A special case of the NPMLE of the joint distribution function with RMCR data.
Abstract
We consider the non-parametric estimation with right-censored competing risks
(RMCR) data and with masked failure cause under the random partition
masking model(see Yu et al. 2009). We show that the naive and common ways
of defining a NPMLE of the joint distribution function F lead to
inconsistency and we propose a consistent NPMLE of F in such a case.
