Performance of some new robust ridge estimators

Multicollinearity among predictors distorts the performance of ordinary least squared estimators. This performance further deteriorates when outliers are also present in the data. Robust ridge regression has been widely used to circumvent this joint problem. In this paper, some new robust ridge estimators based on m-estimators are proposed. The new estimators have been compared with existing ones through Monte Carlo simulation. The simulation results showed that new estimators outperform other considered estimators by considering the criteria of mean squared error. An application has also been presented to support the best performance of new estimators.