论文部分内容阅读
目的:使用BP神经网络结合正交试验优化安痒汤主要药效成分6-姜辣素和盐酸小檗碱提取工艺。方法:采用水回流提取法提取,以6-姜辣素和盐酸小檗碱的含量及干膏得率为评价指标,采用正交试验设计筛选提取工艺,并以正交试验层次分析法得到的综合评分实验数据作为反向传播神经网络的输入,评价指标的综合得分作为网络的输出,对主要影响因素进行仿真优化,得到最优工艺。结果:优化得到的提取工艺条件为10.4倍量水、提取3次、1.3h/次。结论:BP神经网络结合正交试验不需要增加试验次数就能分析试验中提取因素的变化规律和寻找最佳参数,该法可用于安痒汤提取工艺的优化。
OBJECTIVE: To optimize the extraction of 6-gingerol and berberine hydrochloride, the main active ingredient of Anke Soup, using BP neural network and orthogonal test. Methods: The water reflux extraction method to extract 6-gingerol and berberine content and the dry paste yield as the evaluation index, the orthogonal test design screening of the extraction process, and the orthogonal test by the analytic hierarchy process Comprehensive score experimental data is used as the input of backpropagation neural network and the comprehensive score of evaluation index as the output of the network. The main influencing factors are simulated and optimized to obtain the optimal process. Results: The optimized extraction process conditions were 10.4 times the amount of water, extracted three times, 1.3h / time. Conclusion: The BP neural network combined with orthogonal test can analyze the variation law of the extraction factors in the experiment and find the best parameter without increasing the number of tests. This method can be used to optimize the extraction process of Anche Tang.