تقدير انموذج الانحدار باستخدام طريقة المربعات الصغرى الموزونة الحصينة في حالة وجود مشكلة عدم تجانس التباين والقيم الشاذة معاً مع تطبيق عملي

Abstract

The non-Contrasted Homogeneity in an inclination model is considered as one of the supreme problems that one may confront while building such model, especially in economic phenomenon. Thus a method called (WLS)(Weighted Least Squares) which considered as a especial case from (GLS)(General Least  Squares), one of the important Alternative means for (OLS) , in case there is a problem. But the method (WLS) requires weight estimate also requires knowing the natural structure of non –Contrasted Homogeneity, which at most is un known. In (2001) a method made by a number of researchers, known as (MPV) This method can use weight estimate in a way made the (WLS) carry new good characteristics such as ; non-alignment and efficiency. Except that the (MPV) method and other Alternative methods for (OLS) failed enormously, to handle the problem of outliers, beside the problem of non-Contrasted Homogeneity. That made (MPV) method partial and useless. For that reason, in (2008) (robustness) (WLS) methods emerged and gave estimates that is not effected by outliers, and keep their wanted characteristics. This research took method (MPV) robustness (RMPV).The method (RMPV) did not interest Researchers who got all the attention, but both methods were highlighted and presented all the details for applying them on realistic data for a group of data of (income, expenditure). Also an unconventional robustness means were used to find out the outliers and non-Contrasted Homogeneity problem in case of an outlier. All these applications were executed by a ready-made programs like (statistics, Gretel ,  sas).And the they came to a conclusion that the results of (RMPV) are approximate. The (RMPV) was described as a complicated compared to the first method (RMPV).