摘要
目的:建立准确预测反相高效液相色谱线性梯度条件下溶质保留时间的数学模型。方法:应用最小二乘法构建了溶质保留与梯度参数的多元线性回归模型,并采用了两种不同品牌色谱柱及两种中药提取液对该模型进行了验证。结果:两种不同色谱柱的实验数据均能很好拟合预测模型,对粉葛提取液中14个组分和金银花提取液中13个组分保留时间预测总的平均相对偏差分别为(1.64±1.25)%(n=84),(1.28±0.95)%(n=78),明显低于两种常用梯度保留公式的预测偏差。结论:所建立的回归模型计算简单且对复杂体系中弱保留及强保留组分均能准确地预测。
Aim：To build a model predicting accurately the retention time of components under linear gradient elution in RP-HPLC. Methods：The best fitting equation between the retention time and the three parameters was explored by a least squares regression. The regression model is validated by two columns from different manufacturers and two herbal medicine extracts. Results：The results on the two different columns showed solid goodness-of-fit measures and the average relative deviation of the prediction is （ 1.64 ± 1.25 ） % （ n = 84） and （ 1.28 ± 0.95 ） % （ n = 78 ） respectively for 14 components of Pueraria thomsonii extracts and 13 components of Flos Lonicerae extracts using the regression model. The relative deviations are obviously lower than from two known equations. Conclusion： The proposed model is simple and especially suited to the retention prediction of complex mixtures.
出处
《中国药科大学学报》
CAS
CSCD
北大核心
2007年第3期236-242,共7页
Journal of China Pharmaceutical University
关键词
梯度洗脱
保留预测
多元线性回归模型
gradient elution
retention prediction
multiple regression model
