Use types of robust regression in treatment of the outliers in simple linear regression

Abstract EN

In the analysis of the simple linear regression there is only one independent variable.

My be there exist a problem because there are extreme points having higher remains (Residuals) in comparison with those of observations, for there are odd values (outliers ) in the groups of the observations.

Usually least square method are used so as to estimation the parameters of a model.

The analysis of this regressions begins with data designs of those which are remains in the opposite of an independent variety ; also, in the opposite of estimated value of Y to investigation of assumptions of that model so, the robust regression analysis in place of the least square method with outliers.

This research deals with four types of assessment applied to one example along with the use of the criterions (AIC and BIC) applied through goodness fit.

I used statistical program (SAS 9.1) in analyzing the results.