Constructing Separators and Adjustment Sets in Ancestral Graphs

Benito van der Zander; Maciej Liskiewicz; Johannes Textor

Constructing Separators and Adjustment Sets in Ancestral Graphs

Benito van der Zander, Maciej Liskiewicz, Johannes Textor

Institute of Theoretical Computer Science

Abstract

Ancestral graphs (AGs) are graphical causal models that can represent uncertainty about the presence of latent confounders, and can be inferred from data. Here, we present an algorithmic framework for efficiently testing, constructing, and enumerating m-separators in AGs. Moreover, we present a new constructive criterion for covariate adjustment in directed acyclic graphs (DAGs) and maximal ancestral graphs (MAGs) that characterizes adjustment sets as m-separators in a subgraph. Jointly, these results allow to find all adjustment sets that can identify a desired causal effect with multivariate exposures and outcomes in the presence of latent confounding. Our results generalize and improve upon several existing solutions for special cases of these problems.

Original language	English
Title of host publication	Proceedings of the UAI 2014 Conference on Causal Inference: Learning and Prediction - Volume 1274
Number of pages	14
Volume	1274
Place of Publication	Aachen, Germany, Germany
Publisher	CEUR-WS.org
Publication date	27.07.2014
Pages	11-24
Publication status	Published - 27.07.2014
Event	UAI'14 - Quebec, Canada Duration: 23.07.2014 → 27.07.2014 http://auai.org/uai2014/

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

Access to Document

Cite this

@inproceedings{1fc1cddfa44240f788b70dbb022c6edc,

title = "Constructing Separators and Adjustment Sets in Ancestral Graphs",

abstract = "Ancestral graphs (AGs) are graphical causal models that can represent uncertainty about the presence of latent confounders, and can be inferred from data. Here, we present an algorithmic framework for efficiently testing, constructing, and enumerating m-separators in AGs. Moreover, we present a new constructive criterion for covariate adjustment in directed acyclic graphs (DAGs) and maximal ancestral graphs (MAGs) that characterizes adjustment sets as m-separators in a subgraph. Jointly, these results allow to find all adjustment sets that can identify a desired causal effect with multivariate exposures and outcomes in the presence of latent confounding. Our results generalize and improve upon several existing solutions for special cases of these problems.",

author = "{van der Zander}, Benito and Maciej Liskiewicz and Johannes Textor",

year = "2014",

month = jul,

day = "27",

language = "English",

volume = "1274",

series = "CI'14",

publisher = "CEUR-WS.org",

pages = "11--24",

booktitle = "Proceedings of the UAI 2014 Conference on Causal Inference: Learning and Prediction - Volume 1274",

note = "UAI'14 ; Conference date: 23-07-2014 Through 27-07-2014",

url = "http://auai.org/uai2014/",

}

van der Zander, B , Liskiewicz, M & Textor, J 2014, Constructing Separators and Adjustment Sets in Ancestral Graphs. in Proceedings of the UAI 2014 Conference on Causal Inference: Learning and Prediction - Volume 1274. vol. 1274, CI'14, CEUR-WS.org, Aachen, Germany, Germany, pp. 11-24, UAI'14, Quebec, Canada, 23.07.14. <http://dl.acm.org/citation.cfm?id=3020325.3020328>

Constructing Separators and Adjustment Sets in Ancestral Graphs. / van der Zander, Benito ; Liskiewicz, Maciej; Textor, Johannes.
Proceedings of the UAI 2014 Conference on Causal Inference: Learning and Prediction - Volume 1274. Vol. 1274 Aachen, Germany, Germany: CEUR-WS.org, 2014. p. 11-24 (CI'14).

Research output: Chapters in Books/Reports/Conference Proceedings › Conference contribution › peer-review

TY - GEN

T1 - Constructing Separators and Adjustment Sets in Ancestral Graphs

AU - van der Zander, Benito

AU - Liskiewicz, Maciej

AU - Textor, Johannes

PY - 2014/7/27

Y1 - 2014/7/27

N2 - Ancestral graphs (AGs) are graphical causal models that can represent uncertainty about the presence of latent confounders, and can be inferred from data. Here, we present an algorithmic framework for efficiently testing, constructing, and enumerating m-separators in AGs. Moreover, we present a new constructive criterion for covariate adjustment in directed acyclic graphs (DAGs) and maximal ancestral graphs (MAGs) that characterizes adjustment sets as m-separators in a subgraph. Jointly, these results allow to find all adjustment sets that can identify a desired causal effect with multivariate exposures and outcomes in the presence of latent confounding. Our results generalize and improve upon several existing solutions for special cases of these problems.

AB - Ancestral graphs (AGs) are graphical causal models that can represent uncertainty about the presence of latent confounders, and can be inferred from data. Here, we present an algorithmic framework for efficiently testing, constructing, and enumerating m-separators in AGs. Moreover, we present a new constructive criterion for covariate adjustment in directed acyclic graphs (DAGs) and maximal ancestral graphs (MAGs) that characterizes adjustment sets as m-separators in a subgraph. Jointly, these results allow to find all adjustment sets that can identify a desired causal effect with multivariate exposures and outcomes in the presence of latent confounding. Our results generalize and improve upon several existing solutions for special cases of these problems.

M3 - Conference contribution

VL - 1274

T3 - CI'14

SP - 11

EP - 24

BT - Proceedings of the UAI 2014 Conference on Causal Inference: Learning and Prediction - Volume 1274

PB - CEUR-WS.org

CY - Aachen, Germany, Germany

T2 - UAI'14

Y2 - 23 July 2014 through 27 July 2014

ER -

Constructing Separators and Adjustment Sets in Ancestral Graphs

Abstract

UN SDGs

Access to Document

Fingerprint

Cite this