Comparative Studies On Frailties In Survival Analysis

摘要

A mixed survival function is derived through proportional hazards model assuming Gamma or Inverse Gaussian frailty under Weibull baseline via Bayesian approach, and the maximum likelihood estimation (MLE) is proposed to estimate the parameters in the population. Through intensive simulations designed to generate right-censored data from the proposed survival function, we compare the MLE in this article with Cox estimates (in many different versions) for the covariate coefficient. Finally, the proposed survival model is applied to fit Leukemia data from Cox and Oakes (1992).