A general class of hierarchical ordinal regression models with applications to correlated ROC analysis

Hemant Ishwaran, Constantine A. Gatsonis

Research output: Contribution to journalArticle

64 Citations (Scopus)

Abstract

The authors discuss a general class of hierarchical ordinal regression models that includes both location and scale parameters, allows link functions to be selected adaptively as finite mixtures of normal cumulative distribution functions, and incorporates flexible correlation structures for the latent scale variables. Exploiting the well-known correspondence between ordinal regression models and parametric ROC (Receiver Operating Characteristic) curves makes it possible to use a hierarchical ROC (HROC) analysis to study multilevel clustered data in diagnostic imaging studies. The authors present a Bayesian approach to model fitting using Markov chain Monte Carlo methods and discuss HROC applications to the analysis of data from two diagnostic radiology studies involving multiple interpreters.

Original languageEnglish
Pages (from-to)731-750
Number of pages20
JournalCanadian Journal of Statistics
Volume28
Issue number4
StatePublished - Dec 1 2000
Externally publishedYes

Fingerprint

Ordinal Regression
Operating Characteristics
Regression Model
Receiver
Normal cumulative distribution function
Diagnostics
Clustered Data
Finite Mixture
Link Function
Flexible Structure
Location Parameter
Model Fitting
Markov Chain Monte Carlo Methods
Correlation Structure
Receiver Operating Characteristic Curve
Scale Parameter
Bayesian Approach
Correspondence
Imaging
Class

Keywords

  • Bayesian hierarchical model
  • Gibbs sampling
  • HROC model
  • Ordinal categorical data
  • Ordinal regression
  • ROC curve

ASJC Scopus subject areas

  • Statistics and Probability

Cite this

A general class of hierarchical ordinal regression models with applications to correlated ROC analysis. / Ishwaran, Hemant; Gatsonis, Constantine A.

In: Canadian Journal of Statistics, Vol. 28, No. 4, 01.12.2000, p. 731-750.

Research output: Contribution to journalArticle

@article{464e7d8424564973978957f79ddafb61,
title = "A general class of hierarchical ordinal regression models with applications to correlated ROC analysis",
abstract = "The authors discuss a general class of hierarchical ordinal regression models that includes both location and scale parameters, allows link functions to be selected adaptively as finite mixtures of normal cumulative distribution functions, and incorporates flexible correlation structures for the latent scale variables. Exploiting the well-known correspondence between ordinal regression models and parametric ROC (Receiver Operating Characteristic) curves makes it possible to use a hierarchical ROC (HROC) analysis to study multilevel clustered data in diagnostic imaging studies. The authors present a Bayesian approach to model fitting using Markov chain Monte Carlo methods and discuss HROC applications to the analysis of data from two diagnostic radiology studies involving multiple interpreters.",
keywords = "Bayesian hierarchical model, Gibbs sampling, HROC model, Ordinal categorical data, Ordinal regression, ROC curve",
author = "Hemant Ishwaran and Gatsonis, {Constantine A.}",
year = "2000",
month = "12",
day = "1",
language = "English",
volume = "28",
pages = "731--750",
journal = "Canadian Journal of Statistics",
issn = "0319-5724",
publisher = "Statistical Society of Canada",
number = "4",

}

TY - JOUR

T1 - A general class of hierarchical ordinal regression models with applications to correlated ROC analysis

AU - Ishwaran, Hemant

AU - Gatsonis, Constantine A.

PY - 2000/12/1

Y1 - 2000/12/1

N2 - The authors discuss a general class of hierarchical ordinal regression models that includes both location and scale parameters, allows link functions to be selected adaptively as finite mixtures of normal cumulative distribution functions, and incorporates flexible correlation structures for the latent scale variables. Exploiting the well-known correspondence between ordinal regression models and parametric ROC (Receiver Operating Characteristic) curves makes it possible to use a hierarchical ROC (HROC) analysis to study multilevel clustered data in diagnostic imaging studies. The authors present a Bayesian approach to model fitting using Markov chain Monte Carlo methods and discuss HROC applications to the analysis of data from two diagnostic radiology studies involving multiple interpreters.

AB - The authors discuss a general class of hierarchical ordinal regression models that includes both location and scale parameters, allows link functions to be selected adaptively as finite mixtures of normal cumulative distribution functions, and incorporates flexible correlation structures for the latent scale variables. Exploiting the well-known correspondence between ordinal regression models and parametric ROC (Receiver Operating Characteristic) curves makes it possible to use a hierarchical ROC (HROC) analysis to study multilevel clustered data in diagnostic imaging studies. The authors present a Bayesian approach to model fitting using Markov chain Monte Carlo methods and discuss HROC applications to the analysis of data from two diagnostic radiology studies involving multiple interpreters.

KW - Bayesian hierarchical model

KW - Gibbs sampling

KW - HROC model

KW - Ordinal categorical data

KW - Ordinal regression

KW - ROC curve

UR - http://www.scopus.com/inward/record.url?scp=0034557373&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0034557373&partnerID=8YFLogxK

M3 - Article

VL - 28

SP - 731

EP - 750

JO - Canadian Journal of Statistics

JF - Canadian Journal of Statistics

SN - 0319-5724

IS - 4

ER -