EasyABC

About EasyABC

The package EasyABC enables to perform efficient approximate bayesian computation (ABC) sampling schemes by launching a series of simulations of a computer code from the R platform, and to retrieve the simulation outputs in an appropriate format for post-processing treatments.

Package maintainer: Nicolas Dumoulin

Developers: Franck Jabot, Thierry Faure, Nicolas Dumoulin

Development of EasyABC has been supported by the Irstea project DynIndic and by the French National Research Agency (ANR) within the SYSCOMM project DISCO (ANR-09-SYSC-003).

Thanks to R-Forge for hosting the project.

Current features

• Launching of computer simulations to perform sequential or coupled-to-MCMC sampling schemes for Approximate Bayesian Computation.
• EasyABC currently implements four sequential algorithms:
  • Beaumont et al. Biometrika (2009)
  • Drovandi & Pettit Biometrics (2011)
  • Del Moral et al. Statistics and Computing (2012)
  • Lenormand et al. ArXiv (2012)
• EasyABC also implements three coupled-to-MCMC schemes:
  • Marjoram et al. Pnas (2003)
  • Wegmann et al. Genetics (2009)
  • A mix of Marjoram and Wegmann's algorithms (Marjoram with calibration step – Wegmann without PLS)
• EasyABC additionally enables to launch computer simulations in parallel on multiple cores of a multi-core computer.
• EasyABC also contains two R wrappers for binary codes to ease their launching from the R platform.