Central limit theorems for functionals of random hazard rates provide a synthetic picture of the global shape of a given hazard rate model. It has been shown that, for the most popular mixture hazard models with completely random mixing measures, the long-term trend and asymptotic variance of the cumulative hazard are not affected by conditioning on observations, so that the behavior of the posterior for large time horizons is systematically determined by the prior choice. In this paper we derive Berry-Ess\'een bounds for prior and posterior CLTs by using a recent technique based on Stein's method for normal approximation of Poisson functionals. The general result is specialized to various specific kernels and it shows that the data enters the second leading term of the prior bound by increasing it.

On the Asymptotic Behaviour of Random Cumulative Hazards

DE BLASI, Pierpaolo;PRUENSTER, Igor
2010

Abstract

Central limit theorems for functionals of random hazard rates provide a synthetic picture of the global shape of a given hazard rate model. It has been shown that, for the most popular mixture hazard models with completely random mixing measures, the long-term trend and asymptotic variance of the cumulative hazard are not affected by conditioning on observations, so that the behavior of the posterior for large time horizons is systematically determined by the prior choice. In this paper we derive Berry-Ess\'een bounds for prior and posterior CLTs by using a recent technique based on Stein's method for normal approximation of Poisson functionals. The general result is specialized to various specific kernels and it shows that the data enters the second leading term of the prior bound by increasing it.
Joint Statistical Meetings 2010
Vancouver
July 31- August 5
JSM 2010 Proceedings
American Statistical Association
1063
1074
9780979174797
http://www.amstat.org/meetings/jsm/2010/
Bayesian nonparametrics; asymptotics; central limit theorem; mixture hazard rate; survival analysis
P. DE BLASI; G. PECCATI; I. PRUENSTER
File in questo prodotto:
Non ci sono file associati a questo prodotto.

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/2318/80395
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus ND
  • ???jsp.display-item.citation.isi??? ND
social impact