Another proposal about the new two-parameter estimator for linear regression model with correlated regressors
Shakeel Ahmad, Muhammad Aslam
Abstract
In this article, we present a new general class of biased estimators which includes some popular estimators as special cases and discuss its properties for multiple linear regression models when regressors are correlated. This proposal is based on some modification in the existing new two-parameter estimator. Performance of the proposed estimator is compared with many of the leading estimators, using the mean squared error matrix criterion, mitigating the adverse effects of multicollinearity. An extensive simulation study has been provided with a numerical example to illustrate the superiority of the proposed estimator.
Topics & Concepts
MulticollinearityEstimatorMean squared errorMathematicsLinear regressionStatisticsInvariant estimatorMinimum-variance unbiased estimatorLinear modelMinimax estimatorComputer scienceApplied mathematicsEconometricsAdvanced Statistical Methods and ModelsSpectroscopy and Chemometric AnalysesOptimal Experimental Design Methods