Package {zinbcdpmm}

Title: Bayesian ZINB-Ordinal Joint Model with CDPMM Random Effects
Version: 0.1.0
Description: Fits a Bayesian joint model for longitudinal zero-inflated negative binomial (ZINB) and ordinal outcomes with shared random effects under a centered Dirichlet process mixture model (CDPMM) prior. Provides a blocked Gibbs sampler with Polya-Gamma (PG) and Chinese restaurant table data augmentation, plus helpers for simulation and posterior summaries based on the deviance information criterion (DIC), the widely applicable information criterion (WAIC) and the leave-one-out information criterion (LOOIC). The PG augmentation follows Polson, Scott and Windle (2013) <doi:10.1080/01621459.2013.829001>, the zero-inflated negative binomial construction follows Neelon (2019) <doi:10.1214/18-BA1132>, and the centered Dirichlet process mixture prior follows Tang, Tang and Pan (2014) <doi:10.1016/j.csda.2014.02.015>.
License: MIT + file LICENSE
URL: https://github.com/Liang-ShuQing/zinb-ordinal-cdpmm-code
BugReports: https://github.com/Liang-ShuQing/zinb-ordinal-cdpmm-code/issues
Encoding: UTF-8
RoxygenNote: 7.3.2
Depends: R (≥ 4.1.0)
Imports: BayesLogit, mvtnorm, MCMCpack, truncnorm, loo, stats
Suggests: testthat (≥ 3.0.0)
Config/testthat/edition: 3
Config/roxygen2/version: 8.0.0
NeedsCompilation: no
Packaged: 2026-10-12 08:00:56 UTC; lenovo
Author: Shuqing Liang [aut, cre], Jiang Du [aut]
Maintainer: Shuqing Liang <lsq19853500795@163.com>
Repository: CRAN
Date/Publication: 2026-10-21 16:40:02 UTC

zinbcdpmm: Bayesian ZINB–Ordinal Joint Model with CDPMM Random Effects

Description

Fits a Bayesian joint model for longitudinal zero-inflated negative binomial and ordinal outcomes with CDPMM shared random effects via a blocked Gibbs sampler.

Author(s)

Maintainer: Shuqing Liang lsq19853500795@163.com

Authors:

See Also

Useful links:

Fit joint ZINB-ordinal model with CDPMM random effects

Description

Thin wrapper around fit_joint_model.

Usage

fit_zinb_ordinal(data, chain = 2000, burn = 1000, thin = 5, G = 8, ...)

Arguments

data

A list / zinbcdpmm_data from simulate_zinb_ordinal or with the same fields as generate_data output.

chain

MCMC iterations.

burn

Burn-in.

thin

Thinning interval.

G

Truncation level of the stick-breaking CDPMM (default 8).

...

Additional arguments passed to fit_joint_model (e.g. iw_nu0, kappa0, r_update).

Value

An object of class zinbcdpmm_fit.

Examples

set.seed(1)
dat <- simulate_zinb_ordinal(n = 8, nis = 2, scenario = 1)
# Tiny chain for a quick check; use much larger chain/burn in applications.
fit <- suppressWarnings(
  fit_zinb_ordinal(dat, chain = 25, burn = 10, thin = 5, G = 2)
)
print(fit)

Simulate longitudinal ZINB + ordinal data

Description

Thin wrapper around generate_data. Returns an object of class zinbcdpmm_data.

Usage

simulate_zinb_ordinal(
  n = 100,
  nis = 10,
  seed = 2026,
  random_nis = FALSE,
  nis_range = 1:20,
  scenario = 1,
  C = 5,
  re_dist = "normal",
  mvt_df = 4,
  rho = NULL,
  zi_intercept = NULL,
  ...
)

Arguments

n

Number of subjects.

nis

Visits per subject (scalar or length-n vector).

seed

RNG seed.

random_nis

If TRUE, sample visit counts from nis_range.

nis_range

Integer range for random visit counts.

scenario

Simulation scenario (1, 2, or 3).

C

Number of ordinal categories.

re_dist

Random-effects truth: "normal", "mixture", or "mvt".

mvt_df

Degrees of freedom when re_dist = "mvt".

rho

Common RE correlation (default 0.5).

zi_intercept

Optional override of zero-inflation intercept.

...

Passed to generate_data.

Value

A list with class zinbcdpmm_data.

Examples

set.seed(1)
dat <- simulate_zinb_ordinal(n = 8, nis = 2, scenario = 1)
stopifnot(inherits(dat, "zinbcdpmm_data"), dat$n == 8L)

Print and summarize zinbcdpmm fits

Description

Print and summarize zinbcdpmm fits

Usage

## S3 method for class 'zinbcdpmm_fit'
print(x, ...)

## S3 method for class 'zinbcdpmm_fit'
summary(object, ...)

## S3 method for class 'summary.zinbcdpmm_fit'
print(x, digits = 3, ...)

Arguments

x

A zinbcdpmm_fit or its summary.

...

Unused.

object

A zinbcdpmm_fit.

digits

Digits for printing.

Value

print.zinbcdpmm_fit and print.summary.zinbcdpmm_fit invisibly return their input object (they are called for their side effect of printing to the console). summary.zinbcdpmm_fit returns an object of class summary.zinbcdpmm_fit: a list containing posterior means and 95% credible intervals for the zero-inflation (alpha), count (beta) and ordinal (gamma) fixed effects, the thresholds (delta), the negative-binomial size (r), the implied random-effects covariance Sigma and correlations Rho, and the information criteria dic, waic and looic, together with nclust_est.

mirror server hosted at Truenetwork, Russian Federation.