UniDAGs

Overview

UniDAGs: Uniformly distributed samples of directed acyclic graphs for random Bayesian Network generation.

• Source Code: https://github.com/hayesall/UniDAGs

Motivation

When evaluating Bayesian network algorithms, it is convenient if you can check how they perform for known structures.

This implements algorithms for generating trees, multi-DAGs, and polytrees.

Citing

Here’s a BibTeX citation for the author’s paper, “Random Generation of Bayesian Networks.” Also see: https://doi.org/10.1007/3-540-36127-8_35

@inproceedings{ide2002randomgeneration,
  author="Ide, Jaime S. and Cozman, Fabio G.",
  editor="Bittencourt, Guilherme and Ramalho, Geber L.",
  title="Random Generation of Bayesian Networks",
  booktitle="Advances in Artificial Intelligence",
  year="2002",
  publisher="Springer Berlin Heidelberg",
  address="Berlin, Heidelberg",
  pages="366--376",
  isbn="978-3-540-36127-5"
}