Comparing traditional estimators and the estimators of (PSO) algorithm for some growth models of gross domestic product in Iraq

Abstract

Growth models are considered to be one of the most important statistical means that is widely used in the study of the behaviors of different phenomena throughout time, and the estimation of the parameters of these models is considered to be the key element that plays a major role in the inference about these models. In this paper, this problem will be discussed briefly. The aim of the paper is to compare the estimators of some traditional methods and the estimators of the Particle Swarm Optimization (PSO) algorithm for estimating the parameters of some growth models as well as building the best growth model for the Gross Domestic Product (GDP) in Iraq. The growth models that are used in this paper will include three linear models which are Polynomials of order (1, 3, and 5) as well as three nonlinear models which are (The Logistic Model, The Gompertz Model, and The Richards Model). It was concluded that the (PSO) algorithm was better than the traditional methods in estimating the parameters of the growth models, also the fifth degree polynomial was the best model to describe the (GDP) data in Iraq.