Demographic incidence rates and estimation of intensities with incomplete information (Q1067726)

A population process on the individual level can frequently be represented by a time-continuous Markov chain with a finite state space. With complete data, the estimation of transition intensities is straightforward. However, this paper treats the more difficult and important situation, where data are incomplete. Specifically, it is assumed that transitions within a subset of the states are observed in detail, while counts of transition out of this subset are only observed aggregated over the states. The derived methods rest on the assumption of nondifferential risk of transition out of the given subset. The results are given within the framework of counting process theory.