Premium
Estimation of location and scale functionals in nonparametric regression under copula dependent censoring
Author(s) -
Sujica Aleksandar,
van Keilegom Ingrid
Publication year - 2015
Publication title -
canadian journal of statistics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.804
H-Index - 51
eISSN - 1708-945X
pISSN - 0319-5724
DOI - 10.1002/cjs.11250
Subject(s) - estimator , copula (linguistics) , mathematics , covariate , censoring (clinical trials) , nonparametric statistics , statistics , random variable , multivariate random variable , econometrics
Let ( X , Y ) be a random vector, where Y denotes the variable of interest possibly subject to random right censoring, and X is a covariate. The variable Y is a (possible monotone transformation of a) survival time. The censoring time C and the survival time Y are allowed to be dependent, and the dependence is described via a known copula (this also includes the independent case). Under this setting we propose estimators of certain location and scale functionals of Y given X . We derive their asymptotic properties, uniformly over the support of X . In particular we derive an asymptotic representation and the uniform convergence rates for these estimators and their derivatives. We also prove asymptotic results for an estimator of the conditional distribution (the so‐called conditional copula‐graphic estimator), which generalizes previous results obtained by Braekers & Veraverbeke (2005). We also illustrate the results via simulations and the analysis of data on bone marrow transplantation. The Canadian Journal of Statistics 43: 306–335; 2015 © 2015 Statistical Society of Canada