An explicit link between graphical models and Gaussian Markov random fields on metric graphs
(2026) In Stochastic Processes and their Applications 196.- Abstract
We derive an explicit link between Gaussian Markov random fields on metric graphs and graphical models, and in particular show that a Markov random field restricted to the vertices of the graph is, under mild regularity conditions, a Gaussian graphical model. This graphical model has a distribution which is faithful to its pairwise independence graph, that coincides with the neighbor structure of the metric graph. This is used to show that there are no Gaussian random fields on general metric graphs which are both Markov and isotropic in some suitably regular metric on the graph, such as the geodesic or resistance metrics.
Please use this url to cite or link to this publication:
https://lup.lub.lu.se/record/eca7ced3-514f-49d5-98a8-4540c01f5fd8
- author
- Bolin, David LU ; Simas, Alexandre B. LU and Wallin, Jonas LU
- organization
- publishing date
- 2026-06
- type
- Contribution to journal
- publication status
- published
- subject
- keywords
- Gaussian process, GMRF, Graphical models, Markov, Metric graph
- in
- Stochastic Processes and their Applications
- volume
- 196
- article number
- 104925
- publisher
- Elsevier
- external identifiers
-
- scopus:105032385244
- ISSN
- 0304-4149
- DOI
- 10.1016/j.spa.2026.104925
- language
- English
- LU publication?
- yes
- additional info
- Publisher Copyright: © 2026 Elsevier B.V.
- id
- eca7ced3-514f-49d5-98a8-4540c01f5fd8
- date added to LUP
- 2026-04-27 15:48:39
- date last changed
- 2026-04-27 15:49:50
@article{eca7ced3-514f-49d5-98a8-4540c01f5fd8,
abstract = {{<p>We derive an explicit link between Gaussian Markov random fields on metric graphs and graphical models, and in particular show that a Markov random field restricted to the vertices of the graph is, under mild regularity conditions, a Gaussian graphical model. This graphical model has a distribution which is faithful to its pairwise independence graph, that coincides with the neighbor structure of the metric graph. This is used to show that there are no Gaussian random fields on general metric graphs which are both Markov and isotropic in some suitably regular metric on the graph, such as the geodesic or resistance metrics.</p>}},
author = {{Bolin, David and Simas, Alexandre B. and Wallin, Jonas}},
issn = {{0304-4149}},
keywords = {{Gaussian process; GMRF; Graphical models; Markov; Metric graph}},
language = {{eng}},
publisher = {{Elsevier}},
series = {{Stochastic Processes and their Applications}},
title = {{An explicit link between graphical models and Gaussian Markov random fields on metric graphs}},
url = {{http://dx.doi.org/10.1016/j.spa.2026.104925}},
doi = {{10.1016/j.spa.2026.104925}},
volume = {{196}},
year = {{2026}},
}