The hardware and bandwidth for this mirror is donated by dogado GmbH, the Webhosting and Full Service-Cloud Provider. Check out our Wordpress Tutorial.
If you wish to report a bug, or if you are interested in having us mirror your free-software or open-source project, please feel free to contact us at mirror[@]dogado.de.
rvpanaro@gmail.com; license is GPL-3 (removed
the bundled LICENSE copy of the GPL);
Authors@R roles reviewed (single maintainer; thesis
advisors listed as ths, with rev for code
review feedback); package Description citation updated to
<doi:10.48550/arXiv.2003.10548>; set
Config/rstantools/auto_config: FALSE and ship committed
Stan C++ exports (no install-time
configure/rstan_config(), avoiding Windows
R-devel rstan.dll load failures); added
RcppParallel and synced src/Makevars.win TBB
flags with Unix Makevars; restored covr,
rsample, and rsurv in Suggests
for optional tests referenced in tests/..spbp_gamma_information_diagnostics() no longer calls
qr() on non-finite gamma information matrices;
vcov()/survfit() return NA uncertainty with a
warning when that block is unavailable.spsurv
(MLE) and spsurv_bayes for
proportional_hazards(), new
proportional_odds(), and survival_reg();
bp_survival_reg() convenience constructor;
workflow / predict(type = "survival")
support.as_draws_df.spbp(),
spread_surv_draws.spbp(), and S3 hooks for
tidy_draws / spread_draws /
gather_draws on Bayes fits; posterior draw indices
(.chain, .iteration, .draw).predict.spbp() types (survival,
time, linear_pred);
augment.spbp().vignette("tidymodels").vcov.spbp() no longer
applies a diagonal ridge (polish removed); beta standard
errors are unchanged when the Bernstein block is stable and no longer
inflated when it is ill-conditioned.summary.spbp() now includes a
coef_interval component (Wald for MLE, HPD for Bayes) and
prints it when show_intervals = TRUE.inst/ to paper/ (excluded from the package
tarball).Suggests to
packages used in vignettes, tests, or optional integrations.
devtools and roxygen2 are not listed in
Suggests.tidy(),
glance(), logLik(), AIC(),
anova(), estimates(), and se()
methods for spbp objects; bernstein() baseline
specification.predict() and improved
survfit() (tidy output, newdata covariates, Bayesian
monotone bands).residuals() supports
martingale, deviance, and Cox-Snell types; martingale diagnostics fixed
in examples.handlers.R;
validation and setup logic inlined in spbp.default.vcov.spbp.transformed parameters
(only alpha stored); pointwise log_lik moved
to generated quantities; likelihood via
bp_pointwise_log_lik() in the model block.u to avoid
boundary bias; minimum feasible divisor (min_range) only
when range is degenerate.spbp objects: Component
degree on fits and in call$degree; default
Bernstein degree ceiling(sqrt(n)) when omitted (documented
in help, README, and vignettes).anova.spbp: Pairwise and sequential
analysis-of-deviance tables aligned with survival::survreg
(Terms, Test, Df,
Deviance, etc.).bpph / bppo /
bpaft: User-facing match.call()
without unevaluated dist/baseline symbols;
optional degree argument.spbp /
spbp.default help (@details, return
components, vignette links); README usage and model-comparison examples
updated.data("veteran", package = "survival") in docs and
examples..gitignore, .Rbuildignore,
LICENSE).These binaries (installable software) and packages are in development.
They may not be fully stable and should be used with caution. We make no claims about them.
Health stats visible at Monitor.