Hassan H. Razaqa, Hayder R. Talibb, Reem T. Kamilc
a University of Thi-Qar, College of Administration and Economics, Department of Economics, Iraq.
b University of Sumer, College of Administration and Economics, Department of Statistics, Iraq.
c University of Baghdad, College of Physical Education and Sports Sciences for Girls, Department of Theoretical Sciences, Iraq.
This research presents a comparative study of two estimation methods for a semiparametric regression model in the presence of autocorrelation. The model under consideration is a semiparametric partial linear regression model comprising both parametric and nonparametric components. The two approaches evaluated are: the Semiparametric Generalized Least Squares Estimators (SGLSE) method and the Least Squares Estimators Method (LSEM).
Simulation experiments were conducted to assess the performance of both estimators. The results indicate that the SGLSE method provides superior estimation accuracy compared to the LSEM method, as confirmed by the smaller Mean Squared Error (MSE) values obtained. This demonstrates the effectiveness of the semiparametric generalized least squares approach in handling models affected by autocorrelation.
Keywords: Semiparametric partial linear regression model; Semiparametric generalized least squares estimators; Least squares estimators; Autocorrelation problem.
Razaq, H. H., Talib, H. R., & Kamil, R. T. (2023). A comparison between two methods of estimating a semi-parametric regression model in the presence of the autocorrelation problem. Advances in the Theory of Nonlinear Analysis and Its Application, 7(4), 54–59.