Author(s): Mauro Gasparini | Martina Gandini
Journal: Statistica
ISSN 0390-590X
Volume: 71;
Issue: 3;
Start page: 391;
Date: 2011;
Original page
ABSTRACT
We present an application of nonparametric estimation of survival in the presence of left-truncated and right-censored data. We confirm the well-known unstable behavior of the survival estimates when the risk set is small and there are too few early deaths. How ever, in our real scenario where only few death times are necessarily available, the proper nonparametric maximum likelihood estimator, and its usual modification, behave less badly than alternative methods proposed in the literature. The relative merits of the different estimators are discussed in a simulation study extending the settings of the case study to more general scenarios.
Journal: Statistica
ISSN 0390-590X
Volume: 71;
Issue: 3;
Start page: 391;
Date: 2011;
Original page
ABSTRACT
We present an application of nonparametric estimation of survival in the presence of left-truncated and right-censored data. We confirm the well-known unstable behavior of the survival estimates when the risk set is small and there are too few early deaths. How ever, in our real scenario where only few death times are necessarily available, the proper nonparametric maximum likelihood estimator, and its usual modification, behave less badly than alternative methods proposed in the literature. The relative merits of the different estimators are discussed in a simulation study extending the settings of the case study to more general scenarios.
