Lifetime Models with Nonconstant Shape Parameters

Abstract

In its standard form, a lifetime regression model usually assumes that the time until an event occurs has a constant shape parameter and a scale parameter that is a function of covariates. In this paper we consider lifetime models with shape parameter dependent on a vector of covariates. Two special models are considered, the Weibull model and a mixture model incorporating long-term survivors, when we consider that the incidence probability is also dependent on covariates. Classical parameters estimation approach is considered on two real data sets.