Bayesian semiparametric inference for the accelerated failure time model using hierarchical mixture modeling with N-IG priors

We will pursue a Bayesian semiparametric approach for an Accelerated Failure Time regression model, usually consid- ered in survival analysis, when the error distribution is a mix- ture of parametric densities with a nonparametric mixing mea- sure. The Dirichlet process is a popular choice for the mix- ing measure, yielding a Dirichlet process mixture model for the error; the paper considers the same model, but here, as an alternative to the Dirichlet process, the mixing measure is equal to a normalized inverse-Gaussian prior, built from nor- malized inverse-gaussian finite dimensional distributions, as recently proposed in the literature. A comparison between the two models will be carried out. Markov chain Monte Carlo techniques will be used to estimate the predictive distribution of the survival time, along with the posterior distribution of the regression parameters. The efficiency of computational meth- ods will also be compared, using both real and simulated data.