CaltechAUTHORS
  A Caltech Library Service

Causal Reasoning with Ancestral Graphs

Zhang, Jiji (2008) Causal Reasoning with Ancestral Graphs. Journal of Machine Learning Research, 9 . pp. 1437-1474. ISSN 1533-7928. http://resolver.caltech.edu/CaltechAUTHORS:ZHAjmlr08

[img]
Preview
PDF - Published Version
See Usage Policy.

285Kb

Use this Persistent URL to link to this item: http://resolver.caltech.edu/CaltechAUTHORS:ZHAjmlr08

Abstract

Causal reasoning is primarily concerned with what would happen to a system under external interventions. In particular, we are often interested in predicting the probability distribution of some random variables that would result if some other variables were forced to take certain values. One prominent approach to tackling this problem is based on causal Bayesian networks, using directed acyclic graphs as causal diagrams to relate post-intervention probabilities to pre-intervention probabilities that are estimable from observational data. However, such causal diagrams are seldom fully testable given observational data. In consequence, many causal discovery algorithms based on data-mining can only output an equivalence class of causal diagrams (rather than a single one). This paper is concerned with causal reasoning given an equivalence class of causal diagrams, represented by a (partial) ancestral graph. We present two main results. The first result extends Pearl (1995)'s celebrated do-calculus to the context of ancestral graphs. In the second result, we focus on a key component of Pearl's calculus---the property of invariance under interventions, and give stronger graphical conditions for this property than those implied by the first result. The second result also improves the earlier, similar results due to Spirtes et al. (1993).


Item Type:Article
Related URLs:
URLURL TypeDescription
http://jmlr.csail.mit.edu/papers/v9/zhang08a.htmlPublisherUNSPECIFIED
Additional Information:© 2008 Jiji Zhang. Submitted 6/07; Revised 2/08; Published 7/08. Editor: Gregory F. Cooper. I am grateful to Clark Glymour, Thomas Richardson, and Peter Spirtes for their helpful comments on the part of my dissertation this paper is based on. Thanks also to three anonymous referees for helping improve the paper significantly. One of them, especially, made extremely detailed and helpful suggestions.
Subject Keywords:ancestral graphs, causal Bayesian network, do-calculus, intervention
Record Number:CaltechAUTHORS:ZHAjmlr08
Persistent URL:http://resolver.caltech.edu/CaltechAUTHORS:ZHAjmlr08
Usage Policy:No commercial reproduction, distribution, display or performance rights in this work are provided.
ID Code:11657
Collection:CaltechAUTHORS
Deposited By: Archive Administrator
Deposited On:17 Sep 2008 06:15
Last Modified:26 Dec 2012 10:17

Repository Staff Only: item control page