二元型多元回归法在正交试验中的应用

Application of regression analysis in optimum seeking method

目的 比较正交实验的常用分析和多元回归分析两种分析方法的异同。方法 本文以脂质体冻干制备条件正交实验为例。比较正交实验的常用分析方法和多元回归分析的结果。结果 多元回归最佳条件实验结果好于常用的直观分析/方差分析的最佳条件实验结果,但是两者结果差别不显著。结论 正交实验分析结合多元回归分析能更准确地反映实际情况,预测出最佳的因素水平,可更广泛地应用于优选试验中。

OBJECTIVE To compare the two different analysis methods direct analysis and regression analysis in optimum seeking method. METHODS The optimum conditions in the preparation of a liposome sample were selected with orthogonal design. Two different analysis methods direct analysis and regression analysis were used in orthogonal design and their results were contrasted. RESULTS The result obtained from multivariable regression analysis was better than that from direct analysis. But there was no statistical difference between the two analysis. CONCLUSTION The application of orthogonal design combined with multivariable regression method is practicable and successful in optimum seeking method.