Computing Robust Regression Estimators: Developments since Dutter (1977)

Abstract

The proposal ofMestimators for regression (Huber, 1973) and the development of an algorithm for its computation (Dutter, 1977) has lead to an increased activity for further research in this area. New regression estimators were introduced that combine a high level of robustness with high efficiency. Also fast algorithms have been developed and implemented in several software packages. We provide a review of the most important methods, and compare the performance of the algorithms implemented in R .