Structural intervention distance for evaluating causal graphs

Research output: Contribution to journal › Journal article › Research › peer-review

Standard

Structural intervention distance for evaluating causal graphs. / Peters, Jonas Martin; Bühlmann, Peter.

In: Neural Computation, Vol. 27, No. 3, 03.2015, p. 771-799.

Research output: Contribution to journal › Journal article › Research › peer-review

Harvard

Peters, JM & Bühlmann, P 2015, 'Structural intervention distance for evaluating causal graphs', Neural Computation, vol. 27, no. 3, pp. 771-799. https://doi.org/10.1162/NECO_a_00708

APA

Peters, J. M., & Bühlmann, P. (2015). Structural intervention distance for evaluating causal graphs. Neural Computation, 27(3), 771-799. https://doi.org/10.1162/NECO_a_00708

Vancouver

Peters JM, Bühlmann P. Structural intervention distance for evaluating causal graphs. Neural Computation. 2015 Mar;27(3):771-799. https://doi.org/10.1162/NECO_a_00708

Author

Peters, Jonas Martin ; Bühlmann, Peter. / Structural intervention distance for evaluating causal graphs. In: Neural Computation. 2015 ; Vol. 27, No. 3. pp. 771-799.

Bibtex

@article{bc45af1899394c63bb4c74d06f3a77dc,
title = "Structural intervention distance for evaluating causal graphs",
abstract = "Causal inference relies on the structure of a graph, often a directed acyclic graph (DAG). Different graphs may result in different causal inference statements and different intervention distributions. To quantify such differences, we propose a (pre-)metric between DAGs, the structural intervention distance (SID). The SID is based on a graphical criterion only and quantifies the closeness between two DAGs in terms of their corresponding causal inference statements. It is therefore well suited for evaluating graphs that are used for computing interventions. Instead of DAGs, it is also possible to compare CPDAGs, completed partially DAGs that represent Markov equivalence classes. The SID differs significantly from the widely used structural Hamming distance and therefore constitutes a valuable additional measure. We discuss properties of this distance and provide a (reasonably) efficient implementation with software code available on the first author's home page.",
keywords = "Journal Article, Research Support, Non-U.S. Gov't",
author = "Peters, {Jonas Martin} and Peter B{\"u}hlmann",
year = "2015",
month = mar,
doi = "10.1162/NECO_a_00708",
language = "English",
volume = "27",
pages = "771--799",
journal = "Neural Computation",
issn = "0899-7667",
publisher = "M I T Press",
number = "3",

}

RIS

TY - JOUR

T1 - Structural intervention distance for evaluating causal graphs

AU - Peters, Jonas Martin

AU - Bühlmann, Peter

PY - 2015/3

Y1 - 2015/3

N2 - Causal inference relies on the structure of a graph, often a directed acyclic graph (DAG). Different graphs may result in different causal inference statements and different intervention distributions. To quantify such differences, we propose a (pre-)metric between DAGs, the structural intervention distance (SID). The SID is based on a graphical criterion only and quantifies the closeness between two DAGs in terms of their corresponding causal inference statements. It is therefore well suited for evaluating graphs that are used for computing interventions. Instead of DAGs, it is also possible to compare CPDAGs, completed partially DAGs that represent Markov equivalence classes. The SID differs significantly from the widely used structural Hamming distance and therefore constitutes a valuable additional measure. We discuss properties of this distance and provide a (reasonably) efficient implementation with software code available on the first author's home page.

AB - Causal inference relies on the structure of a graph, often a directed acyclic graph (DAG). Different graphs may result in different causal inference statements and different intervention distributions. To quantify such differences, we propose a (pre-)metric between DAGs, the structural intervention distance (SID). The SID is based on a graphical criterion only and quantifies the closeness between two DAGs in terms of their corresponding causal inference statements. It is therefore well suited for evaluating graphs that are used for computing interventions. Instead of DAGs, it is also possible to compare CPDAGs, completed partially DAGs that represent Markov equivalence classes. The SID differs significantly from the widely used structural Hamming distance and therefore constitutes a valuable additional measure. We discuss properties of this distance and provide a (reasonably) efficient implementation with software code available on the first author's home page.

KW - Journal Article

KW - Research Support, Non-U.S. Gov't

U2 - 10.1162/NECO_a_00708

DO - 10.1162/NECO_a_00708

M3 - Journal article

C2 - 25602767

VL - 27

SP - 771

EP - 799

JO - Neural Computation

JF - Neural Computation

SN - 0899-7667

IS - 3

ER -

ID: 165942038