Dependence modeling in Ghent University

Dependence in survival analysis is most frequently modeled by the frailty model or the copula model. In collaboration with the other research groups a number of new results in the research on frailty models and copula models have been published. In this talk we will briefly discuss the development of a shared gamma frailty model that handles data that are simultaneously clustered and interval-censored. In the model frailties can be integrated out analytically and the marginal likelihood can be maximized to obtain parameter estimates. For the remaining time focus will be on a claim often stated in the literature, namely the equivalence between Archimedean copula models and shared frailty models, e.g. between the Clayton-Oakes copula model and the shared gamma frailty model. In both models, there is indeed a well-known equivalence between the copula functions; the modeling of the marginal survival functions, however, is quite different. The latter fact leads to different joint survival functions.

Publication year:2010