校正变换矩阵法用于多组分食用色素的同时测定研究

目的]对胭脂红、苋菜红、日落黄3种食用色素进行不经分离的同时测定。方法]将校正变换矩阵法与可见分光光度法相结合,使用交叉验证法选择主因子数建立了校正模型。结果]胭脂红的回收率在94.98%～113.69%,苋菜红的回收率在79.25%～104.96%,日落黄的回收率在93.22%～109.27%,13个样品中3种组分的表观预测误差（AET）分别为0.392 72、0.272 930、.200 80。结论]校正变换矩阵法用于混合食用色素不经分离同时测定是可行的,其预测准确性与偏最小二乘回归法没有显著性差异。

Study on Application of Calibration Transform Matrix Method with Spectrophotometry for Simultaneous Determination of Food Coloring in Artificial Mixtures

Objective] To simultaneously determine three main food coloring mixtures in the carminum,amaranth and sunset yellow without any separation.Methods] The calibration transform matrix(CTM) multivariate regression combined with spectrophotometry method was established,and calibration sample was established by using cross-validation method to select the appropriate number of factor.Result]The analyte recoveries from synthetic mixtures ranged from 94.98% to 113.69% for carmine,from 79.25% to 104.96% for amaranth,from 93.22% to 109.27% for sunset yellow,respectively.The root mean square of the apparent error in the test vector(AET) for the three analytes from 13 samples were 0.392 72,0.272 93,0.200 80,respectively.Conclusion] It is practicable to use calibration transform matrix method in determining mixed food coloring without any separation,and the forecasting accuracy has not significant difference between calibration transform matrix method and partial least square(PLS) regression.