CMStatistics 2020: Start Registration
View Submission - CMStatistics
Title: Density estimation for censored and contaminated data Authors:  Elif Akca - KU Leuven (Belgium) [presenting]
Ingrid Van Keilegom - KU Leuven (Belgium)
Abstract: A vast literature exists on covariate measurement error correction in a survival context, i.e., a variety of methods are available when an uncontaminated survival outcome is regressed on error-prone covariates. However, it is also possible that the measurements for the survival outcome are not error-free. When those measurements are censored, both censoring and measurement error should be taken into account. A flexible approach for density estimation in the presence of censoring and measurement error is proposed when no auxiliary variable or validation data are available. A classical additive measurement error model with Gaussian noise and a right-censoring scheme is assumed. It is shown that the assumed model is identifiable under certain conditions on the support of the censored-contaminated survival outcome and a methodology using Laguerre polynomials is offered for density estimation. The numerical performance of the proposed methodology is investigated on both simulated and real data.