Prediction function for SVM wrapper — predict.surv.svm • SuperSurv

Obtains predicted survivals from a fitted surv.svm object. Applies the learned Cox calibration to the new raw utility scores.

Usage

# S3 method for class 'surv.svm'
predict(object, newdata, new.times, ...)

Arguments

object: Fitted surv.svm object.
newdata: New covariate data.frame.
new.times: Times at which to obtain predicted survivals.
...: Additional ignored arguments.

Value

A numeric matrix of predicted survival probabilities, where rows correspond to the observations in newdata and columns correspond to the evaluation times in new.times.

Examples

if (requireNamespace("survivalsvm", quietly = TRUE) &&
  requireNamespace("quadprog", quietly = TRUE)) {
  data("metabric", package = "SuperSurv")
  dat <- metabric[1:25, ]
  x_cols <- grep("^x", names(dat))[1:3]
  X <- dat[, x_cols, drop = FALSE]
  newX <- X[1:5, , drop = FALSE]
  times <- seq(50, 150, by = 50)

  fit <- surv.svm(
    time = dat$duration,
    event = dat$event,
    X = X,
    newdata = newX,
    new.times = times
  )

  pred <- predict(fit$fit, newdata = newX, new.times = times)
  dim(pred)
}
#> [1] 5 3