Adaptive Estimations of the Shift and Scale Parameters

Abstract

New robust adaptive estimations (AE) of the shift and scale parameters are synthesized based on the weighted maximum likelihood method. The potential AE efficiencies are determined on classes of local and global Tukey distributions (supermodels). They are compared with the efficiencies of classical and robust estimations on the local and global supermodels of distributions. It is demonstrated that the efficiencies of the adaptive estimations of the shift and scale parameters are significantly higher than of the classical parametrical and robust estimations.