Robust Location Estimation with L2 Distance

Keywords: Robust, L2 Distance

Abstract: Many robust location parameter estimators exist and have been studied. The performance of these estimators has been examined using theory, simulations, and real observations. This study considers the eleven robust estimators and the 24 real datasets presented by Stigler in 1977. We add an additional robust estimator that is based on L2 distance and the details of this method are discussed. Using each of the twelve methods, a location estimate is obtained for each of the 24 datasets. The datasets contain measurements of physical constants and performance is measured by the relative closeness to the true value. We then give a comparison based on the relative performance of each estimator.