Comparison of Some Estimation Methods for the Multiple Linear Regression Model with Existence of Both Auto - Correlation and Multicollinearity Problems

Abstract

One problem that can arise in the analyzing multiple linear regression model like autocorrelation and multicolinearity. The use of the ordinary least squares (OLS) method for estimating the multiple linear regression model parameters that suffers from the existence of these problems leads to incorrect and undesirable estimates. In this paper we consider to use three estimating methods to deal with these problems at once which is ( Trenkler) method , (Kaciranlar) method and (Relaxation For Two Stage Ridge Regression Estimator ) and then compare these methods through the MSE where the simulation results showed that the best method (Kaciranlar’s) method