-
Notifications
You must be signed in to change notification settings - Fork 5
/
README
9 lines (5 loc) · 900 Bytes
/
README
1
2
3
4
5
6
7
8
9
This is an implementation of the probability monad. Inspired by the bher compiler of the MIT Scheme dialect Church (see http://projects.csail.mit.edu/church/wiki/Church) it uses Metropolis Hastings MCMC to sample from probabilistic programs.
The file src/probabilistic_clojure/monadic/api.clj contains the basic API for using the library. In the same directory, the file demos.clj demonstrates how generative models can be implemented. It relies on incanter (available from incanter.org) for graphical output and probability distributions.
For a quick start clone the repository, add the src directory (as well as incanter.jar) to your Java classpath and (use 'probabilistic-clojure.monadic.demos). Then run (test-mixture mixture-mem) to fit a Dirichlet mixture of Gaussian with three components to some test data.
LICENSE
This software is released under the LGPL. See COPYING.LESSER for details.