R语言与正交试验分析处理优化黄芪中黄芪甲苷提取工艺的比较

Comparison of R Language and Orthogonal Experimental Analysis Processing to Optimize the Extraction Process of Astragaloside from Astragalus

目的通过R语言与正交试验分析,优化黄芪中有效成分黄芪甲苷的提取工艺。方法采用醇提法,以黄芪甲苷提取量为指标,采用四因素(乙醇浓度、液料比、提取时间及提取次数)三水平正交试验设计法分组提取。分别采用经典正交分析法、R语言结合BP神经网络及遗传算法进一步目标寻优法优化黄芪甲苷提取工艺,并进行优化工艺对比验证。结果正交分析法所得黄芪中黄芪甲苷的优化提取工艺为:乙醇浓度85%,液料比12∶1,提取3 h,提取次数2次,此条件下每3克黄芪中黄芪甲苷提取量为73.935 mg。R语言得到的优化提取工艺为乙醇浓度75%,乙醇与药材比例12∶1,提取3 h,提取3次,此条件下每3克黄芪中黄芪甲苷提取量为74.057 mg。工艺验证表明,正交分析优化方法RSD为1.37%,R语言处理优化方法RSD为0.944%,与预测结果相对误差仅有0.146%。结论R语言结合BP神经网络模型具有良好的预测性,为中药材提取及相关制剂工艺优化提供了一种新的优化方法。

Objective:To optimize the extraction process of astragaloside from astragalus by R language processing and orthogonal experimental analysis.Methods:With alcohol extraction,astragaloside extraction was taken as the test index,four factors(ethanol concentration,liquid material ratio,extraction time and extraction times)three levels of orthogonal experimental design methods were used to investigate the extraction process.An optimized extraction process was obtained with orthogonal analysis method.Meanwhile,R language combined with BP neural network model and the genetic algorithm was used to optimize the target to obtain another optimized extraction process.Finally,the optimum extraction process of astragaloside was obtained by comparing two optimized processes.Results:The optimization results for astragaloside extraction with orthogonal analysis method was ethanol concentration 85%,ethanol-herbs ratio of 12∶1,extraction time 2 h,extraction 3 times,the extraction of astragaloside in each 3 g of astragalus was 73.935 mg.The optimization extraction results of astragaloside obtained by R language processing was ethanol concentration 75%,ethanol-herbs ratio of 12∶1,extraction time 3 h,extraction 3 times,the extraction of astragaloside in each 3 g of astragalus was 74.057 mg.RSD of the orthogonal analysis method was1.37%;that of R language processing was 0.944%and the relative error was 0.146%.Conclusion:R language combined with BP neural network model has a good predictability,providing a new optimization method for Chinese herbal medicine extraction and optimization.