ELFI - Engine for Likelihood-Free Inference

ELFI is a statistical software package for likelihood-free inference (LFI) such as Approximate Bayesian Computation (ABC). The term LFI refers to a family of inference methods that replace the use of the likelihood function with a data generating simulator function. Other names or related approaches to LFI include simulator-based inference, approximate Bayesian inference, indirect inference, etc.

ELFI features an easy to use syntax and supports parallelized inference out of the box.

MA2 model in ELFI

See the quickstart to get started.

ELFI is licensed under BSD3. The source is in GitHub.

Currently implemented LFI methods:

  • ABC rejection sampler

  • Sequential Monte Carlo ABC sampler

  • ABC-SMC sampler with adaptive distance

  • ABC-SMC sampler with adaptive threshold selection

  • Bayesian Optimization for Likelihood-Free Inference (BOLFI) framework

  • Robust Optimization Monte Carlo (ROMC) framework

  • Bayesian Optimization for Likelihood-Free Inference by Ratio Estimation (BOLFIRE)

  • Bayesian Synthetic Likelihood (BSL)

ELFI also has the following non LFI methods:

Additionally, ELFI integrates tools for visualization, model comparison, diagnostics and post-processing.

Getting started

Usage

Developer documentation

Citation

If you wish to cite ELFI, please use the paper in JMLR:

@article{JMLR:v19:17-374,
author  = {Jarno Lintusaari and Henri Vuollekoski and Antti Kangasr{\"a}{\"a}si{\"o} and Kusti Skyt{\'e}n and Marko J{\"a}rvenp{\"a}{\"a} and Pekka Marttinen and Michael U. Gutmann and Aki Vehtari and Jukka Corander and Samuel Kaski},
title   = {ELFI: Engine for Likelihood-Free Inference},
journal = {Journal of Machine Learning Research},
year    = {2018},
volume  = {19},
number  = {16},
pages   = {1-7},
url     = {http://jmlr.org/papers/v19/17-374.html}
}