Cosmology sampling with Population Monte Carlo (PMC)
CosmoPMC is a Monte-Carlo sampling method to explore the likelihood of various cosmological probes. The sampling engine is implemented with the package pmclib. It is called Population MonteCarlo (PMC), which is a novel technique to sample from the posterior (Cappé et al. 2008). PMC is an adaptive importance sampling method which iteratively improves the proposal to approximate the posterior. This code has been introduced, tested and applied to various cosmology data sets in Wraith, Kilbinger, Benabed et al. (2009). Results on the Bayesian evidence using PMC are discussed in Kilbinger, Wraith, Benabed et al. (2010).
The code consists of two packages:
- pmclib, the library containing the PMC sample engine. This package can be used also as a stand-alone sampler, and contains example code.
- CosmoPMC, the cosmology module. CosmoPMC needs pmclib to run.
Download the latest version (1.2).
- pmclib_v1.01.tar.gz (194 kB)
This patch is necessary for Cosmo_PMC v1.2. See the readme file for instructions how to install the patch.
- CosmoPMC_v1.2.tar.gz (1.7 MB)
- Weak lensing: CFHTLenS 2D cosmic shear data included, see Kilbinger et al. (2012)
- Weak lensing: COSEBIs implemented (Schneider, Eifler & Krause 2011)
- Weak lensing: Cosmology-dependent covariance matrix, see Eifler, Schneider & Hartlap (2009)
- BAO: ratio of sound horizon to low-z distance implemented, with BOSS data from Anderson (2012)
- 2D marginalised contour plots: new R script available, in addition to yorick one
COSMOLOGY MODULES AND DATA
|Cosmic shear||HST-COSMOS cosmic shear||Schrabback et al. (2009)|
|CFHTLenS 2D cosmic shear data||Kilbinger et al. (2012)|
|SNIa||Union sample||Kowalski et al. (2008)|
|BAO distance measures||SDSS||Eisenstein et al. (2005), Reid et al. (2010), Anderson (2012)|
|Galaxy clustering and HOD|
|CMB distance priors||WMAP7||Komatsu et al. 2009|
|Full CMB anisotropies (via CAMB and WMAP likelihood)||WMAP7||Komatsu et al. 2009|
See here for older versions.
Apart from a C-compiler, the libraries gsl (version 1.14 or higher) and fftw are required. These packages can be installed automatically with the pmclib configuration script. For graphics plots, the packages R or yorick are recommended.
The pmclib manual can be found here.
Kilbinger et al. (2011): Cosmo Population Monte Carlo - User's manual
Wraith, Kilbinger, Benabed et al. (2009): Comparison of PMC and MCMC, parameter estimation
Kilbinger, Wraith, Benabed et al. (2010): Bayesian evidence
Karim Benabed, Olivier Cappé, Jean Coupon, Jean-François Cardoso, Gersende Fort, Henry Joy McCracken, Simon Prunet, Christian P. Robert, Darren Wraith
Please feel free to send questions, feedback and bug reports to firstname.lastname@example.org .
If you want to be added to the cosmo_pmc mailing list, to get updates about new versions and bug-fixes, send me a mail at <ahref="mailto:email@example.com?subject=cosmopmc-mailing-list-add">firstname.lastname@example.org.