User guide

A tour of ergmx, from loading a network to comparing models. Every page runs its code when the documentation is built, so the output is what ergmx gives today.

• Installation
  • Install uv
  • Install ergmx from PyPI
  • From GitHub
  • The Rust toolchain
  • Development
  • How the Rust is bundled
  • Troubleshooting
• Quick start
  • A complete ERGM analysis
  • A complete multilevel ERGM analysis
• Under the hood
  • Why not just Python?
  • Why Rust?
  • What runs where
  • The tools
  • From source code to import ergmx
• Networks
  • The bundled networks
  • Vertex attributes
  • Graph attributes
  • What is not supported
• Formulas
  • Operators
  • Statistic names
  • Errors
• Fitting models
  • Dyad-independent models: the exact MLE
  • Dyad-dependent models: the Monte Carlo MLE
  • Seeds and threads
  • Settings
  • Other estimates and starting values
  • Estimating the decay: curved models
  • Fixed coefficients
  • Constraints and missing ties
  • Saving fits
• Constraints
  • Bounded degrees: fixed-choice designs
  • Fixing blocks of dyads
  • Preserving degrees
  • What changes with dyad-dependent constraints
• Missing ties
  • Marking missing dyads
  • Fitting
  • The method
  • Goodness of fit
  • Several networks, and waves
• Multilevel networks
  • Each level, and the ties between them
  • How the levels depend on each other
  • Goodness of fit by level
  • Estimating the decays
  • Directed multilevel networks
  • Densities by kind of tie
  • Closure within levels
  • Given the affiliations
  • Fixed coefficients
• Bipartite networks
  • Declaring the modes
  • Fitting
  • Goodness of fit
• Samples of networks
  • Networks() and N()
  • Effects that depend on the network
  • Curved terms
  • Checking and using the fit
  • Goodness of fit network by network
• Networks over time
  • Formation and persistence
  • Missing dyads
  • Simulating the process
  • A process from one network: the EGMME
• Convergence and MCMC diagnostics
  • Degenerate models
• Goodness of fit
  • Options
  • By level, and network by network
• Model comparison
  • How the log-likelihood is estimated
• Interpreting and reporting results
  • Odds ratios and confidence intervals
  • Tie probabilities
  • Average marginal effects
  • Tables for papers
• Simulation
  • From a fit
  • MCMC settings
  • Over time