Using a semi-parametric regression model to study the most important factors affecting the gross domestic product of oil prices for the period (1980-2020)
Main Article Content
In recent years, researchers' focus at semi-parametric regression models has increased, which is the effect of the integration between the parametric and non-parametric regression models. These models received a clear interest in most research and studies that take a more advanced and sophisticated nature in the process of accurate statistical analysis aimed at obtaining capabilities with a high level of efficiency.
In this research, we deal with methods to address some problems, including the problem of autocorrelation between random errors, by relying on a semi-parametric partial regression model, which consists of two parts, a parametric part and semi-parametric one and merge the two parts in one of semi-parametric. This new regression contain strict constraints of the first and high flexibility of the later. Then we estimate the parameters of the parametric part using the general least squares method (SGLS) and the double residuals method (DR) , also the Generalize Difference Based Liu Method) GDBL) . As for the nonparametric regression model, we applied the core regression method using the Nadaray - Watson estimator and the Smoothing Spline method.
The most important goal of this research is to create a semi-parametric regression model in which the methods used, both parametric and non-parametric, are applied, and make comparison to indicate the preference by relying on the Mean Squared Error (MSE) scale.