Estimation of distribution functions in measurement error models
Authors:
- I. DattnerIDI. Dattner
- B. ReiserBRB. Reiser
Preprint
Abstract
Many practical problems are related to the estimation of distribution functions when data contains measurement errors. For example, consider the estimation of the prevalence of a disease which is determined by some underlying biomarker, measured with error, having value greater than some known constant. We present a data-driven method for estimating distribution functions in measurement error models, illustrate its superiority with respect to other methods both through theory and simulations, and apply it to a real example of estimating hypertension prevalence. In addition, we develop an estimator for the case where the error distribution is not known, but an external sample of measurement errors is available.Keywords
All Related Versions
This publication has 13 references indexed in Scilit:
- Deconvolution with unknown error distributionThe Annals of Statistics, 2009
- Blood pressure measurement: the problem and its solutionJournal of Human Hypertension, 2004
- Accurate test limits under nonnormal measurement errorMetrika, 1998
- Detecting powder mixture inhomogeneity for non-normal measurement errorsPowder Technology, 1997
- On the effect of estimating the error density in nonparametric deconvolutionJournal of Nonparametric Statistics, 1997
- Simulation extrapolation deconvolution of finite population cumulative distribution function estimatorsBiometrika, 1996
- Framingham Study Insights into Hypertensive Risk of Cardiovascular Disease.Hypertension Research, 1995
- A Fourier Approach to Nonparametric Deconvolution of a Density EstimateJournal of the Royal Statistical Society Series B: Statistical Methodology, 1993
- On the Optimal Rates of Convergence for Nonparametric Deconvolution ProblemsThe Annals of Statistics, 1991
- Inversion Formulae for the Distribution of RatiosThe Annals of Mathematical Statistics, 1948