Characterization and Greedy Learning of Interventional Markov Equivalence Classes of Directed Acyclic Graphs (Abstract)
Alain Hauser, Peter Buhlmann
Directed acyclic graphs (DAGs) are commonly used to model causal relationships between random variables. Markov equivalence of DAGs indicates to which extent these causal influences are identifiable from the observational density of the random variables. In this paper, we extend the notion of Markov equivalence of DAGs to the case of interventional distributions arising from multiple intervention experiments which are crucial for improved causal inference, and we generalize the greedy equivalence search (GES) algorithm in order to process observational and interventional data (or data from different interventions) simultaneously.
PDF Link: /papers/11/p851-hauser.pdf
AUTHOR = "Alain Hauser
and Peter Buhlmann",
TITLE = "Characterization and Greedy Learning of Interventional Markov Equivalence Classes of Directed Acyclic Graphs (Abstract)",
BOOKTITLE = "Proceedings of the Twenty-Seventh Conference Annual Conference on Uncertainty in Artificial Intelligence (UAI-11)",
PUBLISHER = "AUAI Press",
ADDRESS = "Corvallis, Oregon",
YEAR = "2011",
PAGES = "851--851"