摘要
以两组分混合物模型体系为例初步研究了多组分体系光谱重叠程度、校正集样品数、测量波长数、组分浓度之比及随机误差等因素对偏最小二乘法和主成分分析计算结果的影响。
In this paper the simulative system of two-component mixture is used to illus-trate the effect of parameters: the spectra over-lapping, the number of calibration sam-ples, the number or wavelengths, the rate of component concentration and random erroron the prediction of results calculated by PLS and PCA.The spectral data of two-compo-nent system of food colorants also is used to illustrate how to apply simulative system toselect the parameters to determine and predict the error.
关键词
成分
化学分析
分光光度法
spectrophotometry
principal component anlysis
partial least squares
