论文部分内容阅读
The determination of enantiomeric composition by partial least squares(PLS) modeling of UV-vis spectral data was investigated for samples of phenylalanine(phe) using sucrose as a chiral auxiliary.And a new data preprocess method,reference band normalization,was introduced to eliminate the spectral variations due to the changes of total concentration of phe.The determination coefficient(R~2) and the standard error of calibration set(SEC) of 13 standard samples are 0.9987 and 0.0128 respectively.The standard error of validation set(SECV) of 7 validation samples is 0.0049.The standard error of predict(SEP) of 6 blind samples for evaluating the robustness of the model is 0.0366.The regression model is robust to determine enantiomeric composition when total concentration varied.It is demonstrated that the reference band normalization is a convenient method of compensating for variations in total concentrations without knowing that in advance.
The determination of enantiomeric composition by partial least squares (PLS) modeling of UV-vis spectral data was investigated for samples of phenylalanine (phe) using sucrose as a chiral auxiliary. And a new data preprocess method, reference band normalization, was introduced to eliminate the spectral variations due to the changes of total concentration of phe. The determination coefficient (R ~ 2) and the standard error of calibration set (SEC) of 13 standard samples are 0.9987 and 0.0128 respectively.The standard error of validation set (SECV) of 7 validation samples is 0.0049. The standard error of predict (SEP) of 6 blind samples for evaluating the robustness of the model is 0.0366. The regression model is robust to determine enantiomeric composition when total concentration varied. It is demonstrated that the reference band normalization is a convenient method of compensating for variations in total concentrations without knowing that in advance.