Logistic regression for spatial Gibbs point processes

We propose a computationally efficient technique, based on logistic regression, for fittingGibbs point process models to spatial point pattern data. The score of the logistic regression is anunbiased estimating function and is closely related to the pseudolikelihood score. Implementationof our techn...

Full description

Bibliographic Details
Main Authors: Baddeley, Adrian, Coeurjolly, J., Rubak, E., Waagepetersen, R.
Format: Journal Article
Published: Oxford University Press 2014
Subjects:
Online Access:http://hdl.handle.net/20.500.11937/39331
_version_ 1848755562414604288
author Baddeley, Adrian
Coeurjolly, J.
Rubak, E.
Waagepetersen, R.
author_facet Baddeley, Adrian
Coeurjolly, J.
Rubak, E.
Waagepetersen, R.
author_sort Baddeley, Adrian
building Curtin Institutional Repository
collection Online Access
description We propose a computationally efficient technique, based on logistic regression, for fittingGibbs point process models to spatial point pattern data. The score of the logistic regression is anunbiased estimating function and is closely related to the pseudolikelihood score. Implementationof our technique does not require numerical quadrature, and thus avoids a source of bias inherentin other methods. For stationary processes, we prove that the parameter estimator is stronglyconsistent and asymptotically normal, and propose a variance estimator. We demonstrate theefficiency and practicability of the method on a real dataset and in a simulation study.
first_indexed 2025-11-14T08:58:17Z
format Journal Article
id curtin-20.500.11937-39331
institution Curtin University Malaysia
institution_category Local University
last_indexed 2025-11-14T08:58:17Z
publishDate 2014
publisher Oxford University Press
recordtype eprints
repository_type Digital Repository
spelling curtin-20.500.11937-393312017-09-13T14:24:48Z Logistic regression for spatial Gibbs point processes Baddeley, Adrian Coeurjolly, J. Rubak, E. Waagepetersen, R. Georgii–Nguyen–Zessin formula Exponential family model Pseudolikelihood Logistic regression Estimating function We propose a computationally efficient technique, based on logistic regression, for fittingGibbs point process models to spatial point pattern data. The score of the logistic regression is anunbiased estimating function and is closely related to the pseudolikelihood score. Implementationof our technique does not require numerical quadrature, and thus avoids a source of bias inherentin other methods. For stationary processes, we prove that the parameter estimator is stronglyconsistent and asymptotically normal, and propose a variance estimator. We demonstrate theefficiency and practicability of the method on a real dataset and in a simulation study. 2014 Journal Article http://hdl.handle.net/20.500.11937/39331 10.1093/biomet/ast060 Oxford University Press restricted
spellingShingle Georgii–Nguyen–Zessin formula
Exponential family model
Pseudolikelihood
Logistic regression
Estimating function
Baddeley, Adrian
Coeurjolly, J.
Rubak, E.
Waagepetersen, R.
Logistic regression for spatial Gibbs point processes
title Logistic regression for spatial Gibbs point processes
title_full Logistic regression for spatial Gibbs point processes
title_fullStr Logistic regression for spatial Gibbs point processes
title_full_unstemmed Logistic regression for spatial Gibbs point processes
title_short Logistic regression for spatial Gibbs point processes
title_sort logistic regression for spatial gibbs point processes
topic Georgii–Nguyen–Zessin formula
Exponential family model
Pseudolikelihood
Logistic regression
Estimating function
url http://hdl.handle.net/20.500.11937/39331