This paper presents an extension to the Conservative PC algorithm which is able to detect violations of adjacency faithfulness under causal sufficiency and triangle faithfulness. Violations can be characterized by pseudo-independent relations and equivalent edges, both generating a pattern of conditional independencies that cannot be modeled faithfully. Both cases lead to uncertainty about specific parts of the skeleton of the causal graph. This is modeled by an f-pattern. We proved that our Very Conservative PC algorithm is able to correctly learn the f-pattern. Experiments based on simulations show that the rate of false edge removals is significantly reduced, at the expense of uncertainty on the skeleton and a higher sensitivity for accidental correlations.
Lemeire, J & Meganck, S 2010, Independence-based Causal Structure Learning in Absence of Faithfulness. in Proceedings of 5èmes Journées Francophones sur les Réseaux Bayésiens (JFRB), Nantes, 10-11 Mai 2010.
Lemeire, J., & Meganck, S. (2010). Independence-based Causal Structure Learning in Absence of Faithfulness. In Proceedings of 5èmes Journées Francophones sur les Réseaux Bayésiens (JFRB), Nantes, 10-11 Mai 2010
@inproceedings{d75f42830bcf447eb32aa98afd89015e,
title = "Independence-based Causal Structure Learning in Absence of Faithfulness",
abstract = "This paper presents an extension to the Conservative PC algorithm which is able to detect violations of adjacency faithfulness under causal sufficiency and triangle faithfulness. Violations can be characterized by pseudo-independent relations and equivalent edges, both generating a pattern of conditional independencies that cannot be modeled faithfully. Both cases lead to uncertainty about specific parts of the skeleton of the causal graph. This is modeled by an f-pattern. We proved that our Very Conservative PC algorithm is able to correctly learn the f-pattern. Experiments based on simulations show that the rate of false edge removals is significantly reduced, at the expense of uncertainty on the skeleton and a higher sensitivity for accidental correlations.",
keywords = "causality, bayesian networks, causal structure learning",
author = "Jan Lemeire and Stijn Meganck",
year = "2010",
language = "English",
booktitle = "Proceedings of 5{\`e}mes Journ{\'e}es Francophones sur les R{\'e}seaux Bay{\'e}siens (JFRB), Nantes, 10-11 Mai 2010",
}