R语言与正交试验对龙生蛭胶囊中水蛭提取工艺的优化比较

Comparison of R Language and Orthogonal Test to Optimize the Extraction Process of Hirudo in Longsheng Leeches Capsules

目的:比较分析正交试验法、R语言结合BP神经网络和遗传算法两种方法对水蛭提取工艺的优化。方法:以水蛭中水蛭素的抗凝血酶活性为指标,采用正交试验设计进行提取考察,同时运用R语言结合BP神经网络及遗传算法对正交试验所得数据进行目标寻优,比较分析两种方法从而得到水蛭的最佳提取工艺条件。结果:正交试验筛选后的最优工艺条件为:浸泡时间120 min、液料比10∶1、提取次数3次。R语言优化后的最佳工艺条件为:浸泡时间90 min、液料比12∶1、提取次数2次。网络预测值为328.37 U/g,验证试验平均值为327.19 U/g,相对误差为0.36%,表明网络预测性良好。结论:该实验同时应用正交设计和R语言数学建模对水蛭提取因素进行考察以及两种处理方法比较分析,R语言环境下数学建模在利用正交试验数据的同时,弥补了正交设计的不足,可用于水蛭素提取工艺的分析和预测,同时为实现中药有效成分的目标寻优提供新的参考。

Objective:To compare ortbogonal test method,R language combined with BP neural network and genetic algorithm in optimizing the extraction process of Hirudo.Methods:Taking the antithrombin activity of hirudin in Hirudo as the detecting indicators, the extraction technology was investigated by orthogonal experiment.In order to obtain the optimization extraction technology of hirudin in Hirudo,R language combined with BP neural network and genetic algorithm had been established to optimize the data of orthogonal test.Results :The optimal process was as follows,the soaking time was 120 min,the ratio of material to liquid was 10:1,extracte three times,and the optimization of extraction technology after R Language optimization was as follows,soaking time was 90 min;liquid-solid ratio was 12:1;the frequency was two times.Under the condition,the model predictive value was 328.37U/g,the experimental average value was 327.19U/g and the relative error was 0.36%,which presented a good network prediction.Conclusion:In this experiment,orthogonal design and R language with mathematical modeling are used to investigate the extraction factors of Hirudo,and two methods are compared and analyzed.The mathematical model which established in this experiment makes up for the deficiency of orthogonal test,and can be used to analyze and predict the extraction technology of hirudin in Hirudo ,which provides a new reference for screening optimization of Chinese medicine active components.