二次回归分析在正交设计中的应用

The Application of Quadratic Regression Analysis in Orthogonal Design

正交设计是立足于方差分析模型的,通过方差分析来估计模型中的未知参数。但是如果因素之间存在交互作用,当因素数目和因素的水平数增加的时候,未知参数的个数将呈指数增长,使得试验次数也大大增加。利用回归分析的方法需要估计的未知参数将会大大减少。本文通过二次回归来建立了正交设计的模型。

Orthogonal design which is based on variance analysis model estimates unknown parameters using variance analysis. If there are some interactions between factors and when the number of factors and the number of levels increase, the number of unknown parameters will increase according to exponential curve and will greatly increase the times of the experiment. Unknown parameters which need to estimate will greatly decrease by using regression analysis. In this paper, the model of orthogonal design is gained by using quadratic regression analysis.