Conferencias
Vicente Garibay Cancho
Instituto de Ciências Matemáticas e de Computação (ICMC-USP), Brazil
In this paper we propose a new lifetime model for multivariate survival data in presence of surviving fractions and examine some of its properties. Its genesis is based on situations in which there are m types of unobservable competing causes, where each cause is related to a time of occurrence of an event of interest. Our model is a multivariate extension of the univariate survival cure rate model proposed by Tsodikov et al. (2003) and Rodrigues et al. (2009). We have discussed inference aspects for the proposed model following both, a classical and Bayesian approach. The inferential classical exploits the maximum likelihood tools. We also perform empirical study of the likelihood ratio test, to test the independence between the observed times of the event of interest. The Bayesian approach via Markov Chain Monte Carlo (MCMC) were considered. The methodology is illustrated on a real data set on customer churn data
