A recursive point process model for infectious diseases (Q2330536)

In the study of the evolution of various events, the Hawkes point processes proved to be useful tools, giving rise to appropriate mathematical models. Such a way, versions of Hawkes point models were used to model seismicity, perturbation in financial markets, crimes, terrorists strikes, invasive plants, etc. In this paper an extension of the Hawkes point processes model to a recursive model is given, that allows for previous disease status to inform a flexible component describing the time intervals between contagious events. The obtained model is said to be recursive, in the sense that the conditional intensity at any time depends on the productivity associated with previous points, and the productivity in turn depends on the conditional intensity at those points. In the paper, some previous results are presented, Hawkes point processes are described, and then the recursive model is given. This model is a point process model where the productivity for a subject infected at time \(t\) is inversely related to the conditional intensity at time \(t\). Basic properties of the obtained recursive model are studied, and differences between Hawkes model and recursive models are highlighted. After that, simulation and estimation of recursive point processes are given and an application to Rocky Mountain Spotted Fever cases in California is analyzed. The paper ends with some concluding remarks on the utility of such recursive point process model.