Sparse Interaction Neighborhood Selection for Markov Random Fields via Reversible Jump and Pseudoposteriors

2022/04/12 by Freguglia, Victor, Garcia, Nancy Lopes
#Computation (stat.CO) #FOS: Computer and information sciences #Machine Learning (stat.ML)

paper · doi:10.48550/arxiv.2204.05933

Abstract

We consider the problem of estimating the interacting neighborhood of a Markov Random Field model with finite support and homogeneous pairwise interactions based on relative positions of a two-dimensional lattice. Using a Bayesian framework, we propose a Reversible Jump Monte Carlo Markov Chain algorithm that jumps across subsets of a maximal range neighborhood, allowing us to perform model selection based on a marginal pseudoposterior distribution of models. To show the strength of our proposed methodology we perform a simulation study and apply it to a real dataset from a discrete texture image analysis.

Related

• Prior specification of neighbourhood and interaction structure in binary Markov random fields
• An Oracle Approach for Interaction Neighborhood Estimation in Random Fields
• Fully Bayesian binary Markov random field models: Prior specification and posterior simulation
• Data-driven neighborhood selection of a Gaussian field
• Approximations and bounds for binary Markov random fields