gdpar: General Dynamic Parameter Models via Reference Anchoring

Implements a unified predictive framework in which individual parameters are decomposed as theta_i equal to theta_ref plus Delta(x_i, theta_ref), with theta_ref a population reference and Delta an explicit deviation function. The decomposition follows the Additive-Multiplicative-Modulated canonical form and is estimated through three complementary paths: hierarchical Bayesian inference via 'Stan', varying-coefficient models via penalized splines, and amortized inference via hypernetworks in 'torch'. The package provides identifiability diagnostics, validity tests for the population reference, and benchmarks against canonical zero-inflated count datasets and avian abundance data from the eBird Status and Trends project. The framework and its estimation paths are described in Gomez Julian (2026) <doi:10.5281/zenodo.21046269>.

Version:	0.1.0
Depends:	R (≥ 4.2.0)
Imports:	posterior, stats, methods, withr
Suggests:	cmdstanr, loo, bayesplot, DHARMa, digest, mgcv, Matrix, pscl, AER, ebirdst, knitr, rmarkdown, testthat (≥ 3.0.0), kableExtra, brms, INLA, rstanarm, scoringRules, gratia, grf, reticulate, mvnfast
Published:	2026-07-15
DOI:	10.32614/CRAN.package.gdpar (may not be active yet)
Author:	José Mauricio Gómez Julián [aut, cre]
Maintainer:	José Mauricio Gómez Julián <isadore.nabi at pm.me>
BugReports:	https://github.com/IsadoreNabi/gdpar/issues
License:	GPL (≥ 3)
URL:	https://github.com/IsadoreNabi/gdpar
NeedsCompilation:	no
Additional_repositories:	https://stan-dev.r-universe.dev, https://inla.r-inla-download.org/R/stable
Materials:	README, NEWS
CRAN checks:	gdpar results

Documentation:

Reference manual:

gdpar.html , gdpar.pdf

Vignettes:

Predictive Models with Dynamic Individual Parameters: A Unifying Conceptual Framework (source, R code) The AMM Canonical Form and Identifiability Conditions (source, R code) Gnoseological Validity Conditions for the Population Reference (source, R code) Standard Predictive Models as Formal Special Cases of the AMM (source, R code) Asymptotic Theory for Path 1 (Hierarchical Bayesian) (source, R code) Asymptotic Theory for Path 2 (Varying-Coefficient via Penalized Splines) (source, R code) Asymptotic Theory for Path 3 (Hypernetwork) (source, R code) Empirical Bayes vs. Fully Bayes Treatment of the Population Reference (source, R code) Empirical Bayes vs. Fully Bayes – Multivariate Extension (source, R code) Positioning AMM relative to the CATE / ITE Literature (source, R code) The T-learner AMM-side Causal Bridge (source, R code) Theoretical Addendum 8.5.B: Comparison against External Meta-learners (source, R code) AMM Sub-phases 8.3.1 to 8.3.10 – Theoretical Canonization (source, R code) Cognitive Motivation: From Driver Prediction to Reference-Anchored Individuation (source, R code) Quickstart: A First Fit in Five Minutes (source, R code) Parametrization Toggle: Operational Guide (source, R code) Arbitrary p: Operational Cookbook for Multivariate Fits (source, R code) Per-group hierarchical anchors: Operational Guide (source, R code) Regression Testing of MCMC Outputs (Experimental) (source, R code) Intermediate AMM Specifications: B-spline W Bases and Heterogeneous Families per Slot (source, R code) Distributional Regression K > 1 and Residual Diagnostics with DHARMa (source, R code) Comparing the AMM-side T-learner against External Meta-learners (source, R code) The Empirical-Bayes Workflow in gdpar (source, R code) Geometric Robustness of Sampling (Block RG) (source, R code) Dependence-Robust Inference in gdpar (Axis 2) (source, R code)

Downloads:

Package source:	gdpar_0.1.0.tar.gz
Windows binaries:	r-devel: not available, r-release: gdpar_0.1.0.zip, r-oldrel: not available
macOS binaries:	r-release (arm64): gdpar_0.1.0.tgz, r-oldrel (arm64): gdpar_0.1.0.tgz, r-release (x86_64): gdpar_0.1.0.tgz, r-oldrel (x86_64): gdpar_0.1.0.tgz

Linking:

Please use the canonical form https://CRAN.R-project.org/package=gdpar to link to this page.