Bayesian Latent Class Analysis with Shrinkage Priors: An Application to the Hungarian Heart Disease Data

Bettina Grün; Gertraud Malsiner-Walli

doi:10.6093/978-88-6887-042-3

Bayesian Latent Class Analysis with Shrinkage Priors: An Application to the Hungarian Heart Disease Data

Bettina Grün, Gertraud Malsiner-Walli

Statistics and Mathematics

Publikation: Beitrag in Buch/Konferenzband › Beitrag in Konferenzband

24 Downloads (Pure)

Abstract

Latent class analysis explains dependency structures in multivariate categorical
data by assuming the presence of latent classes. We investigate the specification of suitable
priors for the Bayesian latent class model to determine the number of classes and perform
variable selection. Estimation is possible using standard tools implementing general purpose
Markov chain Monte Carlo sampling techniques such as the software JAGS. However, class
specific inference requires suitable post-processing in order to eliminate label switching. The
proposed Bayesian specification and analysis method is applied to the Hungarian heart disease
data set to determine the number of classes and identify relevant variables and results are
compared to those obtained with the standard prior for the component specific parameters.

Originalsprache	Englisch
Titel des Sammelwerks	ASMOD 2018 -- Proceedings of the International Conference on Advances in Statistical Modelling of Ordinal Data
Herausgeber*innen	Stefania Capecchi and Francesca Di Iorio and Rosaria Simone
Erscheinungsort	FedOA -- Federico II University Press
Seiten	13 - 24
DOIs	https://doi.org/10.6093/978-88-6887-042-3
Publikationsstatus	Veröffentlicht - 2018

Österreichische Systematik der Wissenschaftszweige (ÖFOS)

101018 Statistik

Zugriff auf Dokument

10.6093/978-88-6887-042-3Lizenz: Nicht angegeben

HeartEndgültige, publizierte Fassung, 3,19 MBLizenz: CC BY

http://www.fedoabooks.unina.it/index.php/fedoapress/catalog/book/91Lizenz: Nicht angegeben

Verfügbarkeit prüfen

Katalog der Universitätsbibliothek

Shrinking and Regularizing Finite Mixture Models
Frühwirth-Schnatter, S. (Projektleitung) & Malsiner-Walli, G. (Forscher*in)
1/11/16 → 30/04/22
Projekt: Forschungsförderung

Zitat

Grün, B., & Malsiner-Walli, G. (2018). Bayesian Latent Class Analysis with Shrinkage Priors: An Application to the Hungarian Heart Disease Data. in Stefania Capecchi and Francesca Di Iorio and Rosaria Simone (Hrsg.), ASMOD 2018 -- Proceedings of the International Conference on Advances in Statistical Modelling of Ordinal Data (S. 13 - 24). https://doi.org/10.6093/978-88-6887-042-3

Grün, Bettina ; Malsiner-Walli, Gertraud. / Bayesian Latent Class Analysis with Shrinkage Priors: An Application to the Hungarian Heart Disease Data. ASMOD 2018 -- Proceedings of the International Conference on Advances in Statistical Modelling of Ordinal Data. Hrsg. / Stefania Capecchi and Francesca Di Iorio and Rosaria Simone. FedOA -- Federico II University Press, 2018. S. 13 - 24

@inproceedings{20a37a1703b944fe90d977a65c3e9022,

title = "Bayesian Latent Class Analysis with Shrinkage Priors: An Application to the Hungarian Heart Disease Data",

abstract = "Latent class analysis explains dependency structures in multivariate categorical data by assuming the presence of latent classes. We investigate the specification of suitable priors for the Bayesian latent class model to determine the number of classes and perform variable selection. Estimation is possible using standard tools implementing general purpose Markov chain Monte Carlo sampling techniques such as the software JAGS. However, class specific inference requires suitable post-processing in order to eliminate label switching. The proposed Bayesian specification and analysis method is applied to the Hungarian heart disease data set to determine the number of classes and identify relevant variables and results are compared to those obtained with the standard prior for the component specific parameters.",

author = "Bettina Gr{\"u}n and Gertraud Malsiner-Walli",

year = "2018",

doi = "10.6093/978-88-6887-042-3",

language = "English",

pages = "13 -- 24",

editor = "{Stefania Capecchi and Francesca Di Iorio and Rosaria Simone}",

booktitle = "ASMOD 2018 -- Proceedings of the International Conference on Advances in Statistical Modelling of Ordinal Data",

}

Grün, B & Malsiner-Walli, G 2018, Bayesian Latent Class Analysis with Shrinkage Priors: An Application to the Hungarian Heart Disease Data. in Stefania Capecchi and Francesca Di Iorio and Rosaria Simone (Hrsg.), ASMOD 2018 -- Proceedings of the International Conference on Advances in Statistical Modelling of Ordinal Data. FedOA -- Federico II University Press, S. 13 - 24. https://doi.org/10.6093/978-88-6887-042-3

Bayesian Latent Class Analysis with Shrinkage Priors: An Application to the Hungarian Heart Disease Data. / Grün, Bettina ; Malsiner-Walli, Gertraud.
ASMOD 2018 -- Proceedings of the International Conference on Advances in Statistical Modelling of Ordinal Data. Hrsg. / Stefania Capecchi and Francesca Di Iorio and Rosaria Simone. FedOA -- Federico II University Press, 2018. S. 13 - 24.

Publikation: Beitrag in Buch/Konferenzband › Beitrag in Konferenzband

TY - GEN

T1 - Bayesian Latent Class Analysis with Shrinkage Priors: An Application to the Hungarian Heart Disease Data

AU - Grün, Bettina

AU - Malsiner-Walli, Gertraud

PY - 2018

Y1 - 2018

N2 - Latent class analysis explains dependency structures in multivariate categorical data by assuming the presence of latent classes. We investigate the specification of suitable priors for the Bayesian latent class model to determine the number of classes and perform variable selection. Estimation is possible using standard tools implementing general purpose Markov chain Monte Carlo sampling techniques such as the software JAGS. However, class specific inference requires suitable post-processing in order to eliminate label switching. The proposed Bayesian specification and analysis method is applied to the Hungarian heart disease data set to determine the number of classes and identify relevant variables and results are compared to those obtained with the standard prior for the component specific parameters.

AB - Latent class analysis explains dependency structures in multivariate categorical data by assuming the presence of latent classes. We investigate the specification of suitable priors for the Bayesian latent class model to determine the number of classes and perform variable selection. Estimation is possible using standard tools implementing general purpose Markov chain Monte Carlo sampling techniques such as the software JAGS. However, class specific inference requires suitable post-processing in order to eliminate label switching. The proposed Bayesian specification and analysis method is applied to the Hungarian heart disease data set to determine the number of classes and identify relevant variables and results are compared to those obtained with the standard prior for the component specific parameters.

U2 - 10.6093/978-88-6887-042-3

DO - 10.6093/978-88-6887-042-3

M3 - Contribution to conference proceedings

SP - 13

EP - 24

BT - ASMOD 2018 -- Proceedings of the International Conference on Advances in Statistical Modelling of Ordinal Data

A2 - Stefania Capecchi and Francesca Di Iorio and Rosaria Simone, null

CY - FedOA -- Federico II University Press

ER -

Grün B , Malsiner-Walli G. Bayesian Latent Class Analysis with Shrinkage Priors: An Application to the Hungarian Heart Disease Data. in Stefania Capecchi and Francesca Di Iorio and Rosaria Simone, Hrsg., ASMOD 2018 -- Proceedings of the International Conference on Advances in Statistical Modelling of Ordinal Data. FedOA -- Federico II University Press. 2018. S. 13 - 24 doi: 10.6093/978-88-6887-042-3

Bayesian Latent Class Analysis with Shrinkage Priors: An Application to the Hungarian Heart Disease Data

Abstract

Österreichische Systematik der Wissenschaftszweige (ÖFOS)

Zugriff auf Dokument

Verfügbarkeit prüfen

Projekte

Shrinking and Regularizing Finite Mixture Models

Zitat