Abstract
The paper uses bootstrap approach to analyze reliability of Least Trimmed Squares, and Least Weighted Squares robust estimators. In Monte Carlo simulation it measures performance of bootstrap con dence intervals.
It also proposes use of further examination of bootstrap population for detecting character and severity of contamination present in data. The main contribution of paper lays in developing basic statistical inference tools for the two estima- tors, for which standard asymptotic approach following fr