论文部分内容阅读
目的利用近红外漫反射技术和化学计量学的方法对天花粉药材的水分进行快速定量分析。方法用积分球漫反射测定样品的近红外漫反射光谱,定量模型的预处理方法为多元散射校正(MSC)和一阶微分处理,波长范围为5 855~5 760 cm~(-1)、5 200~4 832 cm~(-1)和4 512~4 292 cm~(-1),回归方法为偏最小二乘法(PLS)。结果定量模型的最佳主因子数为6,内部交叉验证均方差RMSECV为0.442,决定系数R2为0.9964,外部验证预测均方差(RMSEP)为0.364,系统精密度RSD为1.3%,方法精密度RSD为2.2%。结论建立预测模型对天花粉药材的水分进行测定时可行的,该法分析快速简便,结果准确。
Objective To quantitatively analyze the water content of TCS using near-infrared diffuse reflectometry and chemometrics methods. Methods Diffuse reflectance spectroscopy was used to measure the near-infrared diffuse reflectance spectra of the samples. The pretreatment methods for quantitative models were multiple scattering calibration (MSC) and first-order differential processing with the wavelength range of 5 855-5 760 cm -1 200 ~ 4832 cm ~ (-1) and 4 512 ~ 4 292 cm ~ (-1), respectively. The regression method was partial least squares (PLS). Results The best principal factor of the quantitative model was 6, RMSECV of internal cross-validation was 0.442, coefficient of determination R2 was 0.9964, RMSEP of external validation was 0.364, RSD of system precision was 1.3%, and method precision RSD 2.2%. Conclusion It is feasible to establish the prediction model for the determination of the water content of TCS. The method is rapid and simple, and the results are accurate.