ExtensibleMCMC.jl

A modular implementation of the Markov chain Monte Carlo (MCMC) algorithm guided by the principle of ease of its extensibility.

The package comes with enough functionality to:

• run such MCMC algorithms as:
  • ✔/✘ Metropolis-Hastings algorithm (MH) with
    • ✔ Random walk proposals
      • ✔ drawn from Uniform on additive and log scale (+ adaptation of steps targeting acceptance rate)
      • ✔ drawn from Gaussian on additive and log scale (+ Haario-type adaptation)
    • ✘ Langevin proposals (MALA)
    • ✘ Hamiltionan dynamics
  • ✘ Pseudo-marginal updates
  • ✘ Non-reversible chains
    • ✘ Zig-zag algorithm
• print-out progress information to a console
• save intermediate results to csv files
• provide online diagnostic plots with ExtensibleMCMCPlots.jl extension

The package is structured in a way that allows the user to modify the internal flow of the algorithm in a simple way. As a result, the package admits extensions to compute-intensive problems that don't fit neatly into frameworks of other of Julia's MCMC packages.

Depending on your experience and intended use of this package you might consider starting at different places of this documentation.

• For a quick overview of ExtensibleMCMC.jl's' main functionality see Get started
• For a systematic introduction to all functionality introduced in this package and information about how to extend it see the User Manual
• For a didactic introduction to problems that can be solved using ExtensibleMCMC.jl see the Tutorials
• If you have a problem that you think can be addressed with this package, then check out the How-to guides to see if the answer is already there.