Perbandingan Metode Dua Tahap Durbin dan Theil-Nagar dalam Mengatasi Masalah Autokorelasi.

Abstract

In forming a linear regression model, there are several classical assumptions that must be satisfied. One of the classical assumptions states there is not exist autocorrelation that occurs between disturbance or error. If autocorrelation occured in linear regression model, the regression coefficient estimators no longer have minimum variance so that the method necessary to overcome them. Durbin two-stage and Theil-Nagar method are several methods that can be used to solve autocorrelation problem. In the research, Theil-Nagar method gives the better results compared to Durbin two-stage method because the resulting value of the autocorrelation coefficient have been close to zero.