Ageing and Risk Aspects in Predictive Inference Based on Proportional Hazard Models

Proportional Hazard Models arise from a straightforward generalization of the simple case of conditionally i.i.d., exponentially distributed random variables and, in a sense, can be considered as the idealized models in the statistical analysis of failure and survival data for lifetimes. For these reasons, they have been extensively studied in the literature. Despite of the richness of related contributions, there are still special aspects of these models that are worthwhile focusing. In this discussion paper we aim to present some contributions, in the frame of a Bayesian approach and by using some very basic notions of stochastic ordering.