All known robust location and scale estimators with high breakdown point formultivariate samples are very expensive to compute. In practice, this computation has tobe carried out using an approximate subsampling procedure. In this article we describe analtemative subsampling scheme, applicable to both the Stahel-Donoho estimator and theminimum volume ellipsoid estimator, with the property that the number of subsamplesrequired can be substantially reduced with respect to the standard subsampling proceduresused in both cases. We also discuss some bias and variability properties of the estimatorobtained from the proposed subsampling process
展开▼
