Jeffreys' Priors for Survival Models with Censored Data

Jeffreys priors for survival models withcensored dataF. De SantisUniversitµa di Roma \La Sapienza"J. MorteraUniversitµa Roma TreA. NardiUniversitµa di TeramoAbstractWhen prior information on model parameters is weak or lacking,Bayesian statistical analyses are typically performed with so-called\default" priors. We consider the problem of constructing defaultpriors for the parameters of survival models in the presence of cen-soring, using Je®reys' rule. We compare these Je®reys priors to the\uncensored" Je®reys' priors, obtained without considering censoredobservations, for the parameters of the exponential and log-normalmodels. The comparison is based on the frequentist coverage of theposterior Bayes intervals obtained from these prior distributions.