VCDP

Copyright Pekka Parviainen 2015

VCDP is a software for learning Bayesian networks with bounded vertex cover number using using dynamic programming. For any fixed vertex cover number k, it runs in polynomial time with respect to the number of variables n [1]. VCDP is implemented in Matlab.


Installation
============

Dependencies: VCDP uses functions from the BDAGL toolbox by Daniel Eaton and Kevin Murphy. For convenience, the required files (directory OptimalMAP) are included in this package.


Usage
=====

The main function is find_bounded_vertex_cover_dag.m. The input data is an n*2^n matrix of local scores where n is the number of nodes. For more details, see the comments in the code.


Licence
=======

VCDP is published under GNU General Public Licence v3.0.


References
==========

[1] Janne H. Korhonen and Pekka Parviainen. Tractable Bayesian Network Structure Learning with Bounded Vertex Cover Number. NIPS 2015.
