【作者】 田鸿儒； 李珊山； 王迪； 孟庆繁；

【Author】 TIAN Hong-ru~1 LI Shan-shan~2 WANG Di~1 MENG Qing-fan~(1*) 1.College of Life Science,Jilin University,Changchun 130012,China 2.First Hospital,Jilin University,Changchun 130021,China

【机构】 吉林大学生命科学学院吉林省； 吉林大学第一医院；

【摘要】 采用近红外光谱技术(NIRS)结合径向基神经网络(RBFNN)建立了同时测定利福平片、异烟肼片、吡嗪酰胺片、异福片和异福酰胺片5种抗结核药物中的利福平(RMP)、异烟肼(INH)和吡嗪酰胺(PZA)含量的定量分析模型。模型以交互验证均方根误差(RMSECV)为评价标准,选择最有效光谱区域、对网络结构参数和扩展常数进行优化,得到测定RMP,INH和PEA含量的最优定量分析模型。最优模型的RM- SECV分别为0.0127,0.0104,0.0078,应用最优模型对预测集样品中RMP,1NH和PZA含量进行预测,预测均方根误差(RMSEP)分别为0.0125,0.0109,0.0103。实验表明所建立的RBFNN定量分析方法精密度高,具有方便快捷、非破坏、无污染、可在线检测等优点,对替代常规方法作为药物原位或在线的精确分析具有重要的意义。更多还原

【Abstract】 In this paper,near infrared spectroscopy(NIRS)with radial basis function neural network(RBFNN)was applied to establish the analytical protocol for the simultaneous determination of the contents of rifampicin(RMP),isoniazide(INH)and pyrazinamide(PZA)in five varieties of anti-tuberculosis tablets respectively:Rifampicin tablets,Isoniazide tablets,Pyrazi- namide tablets,Rifampicin and Isoniazide tablets and Rifampicin Isoniazide and Pyrazinamide tablets.The mot mean square error of calibration set obtained by cross-validation(RMSECV)was used as the evaluation of the model.The optimal quantitative a- nalysis models for determination of RMP,INH and PZA was resulted from selecting the most effective spectral region,suitable topological parameter and spread constants.RMSECV of the optimum models were 0.0127,0.0104,0.0078.Using these op- timum models for determination of the RMP,INH and PZA contents in prediction set,the root mean square error of prediction set(RMSEP)were 0.0125,0.0109 and 0.0103.The experiment demonstrate that the RBFNN quantitative analysis method has many advantages,such as high precision,non-destructive,non-polluting,low cost,online detection and etc.It is great sig- nificance for in situ or online drug accurate analysis method and could substitute the conventional methods.更多还原