On the performance of two-parameter ridge estimators for handling multicollinearity problem in linear regression: Simulation and application
M. S. Khan, Amjad Ali, Muhammad Suhail, Fuad A. Awwad, Emad A. A. Ismail, Hijaz Ahmad
Abstract
The inability of ordinary least square estimators against multicollinearity has paved the way for the development of various ridge-type estimators, which are recently classified as one-parameter and two-parameter ridge estimators. In this paper, we offer some efficient two-parameter ridge estimators and evaluate their performance through a simulation study by using the minimum mean square error criterion. Under most of the simulation conditions, our proposed estimators outperformed the existing estimators. Finally, two real-life datasets are used to demonstrate the applications of our proposed estimators.
Topics & Concepts
MulticollinearityEstimatorRidgeExtremum estimatorMean squared errorOrdinary least squaresStatisticsLinear regressionRegressionM-estimatorMathematicsEstimation theoryComputer scienceGeologyPaleontologyAdvanced Statistical Methods and ModelsAdvanced Statistical Process MonitoringStatistical Methods and Inference