Parametric Inference of Recurrent Alternating Event Data

Jun Yan
University of Iowa



Recurrent alternating event data arise from studies in which subjects experience two different types of events alternately, for example, recurrent hospitalization of patients with a chronic disease. There has been few literature addressing this type of data. This research proposes a parametric approach by specifying the transition intensities of alternating events. Maximum likelihood estimators will be developed and its performance for practical sample sizes will be investigated. The research provides a departure point and benchmark for semiparametric and nonparametric approaches.