Encontrado 1 documento, a visualizar página 1 de 1

A non parametric robust method for the detection of outliers in linear models

The detection of outliers for the standard least squares regression is a problem which has been extensevily studied. Lad Regression diagnostics offers alternative approaches whose main feature is the robustness. In this work, we propose a nonparametric method for detecting outliers in LAD regression models and compare t to other classical methods.