/mcmc_diagnostics

Markov chain Monte Carlo general, and Hamiltonian Monte Carlo specific, diagnostics for Stan

Primary LanguageHTMLBSD 3-Clause "New" or "Revised" LicenseBSD-3-Clause

General Markov chain Monte Carlo, and specific Hamiltonian Monte Carlo, diagnostic code compatible with RStan and PyStan2.

The diagnostics can also be fully adapted to any other Hamiltonian Monte Carlo implementation by reimplementing the extract_expectands,
extract_hmc_diagnostics, and plot_inv_metric functions.

Recommendations for code optimization are welcomed and appreciated.

Acknowledgements {-}

I thank Sean Talts for Python code improvements. Raoul Kima originally suggested separating divergent transitions by numerical trajectory length.