Statistics Seminar: A vector of softcore processes for modelling interactions between species using capture-recapture data

Speaker: Alex Diana (University of Kent)

Title: A vector of softcore processes for modelling interactions between species using capture-recapture data

Microsoft Teams Meeting

Summary:

Capture-recapture (CR) data and corresponding models have been used extensively to perform inference on several quantities of interest of a wildlife population. When more than one traps is used, CR models can be extended to spatial CR models to also provide information on the spatial patterns of individual activity centres. Although the spatial patterns of the individual activity centres have usually been assumed to be independent, recent work relaxes this assumption by considering interactions between individuals of a single species. In this paper, we develop a model for CR data that can take into account interactions between and within multiple species. This is achieved by considering a vector of point processes from the general class of interaction point processes. These processes present a challenge from an inferential perspective because of the untractability of the normalising constant of the likelihood function, and hence standard MCMC procedures cannot be applied. Instead, we adopt an inference procedure based on the Monte-Carlo Metropolis Hastings algorithm of Liang and Jin (2013), which scales well also when the number of species is large. We apply the model two CR data-sets on leopards and tigers.