论文部分内容阅读
采集还原型谷胱甘肽片(素片)的近红外漫反射光谱,以HPLC测定结果为参考值,运用偏最小二乘法(PLS)建立含量预测模型。所建模型的交叉验证均方根误差(RMSECV)为0.024 13,预测均方根误差(RMSEP)为0.008 57,预测集的相关系数(R)为0.972 54,证明近红外漫反射光谱法结合多元分析技术可快速测定还原型谷胱甘肽片的含量,并可望用于生产过程控制。
The near-infrared diffuse reflectance spectra of the reduced glutathione tablets (plaques) were collected. The results of HPLC determination were used as a reference value, and the content prediction model was established by partial least squares (PLS). The RMSECV of the model was 0.024 13, the root mean square error of prediction (RMSEP) was 0.008 57, and the correlation coefficient (R) of the prediction set was 0.972 54. It was proved that NIRS combined with multivariate Analytical techniques allow rapid determination of reduced glutathione pellet content and are expected to be used in manufacturing process control.