Bias robust estimation of scale (Q688401)

The problem of robust estimation of the scale of the location residuals is considered when the underlying distribution of the data belongs to a contamination neighborhood of a parametric location-scale family. A class of \(M\)-estimates of scale with general location is defined. It is shown that under certain regularity assumptions, these scale estimates converge to their asymptotic functionals uniformly with respect to the underlying distribution, and with respect to the \(M\)-estimate defining the score function \(\chi\). Expressions are established for the maximum asymptotic bias of the \(M\)- estimates of scale over the contamination neighborhood as a function of the fraction of contamination. Based on these expressions asymptotically min-max bias robust estimates of scale are constructed. In particular, it is shown that a scaled version of the median of absolute residuals about the median is approximately min-max bias robust within the class of Huber's Proposal 2 joint estimates of location and scale. A larger class of \(M\)-estimates of scale with general location is also considered. It is shown that a scaled version of the shortest half of the data is approximately min-max bias robust in this class. Moreover, the results of a Monte Carlo study are presented to show that the shortest half of the data has attractive finite sample size mean squared error properties for contaminated Gaussian data.