Sessõe Temática
Wagner Barreto-Souza
Universidade Federal de Minas Gerais (UFMG), Brazil
A new class of long-term survival models based on the first-activation scheme is proposed. For this new class it is assumed that the number of competing causes follows a generalized Poisson distribution, which is overdispersed. We consider a regression structure for the expected number of competing causes, so allowing to a direct modeling for the cure rate through covariates. Estimation by maximum likelihood and inference for the parameters of the new class are discussed. A small simulation study is conducted to examine the finite-sample performance of the proposed estimators. Applications to biological data sets illustrate the potentiality of the proposed models.
