Frailty Model and its Application to Seizure Data

Publisher Summary This chapter presents a review of both a semi-parametric and a parametric approach to estimating the risk coefficients and the frailty parameters for the gamma frailty, the positive stable frailty and the log-normal frailty models. A regression approach to failure-time data analysis could be either fully parametric or semi-parametric. A parametric approach involves extensions of existing parametric failure-time models, such as exponential, Weibull, and log-normal models by means of re-parameterizations to include covariates. On the other hand, a semiparametric approach is distribution free and involves less stringent assumptions on the underlying failure-time distribution. Because in general, recurrent events are more complicated than parallel data, the chapter reviews frailty models as well as statistical inference for such events. The chapter discusses an application of the frailty model in medical science, which is also applicable to reliability, actuarial science, biological science, and other fields where the multivariate times are observed.