Package: koine 0.3.1

koine: Synthesised Fourth Opinion for Bayesian Triangulation

Synthesises one coherent, independent predictive opinion for a Bayesian model from a curated battery of known and underutilised inference methods, or abstains when no diversity gain exists. Built to serve as a fourth, synthesised witness alongside an independent panel of inference engines: it stacks over model-structure and method diversity in predictive space (leave-one-out stacking, PSIS-gated) on an axis orthogonal to the engine axis, never averaging the engines. Predictive opinions are exchanged through an ArviZ-style 'DataTree' 'NetCDF' interchange so no live 'Python' is required at read time. Abstention is a first-class outcome.

Authors:Max Moldovan [aut, cre]

koine_0.3.1.tar.gz
koine_0.3.1.zip(r-4.7-any)koine_0.3.1.zip(r-4.6-any)koine_0.3.1.zip(r-4.5-any)
koine_0.3.1.tgz(r-4.6-any)koine_0.3.1.tgz(r-4.5-any)
koine_0.3.1.tar.gz(r-4.7-any)koine_0.3.1.tar.gz(r-4.6-any)
koine_0.3.1.tgz(r-4.6-emscripten)
manual.pdf |manual.html✨
DESCRIPTION |NEWS
card.svg |card.png
koine/json (API)

# Install 'koine' in R:

install.packages('koine', repos = c('https://max578.r-universe.dev', 'https://cloud.r-project.org'))

Bug tracker:https://github.com/max578/koine/issues

On CRAN:

3.60 score 41 exports 22 dependencies

Last updated from:f55f0de989 (on main). Checks:9 OK. Indexed: yes.

Target	Result	Time
linux-devel	OK	1171
source / vignettes	OK	232
linux-release	OK	1521
macos-release	OK	1407
macos-oldrel	OK	981
windows-devel	OK	2190
windows-release	OK	1916
windows-oldrel	OK	2151
wasm-release	OK	116

Exports:abstention as_family_spec assemble_loglik evidence_axis family_spec fb_fit fb_fit_diagnostics fit_dram fit_dream fit_glmm_marginal fit_glmm_robust_re fit_horseshoe fit_lasso fit_nimble fit_nimble_glmm fit_pg_gibbs fit_ridge fit_robust_t gate_corroboration gate_psis independence_report is_abstention koine_can_contribute koine_lens_validation_setup koine_opinion_draws koine_pass koine_pass_from_design koine_stage1_registry lens_effective_maturity lens_spec linear_pool_envelope provenance_record read_idata_nc reference_posterior_check resolve_battery sbc_check stack_opinion synthesise validate_fb_fit validate_provenance weight_concentration

Dependencies:abind backports checkmate cli digest distributional generics glue lifecycle loo magrittr matrixStats numDeriv pillar pkgconfig posterior rlang RNetCDF tensorA tibble utf8 vctrs

A second opinion for a mixed model

Rendered fromkoine.Rmdusingknitr::rmarkdown

Last update: 2026-09-02
Started: 2026-05-31

Help page	Topics
Construct a structured abstention	abstention
Adapt a base-R family to an evaluable 'family_spec'	as_family_spec
Assemble per-lens pointwise log-likelihood matrices (Stage-0 SS6.1)	assemble_loglik
Evidence-axis diagnostic (Stage-0 SS9.3)	evidence_axis
Construct an evaluable family specification	family_spec
Construct an 'fb_fit'	fb_fit
Recover convergence diagnostics from an 'fb_fit''s draws	fb_fit_diagnostics
Fit the DRAM lens	fit_dram
Fit the DREAM_ZS lens	fit_dream
Fit the marginal mixed-model lens	fit_glmm_marginal
Fit the heavy-tailed random-effect lens (Student-t random intercepts)	fit_glmm_robust_re
Fit the horseshoe (sparsity) lens	fit_horseshoe
Fit the Bayesian lasso lens	fit_lasso
Fit the NIMBLE-Gibbs lens	fit_nimble
Fit the NIMBLE GLMM lens	fit_nimble_glmm
Fit the Polya-Gamma Gibbs lens (logistic regression)	fit_pg_gibbs
Fit the Bayesian ridge lens	fit_ridge
Fit the robust-t lens (Student-t linear model)	fit_robust_t
The corroboration / maturity gate (Stage-0 SS11)	gate_corroboration
The two-tier gate (Stage-0 SS7) - tier 1 (PSIS reliability)	gate_psis
Build the independence report for the contributing lenses (Stage-0 SS10)	independence_report
Is this object an abstention?	is_abstention
Can koine contribute a diverse opinion for this model? (pre-flight)	koine_can_contribute
Build the validation setup for a koine-authored lens	koine_lens_validation_setup
koine's opinion as a posterior draws array (the stacking mixture)	koine_opinion_draws
Construct a koine 'pass'	koine_pass
Construct a koine 'pass' from design arrays (the IR-targetable builder)	koine_pass_from_design
The Stage-1 enrolment registry (leanest cut)	koine_stage1_registry
Effective maturity tier of a lens after the validation gate	lens_effective_maturity
Construct a lens specification (a registry record)	lens_spec
Linear-pool envelope diagnostic (Stage-0 SS9.2)	linear_pool_envelope
Build a provenance record	provenance_record
Read an ArviZ DataTree NetCDF-4 file into an 'fb_fit'	read_idata_nc
Exact grid-based reference-posterior check (Stage-0 SS11)	reference_posterior_check
Resolve the battery for a pass (family-conditional enrolment, Stage-0 SS3.3)	resolve_battery
Simulation-based calibration of a lens (Stage-0 SS11; Talts et al. 2018)	sbc_check
LOO-stacking of opinion lenses (Stage-0 SS6.2)	stack_opinion
Synthesise the fourth opinion for a pass (Stage-0 SS1 flow)	synthesise
Validate an 'fb_fit' against the field rules	validate_fb_fit
Validate a provenance record's mandatory tuple	validate_provenance
Weight-concentration diagnostic (Stage-0 SS9)	weight_concentration

Package: koine 0.3.1

koine: Synthesised Fourth Opinion for Bayesian Triangulation

Citation

Development and contributors

Readme and manuals

Help Manual

Usage by other packages (reverse dependencies)