Classification of Points in Superpositions of Strauss and Poisson Processes
Artikel i vetenskaplig tidskrift, 2015

Consider a realisation of a point process which is formed as a superposition of a regular point process, here a Strauss process, and some Poisson noise. The aim of the current work is to decide which of the two processes each point belongs to. We construct an MCMC algorithm which estimates the parameters of the superposition model and obtains posterior probabilities for each point of being a Strauss point. The algorithm is evaluated in a simulation study. Finally, it is applied to our motivating data set containing the locations of air bubbles, some of which are noise, in an Antarctic ice core.

Markov chain Monte Carlo

Spatial point process

Noise removal

Parameter estimation

Noise detection

Bayesian inference

Författare

Claudia Redenbach

Technische Universität Kaiserslautern

Martina Sormani

Fraunhofer-Institut fur Techno- und Wirtschaftsmathematk

Technische Universität Kaiserslautern

Aila Särkkä

Chalmers, Matematiska vetenskaper, Matematisk statistik

Göteborgs universitet

Spatial Statistics

2211-6753 (ISSN)

Vol. 12 81-95

Drivkrafter

Hållbar utveckling

Ämneskategorier

Matematik

Styrkeområden

Materialvetenskap

DOI

10.1016/j.spasta.2015.03.003

Mer information

Senast uppdaterat

2020-08-06