استعمال طريقتي التقدير الحصينة المربعات الصغرى المشذبة و M لتقدير معلمات أنموذج الانحدار الخطي بأخطاء عشوائية غير متجانسة التباين

Abstract EN

The problem of heterogeneity of the random error terms variance for the regression model is one of the most important standard problems that most economic and financial data suffer from.

the reliance on Ordinary Least Squares (OLS) estimators in this case lead to shaded results, although these estimators remain unbiased, but they do not have the least variance.

therefore, it is necessary to search for alternative methods that are robust against to the problem of heterogeneity of the variance of the random error terms.

in this research, two robust methods were used, the first is the trimmed Least Squares (TLS) method, and the second is the (M) method.

then a comparison was made between these two methods, through the application on realistic data representing the quantities produced and exported of crude oil in Iraq.

the estimation results for both methods resulted in the preference of the trimmed least squares method, based on the mean of squared error (MSE) criterion.