Optimization of extraction process for Shenxiong Huanglian Jiedu Granules based on AHP-CRITIC hybrid weighting method, grey correlation analysis, and BP-ANN.
10.19540/j.cnki.cjcmm.20250226.302
- Author:
Zi-An LI
1
;
De-Wen LIU
1
;
Xin-Jian LI
1
;
Bing-Yu WU
1
;
Qun LAN
1
;
Meng-Jia GUO
1
;
Jia-Hui SUN
2
;
Nan-Yang LIU
3
;
Hui PEI
3
;
Hao LI
2
;
Hong YI
1
;
Jin-Yu WANG
1
;
Liang-Mian CHEN
1
Author Information
1. Institute of Chinese Materia Medica, China Academy of Chinese Medical Sciences Beijing 100700, China.
2. Wangjing Hospital, China Academy of Chinese Medical Sciences Beijing 100102, China.
3. Xiyuan Hospital, China Academy of Chinese Medical Sciences Beijing 100091, China.
- Publication Type:Evaluation Study
- Keywords:
AHP-CRITIC hybrid weighting method;
Shenxiong Huanglian Jiedu Granule;
backpropagation-artificial neural network;
extraction process;
gray correlation analysis;
orthogonal test
- MeSH:
Drugs, Chinese Herbal/analysis*;
Neural Networks, Computer
- From:
China Journal of Chinese Materia Medica
2025;50(10):2674-2683
- CountryChina
- Language:Chinese
-
Abstract:
By employing the analytic hierarchy process(AHP), the CRITIC method(a weight determination method based on indicator correlations), and the AHP-CRITIC hybrid weighting method, the weight coefficients of evaluation indicators were determined, followed by a comprehensive score comparison. The grey correlation analysis was then performed to analyze the results calculated using the hybrid weighting method. Subsequently, a backpropagation-artificial neural network(BP-ANN) model was constructed to predict the extraction process parameters and optimize the extraction process for Shenxiong Huanglian Jiedu Granules(SHJG). In the extraction process, an L_9(3~4) orthogonal experiment was designed to optimize three factors at three levels, including extraction frequency, water addition amount, and extraction time. The evaluation indicators included geniposide, berberine, ginsenoside Rg_1 + Re, ginsenoside Rb_1, ferulic acid, and extract yield. Finally, the optimal extraction results obtained by the orthogonal experiment, grey correlation analysis, and BP-ANN method were compared, and validation experiments were conducted. The results showed that the optimal extraction process involved two rounds of aqueous extraction, each lasting one hour; the first extraction used ten times the amount of added water, while the second extraction used eight times the amount. In the validation experiments, the average content of each indicator component was higher than the average content obtained in the orthogonal experiment, with a higher comprehensive score. The optimized extraction process parameters were reliable and stable, making them suitable for subsequent preparation process research.
