Paper Image

Estimating linear models with unknown heteroscedastic measurement error

Published on:

21 October 2023

Primary Category:


Paper Authors:

Linh H. Nghiem,

Cornelis J. Potgieter


Key Details

Proposes estimator combining moment correction and phase function methods

Allows for heteroscedastic, non-normal measurement errors

Estimator consistency and asymptotic normality established

Simulation studies show competitive or improved performance

Illustrated on nutrition data from a health survey

AI generated summary

Estimating linear models with unknown heteroscedastic measurement error

This paper proposes a new method to estimate the parameters of a linear regression model where some predictors are measured with error. The errors are assumed to be independent across observations, but can have different variances for each observation. The new method combines existing approaches, using both moment information and asymmetry properties of the data, to achieve an efficient estimator. Simulation studies show it outperforms standard methods, especially when errors are non-normal.

Answers from this paper


No comments yet, be the first to start the conversation...

Sign up to comment on this paper

Sign Up