Bayesian inference for the one-factor copula model

Research output: Contribution to journalArticleResearchpeer-review

Abstract

We develop efficient Bayesian inference for the one-factor copula model with two significant contributions over existing methodologies. First, our approach leads to straightforward inference on dependence parameters and the latent factor; only inference on the former is available under frequentist alternatives. Second, we develop a reversible jump Markov chain Monte Carlo algorithm that averages over models constructed from different bivariate copula building blocks. Our approach accommodates any combination of discrete and continuous margins. Through extensive simulations, we compare the computational and Monte Carlo efficiency of alternative proposed sampling schemes. The preferred algorithm provides reliable inference on parameters, the latent factor, and model space. The potential of the methodology is highlighted in an empirical study of 10 binary measures of socio-economic deprivation collected for 11,463 East Timorese households. The importance of conducting inference on the latent factor is motivated by constructing a poverty index using estimates of the factor. Compared to a linear Gaussian factor model, our model average improves out-of-sample fit. The relationships between the poverty index and observed variables uncovered by our approach are diverse and allow for a richer and more precise understanding of the dependence between overall deprivation and individual measures of well-being.

Original languageEnglish
Pages (from-to)155-173
Number of pages19
JournalJournal of Computational and Graphical Statistics
Volume28
Issue number1
DOIs
Publication statusPublished - 2019

Keywords

  • Model selection
  • Multidimensional poverty index
  • Multivariate analysis
  • Reversible jump MCMC
  • Vine copulas

Cite this

@article{ad0000bfaf4543c9a5e06d48de3c856c,
title = "Bayesian inference for the one-factor copula model",
abstract = "We develop efficient Bayesian inference for the one-factor copula model with two significant contributions over existing methodologies. First, our approach leads to straightforward inference on dependence parameters and the latent factor; only inference on the former is available under frequentist alternatives. Second, we develop a reversible jump Markov chain Monte Carlo algorithm that averages over models constructed from different bivariate copula building blocks. Our approach accommodates any combination of discrete and continuous margins. Through extensive simulations, we compare the computational and Monte Carlo efficiency of alternative proposed sampling schemes. The preferred algorithm provides reliable inference on parameters, the latent factor, and model space. The potential of the methodology is highlighted in an empirical study of 10 binary measures of socio-economic deprivation collected for 11,463 East Timorese households. The importance of conducting inference on the latent factor is motivated by constructing a poverty index using estimates of the factor. Compared to a linear Gaussian factor model, our model average improves out-of-sample fit. The relationships between the poverty index and observed variables uncovered by our approach are diverse and allow for a richer and more precise understanding of the dependence between overall deprivation and individual measures of well-being.",
keywords = "Model selection, Multidimensional poverty index, Multivariate analysis, Reversible jump MCMC, Vine copulas",
author = "Tan, {Ban Kheng} and Anastasios Panagiotelis and George Athanasopoulos",
year = "2019",
doi = "10.1080/10618600.2018.1482765",
language = "English",
volume = "28",
pages = "155--173",
journal = "Journal of Computational and Graphical Statistics",
issn = "1061-8600",
publisher = "Taylor & Francis",
number = "1",

}

Bayesian inference for the one-factor copula model. / Tan, Ban Kheng; Panagiotelis, Anastasios; Athanasopoulos, George.

In: Journal of Computational and Graphical Statistics, Vol. 28, No. 1, 2019, p. 155-173.

Research output: Contribution to journalArticleResearchpeer-review

TY - JOUR

T1 - Bayesian inference for the one-factor copula model

AU - Tan, Ban Kheng

AU - Panagiotelis, Anastasios

AU - Athanasopoulos, George

PY - 2019

Y1 - 2019

N2 - We develop efficient Bayesian inference for the one-factor copula model with two significant contributions over existing methodologies. First, our approach leads to straightforward inference on dependence parameters and the latent factor; only inference on the former is available under frequentist alternatives. Second, we develop a reversible jump Markov chain Monte Carlo algorithm that averages over models constructed from different bivariate copula building blocks. Our approach accommodates any combination of discrete and continuous margins. Through extensive simulations, we compare the computational and Monte Carlo efficiency of alternative proposed sampling schemes. The preferred algorithm provides reliable inference on parameters, the latent factor, and model space. The potential of the methodology is highlighted in an empirical study of 10 binary measures of socio-economic deprivation collected for 11,463 East Timorese households. The importance of conducting inference on the latent factor is motivated by constructing a poverty index using estimates of the factor. Compared to a linear Gaussian factor model, our model average improves out-of-sample fit. The relationships between the poverty index and observed variables uncovered by our approach are diverse and allow for a richer and more precise understanding of the dependence between overall deprivation and individual measures of well-being.

AB - We develop efficient Bayesian inference for the one-factor copula model with two significant contributions over existing methodologies. First, our approach leads to straightforward inference on dependence parameters and the latent factor; only inference on the former is available under frequentist alternatives. Second, we develop a reversible jump Markov chain Monte Carlo algorithm that averages over models constructed from different bivariate copula building blocks. Our approach accommodates any combination of discrete and continuous margins. Through extensive simulations, we compare the computational and Monte Carlo efficiency of alternative proposed sampling schemes. The preferred algorithm provides reliable inference on parameters, the latent factor, and model space. The potential of the methodology is highlighted in an empirical study of 10 binary measures of socio-economic deprivation collected for 11,463 East Timorese households. The importance of conducting inference on the latent factor is motivated by constructing a poverty index using estimates of the factor. Compared to a linear Gaussian factor model, our model average improves out-of-sample fit. The relationships between the poverty index and observed variables uncovered by our approach are diverse and allow for a richer and more precise understanding of the dependence between overall deprivation and individual measures of well-being.

KW - Model selection

KW - Multidimensional poverty index

KW - Multivariate analysis

KW - Reversible jump MCMC

KW - Vine copulas

UR - http://www.scopus.com/inward/record.url?scp=85053268730&partnerID=8YFLogxK

U2 - 10.1080/10618600.2018.1482765

DO - 10.1080/10618600.2018.1482765

M3 - Article

VL - 28

SP - 155

EP - 173

JO - Journal of Computational and Graphical Statistics

JF - Journal of Computational and Graphical Statistics

SN - 1061-8600

IS - 1

ER -