Irrelevance and Independence Relations in Quasi-Bayesian Networks
Abstract
This paper analyzes irrelevance and independence relations in graphical models associated with convex sets of probability distributions (called Quasi-Bayesian networks). The basic question in Quasi-Bayesian networks is, How can irrelevance/independence relations in Quasi-Bayesian networks be detected, enforced and exploited? This paper addresses these questions through Walley's definitions of irrelevance and independence. Novel algorithms and results are presented for inferences with the so-called natural extensions using fractional linear programming, and the properties of the so-called type-1 extensions are clarified through a new generalization of d-separation.
Cite
Text
Cozman. "Irrelevance and Independence Relations in Quasi-Bayesian Networks." Conference on Uncertainty in Artificial Intelligence, 1998.Markdown
[Cozman. "Irrelevance and Independence Relations in Quasi-Bayesian Networks." Conference on Uncertainty in Artificial Intelligence, 1998.](https://mlanthology.org/uai/1998/cozman1998uai-irrelevance/)BibTeX
@inproceedings{cozman1998uai-irrelevance,
title = {{Irrelevance and Independence Relations in Quasi-Bayesian Networks}},
author = {Cozman, Fábio Gagliardi},
booktitle = {Conference on Uncertainty in Artificial Intelligence},
year = {1998},
pages = {89-96},
url = {https://mlanthology.org/uai/1998/cozman1998uai-irrelevance/}
}