an extraction temperature of 72.5 °C, an extraction time of 1.5 h, and a soaking time of 2.5 h. Under these conditions, the baicalin content was 749.0 mg, corresponding to an extraction rate of 14.98%. When the concentration of baicalin was 0.8 mg/mL, its ability to scavenge DPPH free radicals reached 82.83%, indicating that it has good antioxidant activity.
Conclusions
This study combined response surface methodology (RSM) with a hybrid GA-BP neural network model to determine the optimal extraction parameters for baicalin. The GA-BP neural network model demonstrated superior performance to RSM in both prediction and optimization, confirming the immense potential of artificial intelligence methods in complex traditional Chinese medicine extraction systems. This work provides innovative methodological support for advancing the intelligent and precision-oriented development of traditional Chinese medicine manufacturing processes.
Förderung
- Research Project of Zhejiang Chinese Medical University
- Zhejiang Medicine and Health Science and Technology
Zitieren
APA
Yu, Z., Shan, C., Shi, R., Ye, C., Bao, R., Zhang, B., Zhang, Y., & Xu, S. (2026). Optimization of the Baicalin Extraction Process and Study of Its Antioxidant Activity Using the Box–Behnken Response Surface Methodology and Genetic Neural Networks. Separations, 13(11), 319. https://doi.org/10.3390/separations13110319
Vancouver
Yu Z, Shan C, Shi R, Ye C, Bao R, Zhang B, et al. Optimization of the Baicalin Extraction Process and Study of Its Antioxidant Activity Using the Box–Behnken Response Surface Methodology and Genetic Neural Networks. Separations. 2026;13(11):319. doi:10.3390/separations13110319
BibTeX
@article{Yu2026,
author = {Yu, Zebin and Shan, Chuchu and Shi, Rui and Ye, Changyi and Bao, Rou and Zhang, Bin and Zhang, Yuyan and Xu, Shan},
title = {Optimization of the Baicalin Extraction Process and Study of Its Antioxidant Activity Using the Box–Behnken Response Surface Methodology and Genetic Neural Networks},
journal = {Separations},
year = {2026},
volume = {13},
number = {11},
pages = {319},
doi = {10.3390/separations13110319}
}
EUVIMED