Um estudo dos modelos de sobrevivência de longa duração LIGcr e GEPGWcr

Abstract

In this work we study two long-term survival models denomined Logaritmic Inverse Gaussian cure rate (LIGcr) model and Geometric Exponentiated Power Generalized Weibull cure rate (GEPGWcr) model. Both models take into account the existence of unobserved heterogeneity. For this, both models consider a random variable that describes the frailty been discret and starting on zero. For both models studied, the parameters are estimated by the maximum likehood method. In a simulations study, we investigated the averages of the estimates, the biases, the mean squared errors and coverage probabilities. Finally, we show the applicability of the models to real data sets.