论文部分内容阅读
目的采用近红外漫反射光谱技术结合偏最小二乘回归建立了测定头孢丙烯原料药中水分的定量分析模型。方法以头孢丙烯原料药为研究对象,用近红外光纤探头采集其近红外漫反射光谱,用费休氏法测定实际水分含量,运用偏最小二乘回归建立近红外光谱特征值与实际水分含量之间的校正模型,进而对预测集样品进行分析。结果在水分含量0.7%～9.8%范围内,校正集样品经内部交叉检验和检验集检验,相关系数(r)为0.999 6,内部交叉验证均方差(RMSECV)和外部检验均方差(RMSEP)分为0.40和0.41。结论所建模型具有较好的预测能力,结果可靠,为头孢丙烯原料药的质量控制提供了一种快速简便的检测方法。
Objective To establish a quantitative analysis model for determination of water in cefprozil raw material by near-infrared diffuse reflectance spectroscopy combined with partial least-squares regression. Methods The cefprozil crude drug was used as the research object, near-infrared optical fiber probe was used to collect its near-infrared diffuse reflectance spectroscopy, the actual water content was determined by Fisher’s method, partial eigen-regression was used to establish the characteristic value and the actual moisture content Between the calibration model, and then to predict the sample set for analysis. Results In the range of 0.7% ~ 9.8% of water content, the calibration set samples were tested by internal cross test and test set, the correlation coefficient (r) was 0.999 6, RMSECV and RMSEP 0.40 and 0.41. Conclusion The model has good predictive ability and reliable results, which provides a quick and easy method for the quality control of cefprozil raw material.