响应面法(RSM)及响应面-人工神经网络(RSM-ANN)混合模型优化地耳草总黄酮提取工艺  

Application of Response Surface Methodology(RSM)and RSM-Artificial Neural Network(RSM-ANN)Hybrid System for the Optimization of Total Flavonoids Extraction from Hypericum japonicum

作　　者：姜登钊 刘可越 曾明 刘文波 Jiang Dengzhao;Liu Keyue;Zeng Ming;Liu Wenbo(College of Pharmacy and Life Sciences,Jiujiang University,Jiujiang 332005,China)

机构地区：[1]九江学院药学与生命科学学院,江西九江332005

出　　处：《中国野生植物资源》2024年第12期19-26,共8页Chinese Wild Plant Resources

基　　金：“江西省高层次高技能领军人才培养工程”项目(B00381)。

摘　　要：目的:优化地耳草总黄酮的超声波辅助提取工艺。方法:使用超声提取法提取地耳草总黄酮,分别应用响应面法(RSM)及响应面-人工神经网络(RSM-ANN)混合模型优化其总黄酮提取工艺。结果:RSM-ANN模型优化工艺误差更小且获得的提取率更高,表现出更高的预测准确度。地耳草总黄酮最佳提取工艺为液料比38∶1(mL/g),乙醇浓度61%,超声功率250W,超声时间34min,总黄酮提取率为5.43%。结论:RSM-ANN模型能较好优化地耳草总黄酮的提取工艺,展现出卓越的预测能力,具有广阔的应用前景。Objective:To optimize the ultrasonic-assisted extraction process for total flavonoids from Hypericum japonicum.Methods:The total flavonoids from H.japonicum were extracted by ultrasonic extraction.The extraction process was optimized separately using response surface methodology(RSM)and a hybrid model of RSM coupled with artificial neural network(RSM-ANN).Results:It was found that the RSM-ANN model resulted in smaller errors in the optimized process and achieved a higher extraction rate,demonstrating greater prediction accuracy.The optimal extraction process for total flavonoids from H.japonicum,as determined by the RSM-ANN model,involved a liquidto-solid ratio of 38∶1(mL/g),ethanol concentration of 61%,ultrasound power of 250 W,and ultrasound duration of 34 min,yielding an extraction rate of 5.43%.Conclusion:The RSM-ANN model was utilized to effectively optimize the extraction process of total flavonoids from H.japonicum.An outstanding predictive capability was showcased,suggesting promising prospects for further development.

关 键 词：地耳草 总黄酮 超声提取 响应面法 人工神经网络 

分 类 号：R932[医药卫生—生药学]