Modified New Estimator in Linear Regression Model Under Heteroscedastic or Correlated Errors

In this paper, under regression model, a new bias estimator which is called Modified two-parameter unbiased estimator (MTURR) is presented to improve the efficiency of an estimator when the errors term suffers from heteroscedastic or correlated errors, and the independent variables have the multicollinearity problem. After getting the properties of the suggested estimator, its performance is compared with other competing estimators, depending on it has smaller mean squares error. The suggested estimator shows a clear advantage theoretically by obtaining the sufficient and necessary condition that enables it to do so. Its performance is associated with the estimated value of the two parameters k and d, which clearly affects the accuracy of MTURR. By using an applied example, we have shown from the tables and figures that the MTURR can be relied upon and used when the model suffers from these problems in its assumptions.