Mathematics · Statistics and Probability
A Bayesian Framework for Estimating the Concordance Correlation Coefficient Using Skew-elliptical Distributions
Dai Feng, Richard Baumgartner, Vladimir Svetnik
See on the mapAbstract
The concordance correlation coefficient (CCC) is a widely used scaled index in the study of agreement. In this article, we propose estimating the CCC by a unified Bayesian framework that can (1) accommodate symmetric or asymmetric and light- or heavy-tailed data; (2) select model from several candidates; and (3) address other issues frequently encountered in practice such as confounding covariates and missing data. The performance of the proposal was studied and demonstrated using simulated as well as real-life biomarker data from a clinical study of an insomnia drug. The implementation of the proposal is accessible through a package in the Comprehensive R Archive Network.