Modeling Spatial Overdispersion with the Generalized Waring Process

TL;DR

This paper introduces the Generalized Waring process, a novel spatial point process model that overcomes previous limitations by satisfying key properties needed for modeling spatial overdispersion, unifying and extending existing models.

Contribution

The paper constructs a new spatial process based on the Generalized Waring Distribution, satisfying stationarity, ergodicity, and orderliness, and unifying negative binomial and Poisson processes as limits.

Findings

The Generalized Waring process satisfies key properties for spatial overdispersion modeling.

It encompasses negative binomial and Poisson processes as special cases.

The process offers a flexible and tractable alternative to existing models.

Abstract

Modeling spatial overdispersion requires point processes models with finite dimensional distributions that are overdisperse relative to the Poisson. Fitting such models usually heavily relies on the properties of stationarity, ergodicity, and orderliness. And, though processes based on negative binomial finite dimensional distributions have been widely considered, they typically fail to simultaneously satisfy the three required properties for fitting. Indeed, it has been conjectured by Diggle and Milne that no negative binomial model can satisfy all three properties. In light of this, we change perspective, and construct a new process based on a different overdisperse count model, the Generalized Waring Distribution. While comparably tractable and flexible to negative binomial processes, the Generalized Waring process is shown to possess all required properties, and additionally span the negative binomial and Poisson processes as limiting cases. In this sense, the GW process provides an approximate resolution to the conundrum highlighted by Diggle and Milne.