We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

stat.ME

Change to browse by:

References & Citations

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Statistics > Methodology

Title: Bayesian Cumulative Probability Models for Continuous and Mixed Outcomes

Abstract: Ordinal cumulative probability models (CPMs) -- also known as cumulative link models -- such as the proportional odds regression model are typically used for discrete ordered outcomes, but can accommodate both continuous and mixed discrete/continuous outcomes since these are also ordered. Recent papers by Liu et al. and Tian et al. describe ordinal CPMs in this setting using non-parametric maximum likelihood estimation. We formulate a Bayesian CPM for continuous or mixed outcome data. Bayesian CPMs inherit many of the benefits of frequentist CPMs and have advantages with regard to interpretation, flexibility, and exact inference (within simulation error) for parameters and functions of parameters. We explore characteristics of the Bayesian CPM through simulations and a case study using HIV biomarker data. In addition, we provide the package 'bayesCPM' which implements Bayesian CPM models using the R interface to the Stan probabilistic programing language. The Bayesian CPM for continuous outcomes can be implemented with only minor modifications to the prior specification and -- despite several limitations -- has generally good statistical performance with moderate or large sample sizes.
Comments: 30 pages, 25 figures
Subjects: Methodology (stat.ME)
Cite as: arXiv:2102.00330 [stat.ME]
  (or arXiv:2102.00330v1 [stat.ME] for this version)

Submission history

From: Nathan James [view email]
[v1] Sat, 30 Jan 2021 22:25:46 GMT (3512kb,D)
[v2] Fri, 7 Jan 2022 21:37:00 GMT (2530kb,D)

Link back to: arXiv, form interface, contact.