Estimate the General Trend using Semi parametric Regression Models with a Forecast Value of GDP in Iraq

Abstract

In this article we attempt to estimate the general trend component of the gross domestic product (GDP) of Iraq for the period 1981-2010 using different estimation methods depending on the regression models (parametric with multiple regression,semiparametric models with single index model and partial linear model). We use exponential smoothing techniques (Single Linear, Double Linear, Triple Linear and Holt's Linear method) to forecast the Iraqi gross domestic product (GDP) for 2011. The absolute difference criterion has been used for the purpose of comparison between these methods, the results have proved preference the partial linear estimator when we use a triple linear exponential smoothing forecast for GDP, which was very close to real GDP value .