B0664
Title: Asymptotics for M-estimators in linear models with increasing dimension
Authors: Ezequiel Smucler - IMAS, CONICET (Argentina) [presenting]
Abstract: Existing results on the asymptotic properties of M-estimators of regression in linear models with increasing dimension are reviewed. We prove consistency and asymptotic normality for a class of resdescending M-estimators of regression assuming $p/n \rightarrow 0$ and $p^3/n \rightarrow 0$, respectively. This class is wide enough to include several popular high breakdown point regression estimators such as S and MM-estimators.
Journal