{"Package":"spsurv","Type":"Package","Title":"Bernstein Polynomial Based Semiparametric Survival Analysis","Version":"1.1.0","Authors@R":"c(person(given = \"Renato\",\nfamily = \"Panaro\",\nrole = c(\"aut\", \"cre\", \"cph\"),\nemail = \"rvpanaro@gmail.com\",\ncomment = c(ORCID = \"0000-0002-1903-2091\", URL = \"https://rvpanaro.github.io/\")),\nperson(given = \"Fábio\",\nfamily = \"Demarqui\",\nrole = c(\"ths\", \"rev\"),\nemail = \"fndemarqui@gmail.com\",\ncomment = c(ORCID = \"0000-0001-9236-1986\", URL = \"http://www.est.ufmg.br/~fndemarqui/\")),\nperson(given = \"Vinicius\",\nfamily = \"Mayrink\",\nrole = c(\"ths\"),\nemail = \"vdinizm@gmail.com\",\ncomment = c(ORCID = \"0000-0002-5683-8326\", URL = \"http://www.est.ufmg.br/~vdinizm/\")))","Description":"Semiparametric survival analysis based on Bernstein polynomials.\n'spsurv' includes proportional hazards, proportional odds and accelerated\nfailure time frameworks for right-censored data. RV Panaro (2020)\n.","License":"GPL-3","Biarch":"true","Depends":"R (>= 4.1.0), survival (>= 2.44-1.1), coda (>= 0.19-3)","Imports":"methods, Rcpp (>= 0.12.0), RcppParallel (>= 5.0.1), rstan (>=\n2.18.1), generics, viridis","LinkingTo":"BH (>= 1.66.0), Rcpp (>= 0.12.0), RcppEigen (>= 0.3.3.3.0),\nRcppParallel (>= 5.0.1), rstan (>= 2.18.1), StanHeaders (>=\n2.18.0)","Suggests":"testthat (>= 2.1.0), knitr, rmarkdown (>= 2.31), ggplot2 (>=\n3.3.0), KMsurv, parsnip (>= 1.0.0), censored (>= 0.3.0),\nworkflows (>= 1.1.0), tidybayes (>= 3.0.0), posterior (>=\n1.4.0), covr, rsample (>= 1.1.0), rsurv, rstantools (>= 1.5.1)","VignetteBuilder":"knitr","Encoding":"UTF-8","SystemRequirements":"GNU make","BugReports":"https://github.com/rvpanaro/spsurv/issues","URL":"https://github.com/rvpanaro/spsurv,\nhttps://rvpanaro.github.io/spsurv/","Config/roxygen2/version":"8.0.0","Config/rstantools/auto_config":"FALSE","NeedsCompilation":"yes","Packaged":"2026-09-28 07:43:22 UTC; rvpanaro","Author":"Renato Panaro [aut, cre, cph] (ORCID:\n, URL:\nhttps://rvpanaro.github.io/),\nFábio Demarqui [ths, rev] (ORCID:\n, URL:\nhttp://www.est.ufmg.br/~fndemarqui/),\nVinicius Mayrink [ths] (ORCID: ,\nURL: http://www.est.ufmg.br/~vdinizm/)","Maintainer":"Renato Panaro ","Repository":"CRAN","Date/Publication":"2026-09-28 09:00:02 UTC","repoType":"cran","tarballUrl":"ftp://cran.r-project.org/pub/R/src/contrib/spsurv_1.1.0.tar.gz","jsonAuthors":[{"name":"Renato Panaro","email":"rvpanaro@gmail.com","maintainer":true},{"name":"Fábio Demarqui","email":"fndemarqui@gmail.com","maintainer":false},{"name":"Vinicius Mayrink","email":"vdinizm@gmail.com","maintainer":false}],"readme":"## spsurv $\"spsurv$ \n\n[![CRAN status](https://www.r-pkg.org/badges/version/spsurv)](https://cran.r-project.org/package=spsurv) [![CRAN downloads](https://cranlogs.r-pkg.org/badges/spsurv)](https://cran.r-project.org/package=spsurv) [![CRAN downloads](https://cranlogs.r-pkg.org/badges/grand-total/spsurv)](https://cran.r-project.org/package=spsurv) [![R-CMD-check](https://github.com/rvpanaro/spsurv/actions/workflows/R-CMD-check.yaml/badge.svg)](https://github.com/rvpanaro/spsurv/actions/workflows/R-CMD-check.yaml) [![Codecov](https://codecov.io/gh/rvpanaro/spsurv/branch/master/graph/badge.svg)](https://app.codecov.io/gh/rvpanaro/spsurv)\n\n## Overview\n\n**spsurv**: An R package for semi-parametric survival analysis.\n\nThe **spsurv** package was designed to contribute with a flexible set of semi-parametric survival regression options, including **proportional hazards (PH)**, **proportional odds (PO)**, and **accelerated failure time (AFT)** models for right-censored data.\n\nThe package provides:\n\n- Survival classes (PH, PO, AFT) extensions based on a fully likelihood-based approach for either **Bayesian** or **maximum likelihood (ML)** estimation procedures\n- Smooth estimates for the unknown baseline functions based on the **Bernstein polynomial (BP)**\n- Integration with **Stan** for user-defined modeling\n- Six distinct prior specification options in a Bayesian analysis\n\nStan is an open-source platform with its own language and log-probability functions for custom likelihoods and priors. Access to Stan in R is provided via **rstan**; the package uses NUTS (No-U-Turn) sampling by default for Bayesian fits.\n\n## Installation\n\n### From CRAN\n\n```r\ninstall.packages(\"spsurv\")\n```\n\n### Development version\n\n```r\ninstall.packages(\"devtools\")\ndevtools::install_github(\"rvpanaro/spsurv\")\n```\n\n## Usage\n\nFit a BP-based survival regression PH model. If `degree` is omitted, the default\nis `ceiling(sqrt(n))` where `n` is the number of rows in `data`; the resolved\nvalue is stored on the fit as `fit$degree`.\n\n```r\nlibrary(\"KMsurv\")\ndata(\"larynx\")\nlarynx$stage <- factor(larynx$stage)\n\nlibrary(spsurv)\n\nfit <- bpph(Surv(time, delta) ~ age + stage, data = larynx, approach = \"mle\")\nsummary(fit)\nfit$degree # Bernstein polynomial degree used\n```\n\nAlternatively, use the `spbp` function or set the degree with `bernstein(m)`:\n\n```r\nfit <- spbp(Surv(time, delta) ~ age + stage, model = \"ph\", data = larynx,\n dist = bernstein(5), approach = \"mle\")\nsummary(fit)\n```\n\nBayesian analysis with the `approach` argument:\n\n```r\nfit2 <- spbp(Surv(time, delta) ~ age + stage,\n approach = \"bayes\", data = larynx,\n iter = 2000, chains = 1, warmup = 1000)\nsummary(fit2)\n```\n\nNested model comparison (MLE) follows the same `anova()` patterns as\n`survival::survreg`:\n\n```r\nfit0 <- bpph(Surv(time, delta) ~ 1, data = larynx, approach = \"mle\", degree = 5)\nfit1 <- bpph(Surv(time, delta) ~ age, data = larynx, approach = \"mle\", degree = 5)\nfit2 <- bpph(Surv(time, delta) ~ age + stage, data = larynx, approach = \"mle\", degree = 5)\n\nanova(fit0, fit1)\nanova(fit1, fit2)\nanova(fit2) # sequential term-wise table\n```\n\nSee the [reference manual](https://rvpanaro.github.io/spsurv/reference/index.html) and\nvignette `vignette(\"getting-started\", package = \"spsurv\")` for more examples.\n\n### Tidy model summaries\n\nCoefficient and model-level summaries follow the [generics](https://generics.r-lib.org/) / broom convention:\n\n```r\nlibrary(generics)\nlibrary(ggplot2)\n\nfit <- bpph(Surv(time, delta) ~ age + stage, data = larynx, approach = \"mle\")\n\n# One row per term (hazard ratios with 95% CI)\ntd <- tidy(fit, conf.int = TRUE, exponentiate = TRUE)\nprint(td)\n\n# One-row model summary (n, log-likelihood, global LR test, AIC, ...)\nglance(fit)\n\n# Forest plot\nggplot(td, aes(x = estimate, y = term)) +\n geom_point() +\n geom_errorbarh(aes(xmin = conf.low, xmax = conf.high), height = 0.2) +\n geom_vline(xintercept = 1, linetype = \"dashed\") +\n labs(x = \"Hazard ratio\", y = NULL)\n```\n\nFor publication-ready tables, pipe `tidy()` output into [gt](https://gt.rstudio.com/) or similar packages.\n\n### Baseline inference, diagnostics, and model ranking\n\n```r\n# Baseline gamma weights with Wald table (survstan-style summary)\nsummary(fit, show_baseline = TRUE)\nprint(fit, show_baseline = TRUE) # same via print()\nprint(fit, bp.param = TRUE) # baseline-only table\n\n# Tidy baseline weights\ntidy(fit, component = \"baseline\", conf.int = TRUE)\nconfint(fit, parm = names(fit$bp.param))\n\n# Rank nested models by AIC\nf0 <- bpph(Surv(time, delta) ~ 1, data = larynx, degree = 5)\nf1 <- bpph(Surv(time, delta) ~ age, data = larynx, degree = 5)\nrank_models(f0, f1) # or AIC(f0, f1)\n\n# Baseline survival S0(t) at event times\nsurvfit(fit, baseline = TRUE, tidy = TRUE)\n\n# ggplot2 residual plots (requires ggplot2)\nggresiduals(fit, type = \"martingale\")\n```\n\nIf you also use the **survstan** package, qualify spsurv helpers after\n`library(survstan)` — e.g. `spsurv::estimates(fit)` and `spsurv::se(fit)`.\n\n### tidymodels and tidybayes\n\nLoad **parsnip** (or **tidymodels**) before **spsurv** so censored-regression\nengines register. Use `scale = FALSE` when preprocessing with **recipes**.\n\n```r\nlibrary(parsnip)\nlibrary(censored)\nlibrary(workflows)\nlibrary(spsurv)\n\nwf <- workflow() |>\n add_formula(Surv(time, status) ~ karno) |>\n add_model(\n proportional_hazards() |>\n set_engine(\"spsurv\", degree = 5, scale = FALSE, init = 0)\n )\n\nfit_wf <- fit(wf, data = veteran)\npredict(fit_wf, veteran[1:2, ], type = \"survival\", eval_time = 100)\n```\n\nBayesian fits support **tidybayes** via `as_draws_df.spbp()`, `spread_draws()`,\nand `spread_surv_draws.spbp()`. See `vignette(\"tidymodels\", package = \"spsurv\")`.\n\n## Troubleshooting\n\n### Bayesian convergence checks\n\nFor Bayesian fits, always check:\n\n- divergent transitions (target: 0)\n- split-\$\\hat R\$ (target: close to 1, typically < 1.01)\n- effective sample sizes (avoid very low bulk/tail ESS)\n\nIf warnings appear, use this escalation order:\n\n1. increase `adapt_delta` (for example, from 0.8 to 0.9 or 0.95)\n2. increase `iter` and `warmup`\n3. re-check divergences, `Rhat`, and ESS before interpretation\n\nHigher `adapt_delta` usually reduces divergences but increases runtime,\nso tuning should balance stability and compute cost.\n\nPlease report issues at [https://github.com/rvpanaro/spsurv/issues](https://github.com/rvpanaro/spsurv/issues) or contact the maintainer (see DESCRIPTION).\n\n## Links\n\n- [Download from CRAN](https://cran.r-project.org/package=spsurv)\n- [Browse source code](https://github.com/rvpanaro/spsurv)\n- [Report a bug](https://github.com/rvpanaro/spsurv/issues)\n\n## License\n\nGPL-3","jobInfo":{"package":"spsurv","version":"1.1.0","parsingStatus":"success","parserVersion":2,"parsedAt":"2026-09-28T09:30:14+0000"}}

Description

Usage

Arguments