Skip to contents

Changelog

Source: NEWS.md

bayesDP 1.3.8

Bug fixes

  • Fixed bdplm() and bdplogit() producing invalid historical borrowing when covariates were not mean-centered. Because the models use an intercept-free parameterization (separate treatment and control means), uncentered covariates made the arm-mean estimators strongly correlated and inflated their standard errors as extrapolation errors at covariate = 0, corrupting the (diagonal) discount prior. Both functions now automatically mean-center covariates on their pooled (current plus historical) mean and back-transform the reported intercept, so estimates are invariant to covariate location shifts (#1)
  • Fixed a bug where the summary methods for bdpnormal and bdpbinomial read the one-/two-arm indicator from the wrong list element (args instead of args1), causing two-arm fits to be summarised as one-arm
  • Fixed a bug in the two-arm bdpsurvival summary that errored when current control data were absent
  • Fixed an invalid matrix index (Y[, 0]) used when computing the default surv_time in bdpsurvival
  • Fixed the compare argument being silently dropped (passed into paste()) rather than stored in the bdpnormal and bdpbinomial fit objects
  • Fixed the mc discount-weight Z-statistic in bdplm to divide by the standard error rather than the variance
  • Fixed plot methods hanging on the interactive “Hit ” prompt in non-interactive sessions (e.g. tests, CI); par(ask = ...) now respects interactive()
  • Fixed bdplogit() failing during its main model fit because the analysis data passed to MCMClogit() omitted the response variable. The discounted prior precision matrix is now also passed to MCMClogit() correctly (#12).
  • Fixed alpha_discount() so alpha_max is respected when discount_function = "identity" (#6).
  • Fixed bdpnormal() one-arm normal fits with only one source of data for an arm (current-only or historical-only internally) returning an over-dispersed posterior_mu. These branches now return the conjugate posterior of the mean rather than adding an extra observation-level draw (posterior-predictive-like variance) (#15).

Tests

  • Re-enabled the testthat test harness (tests/testthat.R)
  • Rewrote the test suite with expect_*() assertions: augmented one-arm binomial and normal posterior means are pinned against their closed-form conjugate values, and the fixed bugs (one-/two-arm dispatch, stored compare flag, default survival time, two-arm survival summary) are now guarded by tests. Plot calls in tests pass an explicit type so they no longer prompt for input.
  • Expanded test coverage from ~60% to ~76%, adding tests for alpha_discount() and probability_discount() (both now fully covered), the ppexp() vector and matrix paths, the print methods (now fully covered), additional plot branches, input-validation paths, the bdplogit() main fit path, factor-covariate handling in bdplm() and bdplogit(), and the mc discounting method for bdpnormal and bdpbinomial
  • Added regression tests pinning the bdpnormal flat-prior draw of the mean (posterior_flat_mu) and the fixed current-only posterior_mu against their closed-form conjugate (Student-t) variance
  • Expanded method="mc" documentation in the binomial, normal, and survival interfaces/vignettes to note that per-iteration recomputation of the stochastic comparison yields a random alpha_discount sequence that can show noticeable Monte Carlo variability (#4)
  • Guarded the plotting tests with a null graphics device so they no longer write a stray Rplots.pdf

Documentation

  • Clarified the prior_covariate_sd documentation in bdplm() and bdplogit() to note that covariate effects carry an intentional near-zero discount weight, making their priors effectively flat. The supplied value has negligible influence on the posterior, and the effective prior standard deviation is roughly 1e6 larger than the nominal value at the default (#2)

Housekeeping

  • Removed AppVeyor continuous integration
  • Updated GitHub Actions workflows to the latest r-lib/actions examples (actions/checkout@v6, codecov/codecov-action@v6, actions/upload-artifact@v7)
  • Made Codecov upload issues non-fatal in the coverage workflow so tests and coverage generation remain the CI gate while Codecov service/signature failures do not fail the build
  • Added a pkgdown website and accompanying GitHub Actions workflow
  • Replaced deprecated ggplot2::aes_string() with aes() and the .data pronoun in all plot methods
  • Replaced deprecated size aesthetic with linewidth in geom_line() calls
  • De-duplicated the internal model.matrixBayes() helper (previously defined identically in both bdplm and bdplogit) into a single internal file
  • Removed leftover commented-out debugging code
  • Collapsed a redundant conditional in posterior_survival() where both branches initialized identical hazard matrices (#8)
  • Tidied the mc sigma2 sampling in bdplm() and kept the sampling weights aligned with the candidate grid when some marginal log-likelihoods are non-finite (#9)
  • ppexp() now validates its x argument and errors with an informative message when it is neither a numeric vector nor a matrix (#10)
  • Survival curves in the plot and summary methods are now computed with a vectorised C++ routine (ppexpMV) that transposes the hazard matrix once across all time points, instead of looping ppexp() per time point (#11)
  • Avoided recomputing the per-interval sufficient statistics in posterior_survival(); the augmentation step now reuses the values already computed during the discount phase (#7)
  • Removed a redundant useDynLib() directive in the package namespace
  • Added contributor guidance documenting the NEWS.md subsection convention for future releases
  • Expanded README.md with links, supported analyses, examples, and citation guidance (#13)
  • Added the CRAN package URL to DESCRIPTION (#14)
  • Clarified in posterior_normal() that the flat-prior draw of the mean is the conjugate posterior (scale sqrt(sigma^2 / N)), not the posterior predictive
  • Gave each vignette a descriptive title (previously all titled “BayesDP”) and removed unused params/EVAL scaffolding from the vignette headers

bayesDP 1.3.7

CRAN release: 2025-01-12

  • Updated GitHub actions workflows
  • Updated README badges
  • Fixed .Rd file itemize list for bdpbinomial, bdpnormal, and bdpsurvival
  • Fixed logical check in bdpsurvival
  • Fixed spelling mistakes in documentation
  • Minor formatting updates
  • Add reverse dependency checks

bayesDP 1.3.6

CRAN release: 2022-01-30

  • Fixed CRAN CMD warnings for S4 generics

bayesDP 1.3.5

CRAN release: 2021-11-16

  • Fixed CRAN CMD warnings for S4 generics

bayesDP 1.3.4

CRAN release: 2021-01-06

  • Updates to README and DESCRIPTION
  • Updates to .gitignore file
  • Add codecov for coverage checking
  • Updates to R code formatting
  • Added GitHub Actions CI

bayesDP 1.3.3

CRAN release: 2020-02-03

  • New package maintainer (Graeme L. Hickey) since package was orphaned
  • Updates to README, DESCRIPTION, NAMESPACE
  • Added stop break to discount_logit for method = mc

bayesDP 1.3.2

CRAN release: 2018-07-10

  • Minor bdplm vignette typo fixes

bayesDP 1.3.1

CRAN release: 2018-04-11

Major new features

  • Changes to inputs for bdpsurvival
    • Current and (optional) historical data are specified in separate data frames
  • Updated normal approximation used for method = "mc" of the bdpbinomial and bdpnormal functions

Bug fixes and minor improvements

  • Summary method for bdplm now exists and mimics lm
  • Removed bdpbinomial vignette language around success (vs event)
  • Reported one-arm sample size for bdpsurvival print method adjusted to current data only

bayesDP 1.3.0

CRAN release: 2017-12-07

Major new features

  • Addition of the bdplm function for two-arm trials
  • Users can now choose between 3 discount functions via the discount_function input:
    • Weibull CDF
    • Scaled Weibull CDF - scales the Weibull CDF so that the max possible value is 1
    • Identity - sets the discount weight to the posterior probability
  • Removal of bdpregression

Bug fixes and minor improvements

  • Removed two-sided and one-sided function inputs to avoid confusion
  • Posterior probabilities for method = "mc" switched from pshisq to pnorm
  • Updated vignettes to reflect new features

bayesDP 1.2.0

CRAN release: 2017-07-10

Major new features

  • Supports one-arm regression analysis
  • Two additional modular functions
  • Implementation of Monte Carlo-based estimation of alpha discount

Bug fixes and minor improvements

  • Fixes to class slots
  • Added print input to plot method

bayesDP 1.1.0

CRAN release: 2017-05-03

Major new features

  • Supports two-arm survival analysis via hazard rate comparisons
  • Completely revamped summary and print methods to produce better formatted results
  • Plot method allows users to specify a type
  • Added vignettes for each of bdpbinomial, bdpnormal, and bdpsurvival
  • Implemented the fix_alpha input which allows users to set the historical data weight at alpha_max

Bug fixes and minor improvements

  • Fixed error with two-arm analysis where models did not fit if either the current or historical control data were not input
  • Changed two_side input to logical
  • Consolidated several internal functions into a single function for computational efficiency gains

bayesDP 1.0.3

  • README update
  • Added plot types
  • Added Vignettes
  • Added logo
  • Improved documentation
  • Updated print, summary, plot methods
  • Refactored bdpnormal / bdpbinomial

bayesDP 1.0.2

CRAN release: 2017-04-14

  • Crucial bugfixes

bayesDP 1.0.1

CRAN release: 2017-04-01

  • User requested bugfixes

bayesDP 1.0.0

CRAN release: 2017-03-22

  • Initial CRAN release with normal, binomial and survival functions