QUASI-EMPIRICAL BAYES MODELING OF MEASUREMENT ERROR MODELS AND R-ESTIMATION OF THE REGRESSION PARAMETERS

Authors

  • A.K.Md. Ehsanes Saleh

Keywords:

ME model, asymptotic relative eciency, rank estimators, quasi-empirical Bayes model, Theil-Sen estimator, asymptotic normality.

Abstract

This paper deals with the R-estimation of the regression parameters of a mea-
surement error model: yi = 0+1xi+ei and x0i
= xi+ui; i = 1; : : : ; n. By com-
bining the two sets of the information, an emaculate regression model is obtained
using \quasi-empirical Bayes" estimates of the unknown covariates x1; : : : ; xn.
The model produces consistent estimates of the attenuated slope and the inter-
cept parameters and applies to broad range of regression problems. Asymptotic
properties of the R-estimators are provided based on the \quasi-Bayes regression
model". Some simulated results are presented as evidence of the performances of
the estimators.

Downloads

Issue

Vol. 44 No. 2 (2010)

Section

Articles