Modeling confounding by half-sibling regression

Bernhard Schölkopf, David W. Hogg, Dun Wang, Daniel Foreman-Mackey, Dominik Janzing, Carl Johann Simon-Gabriel, Jonas Peters

    Research output: Contribution to journalArticle

    Abstract

    We describe a method for removing the effect of confounders to reconstruct a latent quantity of interest. The method, referred to as "half-sibling regression," is inspired by recent work in causal inference using additive noise models. We provide a theoretical justification, discussing both independent and identically distributed as well as time series data, respectively, and illustrate the potential of the method in a challenging astronomy application.

    Original languageEnglish (US)
    Pages (from-to)7391-7398
    Number of pages8
    JournalProceedings of the National Academy of Sciences of the United States of America
    Volume113
    Issue number27
    DOIs
    StatePublished - Jul 5 2016

    Fingerprint

    Additive noise
    Astronomy
    Time series

    Keywords

    • Astronomy
    • Causal inference
    • Exoplanet detection
    • Machine learning
    • Systematic error modeling

    ASJC Scopus subject areas

    • General

    Cite this

    Schölkopf, B., Hogg, D. W., Wang, D., Foreman-Mackey, D., Janzing, D., Simon-Gabriel, C. J., & Peters, J. (2016). Modeling confounding by half-sibling regression. Proceedings of the National Academy of Sciences of the United States of America, 113(27), 7391-7398. https://doi.org/10.1073/pnas.1511656113

    Modeling confounding by half-sibling regression. / Schölkopf, Bernhard; Hogg, David W.; Wang, Dun; Foreman-Mackey, Daniel; Janzing, Dominik; Simon-Gabriel, Carl Johann; Peters, Jonas.

    In: Proceedings of the National Academy of Sciences of the United States of America, Vol. 113, No. 27, 05.07.2016, p. 7391-7398.

    Research output: Contribution to journalArticle

    Schölkopf, B, Hogg, DW, Wang, D, Foreman-Mackey, D, Janzing, D, Simon-Gabriel, CJ & Peters, J 2016, 'Modeling confounding by half-sibling regression', Proceedings of the National Academy of Sciences of the United States of America, vol. 113, no. 27, pp. 7391-7398. https://doi.org/10.1073/pnas.1511656113
    Schölkopf B, Hogg DW, Wang D, Foreman-Mackey D, Janzing D, Simon-Gabriel CJ et al. Modeling confounding by half-sibling regression. Proceedings of the National Academy of Sciences of the United States of America. 2016 Jul 5;113(27):7391-7398. https://doi.org/10.1073/pnas.1511656113
    Schölkopf, Bernhard ; Hogg, David W. ; Wang, Dun ; Foreman-Mackey, Daniel ; Janzing, Dominik ; Simon-Gabriel, Carl Johann ; Peters, Jonas. / Modeling confounding by half-sibling regression. In: Proceedings of the National Academy of Sciences of the United States of America. 2016 ; Vol. 113, No. 27. pp. 7391-7398.
    @article{54c1c4af71d243ffadc59ab8ae511d3a,
    title = "Modeling confounding by half-sibling regression",
    abstract = "We describe a method for removing the effect of confounders to reconstruct a latent quantity of interest. The method, referred to as {"}half-sibling regression,{"} is inspired by recent work in causal inference using additive noise models. We provide a theoretical justification, discussing both independent and identically distributed as well as time series data, respectively, and illustrate the potential of the method in a challenging astronomy application.",
    keywords = "Astronomy, Causal inference, Exoplanet detection, Machine learning, Systematic error modeling",
    author = "Bernhard Sch{\"o}lkopf and Hogg, {David W.} and Dun Wang and Daniel Foreman-Mackey and Dominik Janzing and Simon-Gabriel, {Carl Johann} and Jonas Peters",
    year = "2016",
    month = "7",
    day = "5",
    doi = "10.1073/pnas.1511656113",
    language = "English (US)",
    volume = "113",
    pages = "7391--7398",
    journal = "Proceedings of the National Academy of Sciences of the United States of America",
    issn = "0027-8424",
    number = "27",

    }

    TY - JOUR

    T1 - Modeling confounding by half-sibling regression

    AU - Schölkopf, Bernhard

    AU - Hogg, David W.

    AU - Wang, Dun

    AU - Foreman-Mackey, Daniel

    AU - Janzing, Dominik

    AU - Simon-Gabriel, Carl Johann

    AU - Peters, Jonas

    PY - 2016/7/5

    Y1 - 2016/7/5

    N2 - We describe a method for removing the effect of confounders to reconstruct a latent quantity of interest. The method, referred to as "half-sibling regression," is inspired by recent work in causal inference using additive noise models. We provide a theoretical justification, discussing both independent and identically distributed as well as time series data, respectively, and illustrate the potential of the method in a challenging astronomy application.

    AB - We describe a method for removing the effect of confounders to reconstruct a latent quantity of interest. The method, referred to as "half-sibling regression," is inspired by recent work in causal inference using additive noise models. We provide a theoretical justification, discussing both independent and identically distributed as well as time series data, respectively, and illustrate the potential of the method in a challenging astronomy application.

    KW - Astronomy

    KW - Causal inference

    KW - Exoplanet detection

    KW - Machine learning

    KW - Systematic error modeling

    UR - http://www.scopus.com/inward/record.url?scp=84977275424&partnerID=8YFLogxK

    UR - http://www.scopus.com/inward/citedby.url?scp=84977275424&partnerID=8YFLogxK

    U2 - 10.1073/pnas.1511656113

    DO - 10.1073/pnas.1511656113

    M3 - Article

    VL - 113

    SP - 7391

    EP - 7398

    JO - Proceedings of the National Academy of Sciences of the United States of America

    JF - Proceedings of the National Academy of Sciences of the United States of America

    SN - 0027-8424

    IS - 27

    ER -