دراسة مقارنة لالنحدار الحصين : دراسة باستخدام المحاكاة

Abstract EN

The idea of the research is to find robust estimator for linear regression model in the case of presence of outliers value that make traditional methods like OLS sensitive towards outliers and non resistance, an estimation methods are used traditional method OLS and robust method M, MM, S, R, LTS, LMS, GM we find that OLS, M, MM methods has breakdown point zero were other robust methods has 0.5 breakdown point which is R, S, LTS, LMS, GM and there variance not high and minimum MSE were the best to use them in the stat of outliers for all size of samples.