Abstract
In the linear regression model, least-squares (LS) estimator is usually used for estimating regression parameters. LS is an unreliable and unfavourable estimator when multicollinearity and outlier problems exist in the model. Therefore, we propose a new robust regression estimator for solving the abovementioned problems simultaneously. We conducted theoretical comparisons and different scenarios of simulation studies, and a real-life dataset was employed to show the performance of the proposed estimator. Results showed that the proposed estimator performs better than other estimators when multicollinearity and outlier problems occur simultaneously in the model.
Disclosure statement
No potential conflict of interest was reported by the author(s).