Home > Research > Publications & Outputs > A comparative review of dimension reduction met...

Research

Electronic data

Links

Text available via DOI:

Keywords

View graph of relations

A comparative review of dimension reduction methods in approximate Bayesian computation

Research output: Contribution to Journal/Magazine › Journal article › peer-review

Published

Standard

A comparative review of dimension reduction methods in approximate Bayesian computation. / Blum, M. G. B.; Nunes, Matthew; Prangle, Dennis et al.
In: Statistical Science, Vol. 28, No. 2, 05.2013, p. 189-208.

Research output: Contribution to Journal/Magazine › Journal article › peer-review

Harvard

Blum, MGB, Nunes, M, Prangle, D & Sisson, SA 2013, 'A comparative review of dimension reduction methods in approximate Bayesian computation', Statistical Science, vol. 28, no. 2, pp. 189-208. https://doi.org/10.1214/12-STS406SUPP

APA

Blum, M. G. B., Nunes, M., Prangle, D., & Sisson, S. A. (2013). A comparative review of dimension reduction methods in approximate Bayesian computation. Statistical Science, 28(2), 189-208. https://doi.org/10.1214/12-STS406SUPP

Vancouver

Blum MGB, Nunes M, Prangle D, Sisson SA. A comparative review of dimension reduction methods in approximate Bayesian computation. Statistical Science. 2013 May;28(2):189-208. Epub 2012 Nov 1. doi: 10.1214/12-STS406SUPP

Author

Blum, M. G. B. ; Nunes, Matthew ; Prangle, Dennis et al. / A comparative review of dimension reduction methods in approximate Bayesian computation. In: Statistical Science. 2013 ; Vol. 28, No. 2. pp. 189-208.

Bibtex

@article{73d8a83047654102a4a86f349c464af3,
title = "A comparative review of dimension reduction methods in approximate Bayesian computation",
abstract = "Approximate Bayesian computation (ABC) methods make use of comparisons between simulated and observed summary statistics to overcome the problem of computationally intractable likelihood functions. As the practical implementation of ABC requires computations based on vectors of summary statistics, rather than full datasets, a central question is how to derive low dimensional summary statistics from the observed data with minimal loss of information. In this article we provide a comprehensive review and comparison of the performance of the principal methods of dimension reduction proposed in the ABC literature. The methods are split into three non-mutually exclusive classes consisting of best subset selection methods, projection techniques and regularisation. In addition, we introduce two new methods of dimension reduction. The first is a best subset selection method based on Akaike and Bayesian information criteria, and the second uses ridge regression as a regularisation procedure. We illustrate the performance of these dimension reduction techniques through the analysis of three challenging models and data sets.",
keywords = "Approximate Bayesian computation , dimension reduction , likelihood-free inference , regularization , variable selection",
author = "Blum, {M. G. B.} and Matthew Nunes and Dennis Prangle and Sisson, {S. A.}",
year = "2013",
month = may,
doi = "10.1214/12-STS406SUPP",
language = "English",
volume = "28",
pages = "189--208",
journal = "Statistical Science",
issn = "0883-4237",
publisher = "Institute of Mathematical Statistics",
number = "2",

}

RIS

TY - JOUR

T1 - A comparative review of dimension reduction methods in approximate Bayesian computation

AU - Blum, M. G. B.

AU - Nunes, Matthew

AU - Prangle, Dennis

AU - Sisson, S. A.

PY - 2013/5

Y1 - 2013/5

N2 - Approximate Bayesian computation (ABC) methods make use of comparisons between simulated and observed summary statistics to overcome the problem of computationally intractable likelihood functions. As the practical implementation of ABC requires computations based on vectors of summary statistics, rather than full datasets, a central question is how to derive low dimensional summary statistics from the observed data with minimal loss of information. In this article we provide a comprehensive review and comparison of the performance of the principal methods of dimension reduction proposed in the ABC literature. The methods are split into three non-mutually exclusive classes consisting of best subset selection methods, projection techniques and regularisation. In addition, we introduce two new methods of dimension reduction. The first is a best subset selection method based on Akaike and Bayesian information criteria, and the second uses ridge regression as a regularisation procedure. We illustrate the performance of these dimension reduction techniques through the analysis of three challenging models and data sets.

AB - Approximate Bayesian computation (ABC) methods make use of comparisons between simulated and observed summary statistics to overcome the problem of computationally intractable likelihood functions. As the practical implementation of ABC requires computations based on vectors of summary statistics, rather than full datasets, a central question is how to derive low dimensional summary statistics from the observed data with minimal loss of information. In this article we provide a comprehensive review and comparison of the performance of the principal methods of dimension reduction proposed in the ABC literature. The methods are split into three non-mutually exclusive classes consisting of best subset selection methods, projection techniques and regularisation. In addition, we introduce two new methods of dimension reduction. The first is a best subset selection method based on Akaike and Bayesian information criteria, and the second uses ridge regression as a regularisation procedure. We illustrate the performance of these dimension reduction techniques through the analysis of three challenging models and data sets.

KW - Approximate Bayesian computation

KW - dimension reduction

KW - likelihood-free inference

KW - regularization

KW - variable selection

U2 - 10.1214/12-STS406SUPP

DO - 10.1214/12-STS406SUPP

M3 - Journal article

VL - 28

SP - 189

EP - 208

JO - Statistical Science

JF - Statistical Science

SN - 0883-4237

IS - 2

ER -