A note on the EM algorithm for estimation in the destructive negative binomial cure rate model

Abstract

In this note we present a modification in the EM algorithm for the destructive negative binomial cure rate model. This alteration enables us to obtain the estimates of the whole parameter vector from the complete loglikelihood function, avoiding the corresponding observed log-likelihood function, which is more involved. To achieve this goal, we resort to the mixture representation of the negative binomial distribution in terms of the Poisson and gamma distributions