Analysis of Robust L-Estimators for Location Parameter

International Conference "Probability Theory, Mathematical Statistics and Their Applications": the Collection of Scientific Papers (Proceedings of the International Conference) 2010
Oļegs Užga-Rebrovs, Gaļina Kuļešova

Methods of robust statistics are successfully used in the situations when certain sample values differ greatly from the rest of data. There are a lot of reasons why outliers of that kind occur in the data. In most of the cases outliers are erroneous data and - provided that methods of classical statistics are used- cause incorrect estimations of location parameter and relevant distribution scale parameter. Nowadays plenty of robust methods for estimating location and scale parameters are available. This paper analyses widespread kinds of robust L-estimators for location parameter and provides some general recommendations how to use robust statistics techniques.

Atslēgas vārdi
Order statistics, median, median of deviations from the median, fourth spread, cutoffs, robust mean values

Užga-Rebrovs, O., Kuļešova, G. Analysis of Robust L-Estimators for Location Parameter. No: International Conference "Probability Theory, Mathematical Statistics and Their Applications": the Collection of Scientific Papers (Proceedings of the International Conference), Baltkrievija, Minsk, 22.-25. februāris, 2010. Minsk: Belarus State University, 2010, 350.-355.lpp. ISBN 978-985-500-335-0.

Publikācijas valoda
English (en)