A modified ridge m-estimator for linear regression model with multicollinearity and outliers |
| |
Authors: | Hasan Ertaş |
| |
Institution: | Artvin Coruh University, Faculty of Forestry, Department of Forest Engineering, Artvin, Turkey |
| |
Abstract: | The ordinary least-square estimators for linear regression analysis with multicollinearity and outliers lead to unfavorable results. In this article, we propose a new robust modified ridge M-estimator (MRME) based on M-estimator (ME) to deal with the combined problem resulting from multicollinearity and outliers in the y-direction. MRME outperforms modified ridge estimator, robust ridge estimator and ME, according to mean squares error criterion. Furthermore, a numerical example and a Monte Carlo simulation experiment are given to illustrate some of the theoretical results. |
| |
Keywords: | Biased estimator Multicollinearity Outliers Robust estimator |
|
|