Cobin and micobin regression models are scalable and robust alternative to beta regression model for continuous proportional data. See the following paper for more details:

Lee, C. J., Dahl, B. K., Ovaskainen, O., Dunson, D. B. (2026). Scalable and robust regression models for continuous proportional data. Journal of the American Statistical Association, in press.

Preprint is available at https://arxiv.org/abs/2504.15269 as well as a journal version at https://doi.org/10.1080/01621459.2026.2626081.

A dedicated Github repository for reproducing the analysis in the paper is available at https://github.com/changwoo-lee/cobin-reproduce. This R package repository contains the functions for the cobin and micobin regression models, as well as sampler for Kolmogorov-Gamma random variables.

Install the package:

install.packages("cobin") # from CRAN
# or the development version from GitHub
# install.packages("devtools")
devtools::install_github("changwoo-lee/cobin")

Glossaries: GLM: generalized linear model; GLMM: generalized linear mixed model; GP: Gaussian process; NNGP: nearest neighbor Gaussian process; cobin: continuous binomial; micobin: mixture of continuous binomial;

vignette

Comparison of cobin and beta density

Comparison of micobin and beta density

Please see MMI data analysis code corresponding to the Section 5 of the paper(https://doi.org/10.1080/01621459.2026.2626081). More detailed examples TBA.

Code structure

Basic functions

• cobin.R:
  • dcobin(x, theta, lambda): Density of