引入模型误差的线性回归模型

The introduction of model error to linear regression model

在线性回归模型中,经常出现病态问题,严重影响计算结果的精度。文中提出一种引入模型误差求解部分参数的方法,此方法可以改变病态问题,提高部分参数的求解精度,且通过算例检验这种方法的可行性,并探讨它的使用范围。

Ill-conditioned problems widely exist in general linear regression model and seriously affect the accuracy of results. In order to change pathological problems, a new method is presented by solving some of the parameters with the introduction to the function model error. The solution enhances the accuracy of parameters, and the feasibility of the method is tested by using an example.