This is the development version of the pyhm software package.

pyhm is a Python package for defining hierarchical Bayesian models, with routines to optimise and sample from the posterior distribution. Currently, pyhm only supports the scipy optimisation routines and standard Metropolis-Hastings sampling; however, pyhm was built with extensibility in mind, and adding further capabilities will be trivial due to the modular structure of the software. In particular, implementing a nested sampling option is a high priority. This software drew a lot of inspiration from the pymc package.

For more information and a tarball of the latest release, see the documentation at:

http://www.physics.ox.ac.uk/users/evanst/code/pyhm/index.html