Flexible Parametric Distribution Survival Learner

Fits a weighted parametric survival model using flexsurv::flexsurvreg(). Native survival probabilities are used without risk-score calibration.

Usage

surv.flexsurvreg(
  time,
  event,
  X,
  newdata = NULL,
  new.times,
  obsWeights = NULL,
  id = NULL,
  dist = "gengamma",
  ...
)

Arguments

time

Observed follow-up time.

event

Observed event indicator.

X

Training covariate data frame.

newdata

Covariate data frame used for prediction.

new.times

Times at which survival probabilities are requested.

obsWeights

Optional non-negative observation weights.

id

Currently ignored.

dist

Distribution: "gengamma" (generalized gamma, default), "gompertz", "gamma", "weibull", "exp", "lnorm", or "llogis".

...

Additional named arguments to flexsurv::flexsurvreg(), such as inits, anc, or optimizer control. Outcome, data, weight, truncation, and relative-survival arguments are managed or excluded by this adapter.

Value

A list with numeric survival matrix pred and fitted object fit.

Details

Only single-event, right-censored outcomes are supported. Covariates enter the distribution's location parameter by default; anc can specify covariates on ancillary parameters. Generalized gamma may require suitable initial values, especially in small training folds. Failed optimization is reported as an error rather than silently accepted.

Examples

if (requireNamespace("flexsurv", quietly = TRUE)) {
  data("metabric", package = "SuperSurv")
  dat <- metabric[1:80, ]
  X <- dat[, "x1", drop = FALSE]
  fit <- surv.flexsurvreg(dat$duration, dat$event, X,
                         X[1:3, , drop = FALSE], c(50, 100), dist = "weibull")
  predict(fit$fit, X[1:3, , drop = FALSE], new.times = c(25, 50, 100))
}
#>           [,1]      [,2]      [,3]
#> [1,] 0.9331806 0.8539393 0.6973283
#> [2,] 0.9479166 0.8850395 0.7566717
#> [3,] 0.9227779 0.8323605 0.6577469