MCMCDiagnostics.jl

Markov Chain Monte Carlo convergence diagnostics in Julia
Author tpapp
Popularity
18 Stars
Updated Last
1 Year Ago
Started In
June 2017

MCMCDiagnostics.jl

Markov Chain Monte Carlo convergence diagnostics in Julia.

lifecycle build codecov.io

Overview

This package contains two very useful diagnostics for Markov Chain Monte Carlo:

  1. potential_scale_reduction(chains...), which estimates the potential scale reduction factor, also known as Rhat, for multiple scalar chains,

  2. effective_sample_size(chain), which calculates the effective sample size for scalar chains.

These are intended as building blocks, to be used by other libraries, and were organized into a separate library for testing and DRY.

Installation

The package is registered. You can install it with

Pkg.add("MCMCDiagnostics")

Related

You may find my other packages for MCMC interesting. See the documentation of DynamicHMC.jl for details.

Bibliography

Gelman, A., & Rubin, D. B. (1992). Inference from iterative simulation using multiple sequences. Statistical science, 457-472.

Gelman, A., Carlin, J. B., Stern, H. S., & Rubin, D. B. (2013). Bayesian data analysis (3rd edition). Chapman & Hall/CRC.

Stan Development Team. (2017). Stan Modeling Language Users Guide and Reference Manual, Version 2.15.0. http://mc-stan.org