论文部分内容阅读
目的:探讨神经网络光度法用于复方制剂的含量测定。方法:训练集为按L_(25)(5~6)正交表制备的25组标准混合液的吸光度数据和各组分的浓度数据,混合液中各组分的5个浓度水平分别为80％、90％、100％,110％和120％。预报集采用复方制剂的吸光度数据。网络的输入为混合物的吸光度,网络的输出为各组分的浓度。分别用径向基函数网络和Levenberg-Marqurdt优化算法的BP网络处理数据。结果:复方阿司匹林片和联磺甲氧苄啶片的紫外分光光度法测定结果表明,径向基函数网络在网络训练时间和测定精度等方面好于Levenberg-Marqurdt优化算法的BP网络。结论:径向基函数网络光度法测定复方制剂简便,准确。
Objective: To explore the method of neural network spectrophotometry for the determination of compound preparations. Methods: The training set consisted of the absorbance data of 25 standard mixed solutions prepared according to the L_ (25) (5 ~ 6) orthogonal table and the concentration data of each component. The five concentration levels of each component in the mixed solution were 80 %, 90%, 100%, 110% and 120%. The forecast set uses the absorbance data of the compound preparation. The input of the network is the absorbance of the mixture, and the output of the network is the concentration of the individual components. The data were processed by BP neural network with RBF network and Levenberg-Marqurdt optimization algorithm respectively. Results: The results of UV spectrophotometry of compound aspirin tablets and trimethoprim tablets showed that radial basis function network was better than Levenberg-Marqurdt BP algorithm in network training time and measurement precision. Conclusion: Radial Basis Function Spectrophotometry is simple and accurate for the determination of compound preparations.