Identifiability and estimation of recursive max‐linear models

Publikation: Bidrag til tidsskrift › Tidsskriftartikel › Forskning › fagfællebedømt

Standard

Identifiability and estimation of recursive max‐linear models. / Gissibl, Nadine; Klüppelberg, Claudia; Lauritzen, Steffen.

I: Scandinavian Journal of Statistics, Bind 48, Nr. 1, 2021, s. 188-211.

Publikation: Bidrag til tidsskrift › Tidsskriftartikel › Forskning › fagfællebedømt

Harvard

Gissibl, N, Klüppelberg, C & Lauritzen, S 2021, 'Identifiability and estimation of recursive max‐linear models', Scandinavian Journal of Statistics, bind 48, nr. 1, s. 188-211. https://doi.org/10.1111/sjos.12446

APA

Gissibl, N., Klüppelberg, C., & Lauritzen, S. (2021). Identifiability and estimation of recursive max‐linear models. Scandinavian Journal of Statistics, 48(1), 188-211. https://doi.org/10.1111/sjos.12446

Vancouver

Gissibl N, Klüppelberg C, Lauritzen S. Identifiability and estimation of recursive max‐linear models. Scandinavian Journal of Statistics. 2021;48(1):188-211. https://doi.org/10.1111/sjos.12446

Author

Gissibl, Nadine ; Klüppelberg, Claudia ; Lauritzen, Steffen. / Identifiability and estimation of recursive max‐linear models. I: Scandinavian Journal of Statistics. 2021 ; Bind 48, Nr. 1. s. 188-211.

Bibtex

@article{ebdb7a02214e45569d7a1c8d93157fc0,
title = "Identifiability and estimation of recursive max‐linear models",
abstract = "We address the identifiability and estimation of recursive max‐linear structural equation models represented by an edge‐weighted directed acyclic graph (DAG). Such models are generally unidentifiable and we identify the whole class of DAG s and edge weights corresponding to a given observational distribution. For estimation, standard likelihood theory cannot be applied because the corresponding families of distributions are not dominated. Given the underlying DAG, we present an estimator for the class of edge weights and show that it can be considered a generalized maximum likelihood estimator. In addition, we develop a simple method for identifying the structure of the DAG. With probability tending to one at an exponential rate with the number of observations, this method correctly identifies the class of DAGs and, similarly, exactly identifies the possible edge weights.",
author = "Nadine Gissibl and Claudia Kl{\"u}ppelberg and Steffen Lauritzen",
year = "2021",
doi = "10.1111/sjos.12446",
language = "English",
volume = "48",
pages = "188--211",
journal = "Scandinavian Journal of Statistics",
issn = "0303-6898",
publisher = "Wiley-Blackwell",
number = "1",

}

RIS

TY - JOUR

T1 - Identifiability and estimation of recursive max‐linear models

AU - Gissibl, Nadine

AU - Klüppelberg, Claudia

AU - Lauritzen, Steffen

PY - 2021

Y1 - 2021

N2 - We address the identifiability and estimation of recursive max‐linear structural equation models represented by an edge‐weighted directed acyclic graph (DAG). Such models are generally unidentifiable and we identify the whole class of DAG s and edge weights corresponding to a given observational distribution. For estimation, standard likelihood theory cannot be applied because the corresponding families of distributions are not dominated. Given the underlying DAG, we present an estimator for the class of edge weights and show that it can be considered a generalized maximum likelihood estimator. In addition, we develop a simple method for identifying the structure of the DAG. With probability tending to one at an exponential rate with the number of observations, this method correctly identifies the class of DAGs and, similarly, exactly identifies the possible edge weights.

AB - We address the identifiability and estimation of recursive max‐linear structural equation models represented by an edge‐weighted directed acyclic graph (DAG). Such models are generally unidentifiable and we identify the whole class of DAG s and edge weights corresponding to a given observational distribution. For estimation, standard likelihood theory cannot be applied because the corresponding families of distributions are not dominated. Given the underlying DAG, we present an estimator for the class of edge weights and show that it can be considered a generalized maximum likelihood estimator. In addition, we develop a simple method for identifying the structure of the DAG. With probability tending to one at an exponential rate with the number of observations, this method correctly identifies the class of DAGs and, similarly, exactly identifies the possible edge weights.

U2 - 10.1111/sjos.12446

DO - 10.1111/sjos.12446

M3 - Journal article

VL - 48

SP - 188

EP - 211

JO - Scandinavian Journal of Statistics

JF - Scandinavian Journal of Statistics

SN - 0303-6898

IS - 1

ER -

ID: 240251371