Goodness-of-fit of the distribution of time-to-first-occurrence in recurrent event models.

摘要: Imperfect repair models are a class of stochastic that deal with recurrent phenomena. This article focuses on the Block, Borges, and Savits (1985) age-dependent minimal model (the BBS model) in which system fails at time t undergoes one two types repair: probability p(t), perfect is performed, or 1-p(t), performed. The goodness-of-fit problem interest concerns initial distribution failure ages. In particular, testing null hypothesis hazard rate function time-to-first-event-occurrence, λ(·), equal to prespecified λ0(·). paper extends hazard-based smooth tests introduced Pena (1998a) case where data accrual from model. score derived by reformulating Neyman's idea terms functions. Omnibus as well directional developed simulation results presented illustrate sensitivities proposed for certain alternatives.