Prior specification of neighbourhood and interaction structure in binary Markov random fields

TL;DR

This paper introduces a novel prior for the structure of binary Markov random fields, including higher order interactions, and develops an RJMCMC algorithm for posterior sampling, demonstrated on simulations and real data.

Contribution

It proposes a new prior for MRF structure and interactions, incorporating higher order terms and parameter constraints, with an efficient sampling algorithm.

Findings

Effective in modeling complex dependencies in MRFs

Reduces free parameters by allowing parameter equality

Demonstrates applicability on simulated and real data

Abstract

In this paper we propose a prior distribution for the clique set and dependence structure of binary Markov random fields (MRFs). In the formulation we allow both pairwise and higher order interactions. We construct the prior by first defining a prior for the neighbourhood system of the MRF, and conditioned on this we define a prior for the appearance of higher order interactions. Finally, for the parameter values we adopt a prior that allows for parameter values to equal, and in this way we reduce the effective number of free parameters. To sample from a resulting posterior distribution conditioned on an observed scene we construct a reversible jump Markov chain Monte Carlo (RJMCMC) algorithm. We circumvent evaluations of the intractable normalising constant of the MRF when running this algorithm by adopting a previously defined approximate auxiliary variable algorithm. We demonstrate the usefulness of our prior in two simulation examples and one real data example.