The research about independence of random errors in quasi-linearity regression model

How to construct a regression model with good operability is the core of regression analysis. In this paper, by analyzing the efficiencies of the existing regression methods, we establish quasi-linear regression model (denoted by QRM for short) which has good structure and approximation properties; further we put forward the plan to test the independence of random errors based on residual analysis, and MATLAB gives the strategies; finally, we analyze the effectiveness of these discussions by an illustrative example.