User’s Guide

The User’s Guide covers conceptual background and practical tutorials for bgms. It is organized into four parts: background concepts, model types, analysis workflows, and MCMC output.

Background

• Getting Started — Installation and first model fit
• Graphical Models — Nodes, edges, and conditional independence
• The Bayesian Approach — Priors, posteriors, and Bayes factors
• Prior Basics — How bgms uses priors and when to change defaults

Models

• Gaussian Graphical Model — Continuous data, partial associations, precision matrix
• Ordinal MRF — Binary and ordinal data, category thresholds, partial associations
• Mixed MRF — Datasets with both continuous and discrete variables

Analyses

• Edge Selection — Spike-and-slab priors for structure learning
• Group Comparison — Comparing networks across groups with bgmCompare()
• Edge Clustering — Stochastic block model prior for discovering edge clusters
• Missing Data — Handling missing values via listwise deletion or MCMC imputation

MCMC Output

• MCMC Output — What the sampler returns and how to access it
• MCMC Diagnostics — Convergence checks: ESS, R-hat, trace plots, divergences

Prerequisites

This guide assumes basic familiarity with R programming and the goals of network analysis. The Background chapters introduce the necessary statistical concepts.