Software

BipartiteGMRF.jl

Julia package for fitting bipartite-graph Gaussian Markov random-field random-effects models. GitHub / DOI

BipartiteGMRF.jl provides tools for estimating random-effects models on matched bipartite data, such as worker-firm, student-school, or patient-provider networks. The package places a joint Gaussian prior on the two sets of latent node effects and estimates variance components and local dependence using sparse linear algebra. It was developed as the computational backbone for A Random-Effects Model Reveals Strong Positive Sorting in CEO Labor Markets, and is intended for reusable likelihood-based estimation, variance decomposition, and covariance extraction in sparse mobility networks.