The Multivariate $S_n$ Estimator
Abstract
In this note we introduce the M estimator (for Multivariate ) a new robust estimator of multivariate ranking. Like MVE and MCD it searches for an -subset which minimizes a criterion. The difference is that the new criterion measures the degree of overlap between univariate projections of the data. A primary advantage of this new criterion lies in its relative independence from the configuration of the outliers. A second advantage is that it easily lends itself to so-called "symmetricizing" transformations whereby the observations only enter the objective function through their pairwise differences: this makes our proposal well suited for models with an asymmetric distribution. M is, therefore, more generally applicable than either MVE, MCD or SDE. We also construct a fast algorithm for the M estimator, and simulate its bias under various adversary configurations of outliers.
Cite
@article{arxiv.1208.3121,
title = {The Multivariate $S_n$ Estimator},
author = {Kaveh Vakili},
journal= {arXiv preprint arXiv:1208.3121},
year = {2014}
}
Comments
12 pages, 6 figures This paper has been withdrawn by the author due to a crucial error in equation 7 (said equation did not correspond with computer code used in the simulation, the computer code being the correct one.)