B0229
Title: Estimation of the density for censored and contaminated data
Authors: Ingrid Van Keilegom - KU Leuven (Belgium) [presenting]
Abstract: A vast literature exists on covariate measurement error correction in a survival context. In other words, plenty of methods are available when an uncontaminated survival outcome is regressed on error-prone covariates. However, it is possible that the measurements for the survival outcome are themselves prone to measurement error. When those measurements are also subject to censoring, both censoring and measurement error should be taken into account. A classical additive measurement error model is assumed with Gaussian noise unknown error variance and a random right censoring scheme. Under this setup, a flexible approach is proposed for the estimation of the error variance and the density of the survival time when no auxiliary variables or validation data are available. It is proven that the assumed model is identifiable and offers a flexible estimation strategy using Laguerre polynomials for the estimation of both quantities. The asymptotic normality of the proposed estimators is established, and the numerical performance of the methodology is investigated on both simulated and real data on gestational age.