摘要
利用改进BP网络同时测定了镇咳新药联邦镇咳露中的三组分含量,考察了网络拓扑模型对于分析结果的影响,通过选择合适的网络参数,同时测定模拟样品中的三组分,其中磷酸可待因、扑尔敏和盐酸麻黄碱的平均回收率分别为100.03%,100.10%和100.07%,相对标准偏差分别为0.54%,0.65%,0.51%;与非线性偏最小二乘法结果相比,本文结果更为满意。
The effect of network topologic structure on the analytical result and calculating ability in the determination of the three components in Lian bang Zhenkelu by using modified BP network is studied. Experimental result indicates that proporal network parameter helps to determine simultaneously three components in the simulated sample. The average recoveries of codeine phosphas, chlorpheniraminum and ephedrini hydrochloridum only 100.03%, 100.10%, 9910% respectively. The relative standard deviations are 0.54%, 0.65% and 0.51%, respectively.
出处
《中国药科大学学报》
CAS
CSCD
北大核心
1997年第6期342-344,共3页
Journal of China Pharmaceutical University
关键词
人工神经网络
BP算法
联邦镇咳露
止咳药
Artificial neural network, Codeine phosphas, Chlorpheniraminum, Ephedrini hydrochloridum
Lianbang Zhengkelu