BNet is an exact Bayesian inference library for C++.
This library implements the junction tree algorithm for discrete variable Bayesian networks.
Examples are provided to illustrate its capabilities.
type 'make' to compile everything
type 'make doc' to compile the documentation (then you will have doc/BNET.pdf)
you will need a properly installed Linux system.
Requirements : gcc C++ compiler, latex (for the documentation)
It has been tested on Debian i(32 bits, P3) with gcc 3.3.5 and Fedora Core 4 (32 bits, Sempron) with gcc 4.0.1.
As usual, your mileage may vary. Please report bugs to firstname.lastname@example.org if it fails to compile.