Add to list

[Gabriel-p]

• pyABC https://github.com/icb-dcm/pyabc
• PTMCMCSampler https://github.com/jellis18/PTMCMCSampler
• ptmpi https://github.com/chris-n-self/ptmpi (blog entry https://chrisdoesscience.wordpress.com/2016/07/17/parallelised-parallel-tempering-with-mpi/)
• PyMultiNest https://github.com/JohannesBuchner/PyMultiNest
• dynesty https://github.com/joshspeagle/dynesty
• UltraNest https://github.com/JohannesBuchner/UltraNest
• XHMC (https://arxiv.org/abs/1601.00225) No repo yet (present in Stan)
• A-NICE-MC (https://arxiv.org/abs/1706.07561) https://github.com/ermongroup/a-nice-mc
• pymcmcstat (https://github.com/prmiles/pymcmcstat)
• ABCpy
• dyPolyChord (https://github.com/ejhigson/dyPolyChord)
• pocoMC (https://github.com/minaskar/pocomc)
• harmonic (https://github.com/astro-informatics/harmonic)
• PyVBMC (https://api.semanticscholar.org/CorpusID:257557304, https://github.com/acerbilab/pyvbmc, https://acerbilab.github.io/pyvbmc/index.html)
• approxbayescomp (https://github.com/Pat-Laub/approxbayescomp)

General list of python MCMC packages in Github: https://github.com/search?l=Python&o=asc&q=mcmc&s=updated&type=Repositories

[LoryPack]

Hello,

Could you please add ABCpy to the list? It is a Python library mainly for Approximate Bayesian Computation, implementing several algorithms. Features include: automatic summary statistics learning with linear regression or neural networks, easy parallelization with MPI or Spark and hierarchical model support.

Repository: https://github.com/eth-cscs/abcpy
Docs: https://abcpy.readthedocs.io/en/v0.5.7/
Article: https://arxiv.org/abs/1711.04694

[Gabriel-p]

Hi @LoryPack, the package has been added. Thank you for reporting it!

Cheers

[AdityaSavara]

This is a really nice list!
CheKiPEUQ is a pythonMCMC code for Parameter estimation for complex physical problems. The CheKiPEUQ software provides tools for finding physically realistic parameter estimates, graphs of the parameter estimate positions within parameter space, and plots of the final simulation results.

Features:

• Entirely implemented in python and can take external functions like a blackbox.
• Easily define custom log-likelihood, set variables as constants, or look at only a subset of responses.
• Parallel exploration by MPI
• Built-in Metropolis Hastings, Built-In Grid/Permutation Sampling, 'Built-in' Ensemble Slice Sampling by Zeus
• Bayesian Design of Experiments
• Customizable font sizes etc. for plots
• Many examples.

Repository: https://github.com/AdityaSavara/CheKiPEUQ
Article: https://chemistry-europe.onlinelibrary.wiley.com/doi/abs/10.1002/cctc.202000953

[Gabriel-p]

Done, thank you for the contribution @AdityaSavara!