Determines the number of boosting steps for a survival model fitted by CoxBoost via integrated prediction error curve (IPEC) estimates, conforming to the calling convention required by argument complexity in peperr call.
complexity.ipec.CoxBoost(response, x, boot.n.c = 10, boost.steps = 100,
eval.times = NULL, smooth = FALSE, full.data, ...)complexity.ripec.CoxBoost(response, x, boot.n.c = 10, boost.steps = 100,
eval.times = NULL, smooth = FALSE, full.data, ...)
a survival object (with Surv(time, status)).
n*p matrix of covariates.
number of bootstrap samples.
maximum number of boosting steps, i.e. number of boosting steps is selected out of interval (1, boost.steps).
vector of evaluation time points.
logical. Shall prediction error curve be smoothed by local polynomial regression before integration?
Data frame containing response and covariates of the full data set.
additional arguments passed to CoxBoost call.
Scalar value giving the number of boosting steps.
Plotting the .632+ estimator for each time point given in eval.times results in a prediction error curve. A summary measure can be obtained by integrating over time. complexity.ripec.CoxBoost computes a Riemann integral, while complexity.ipec.CoxBoost uses a Lebesgue-like integral taking Kaplan-Meier estimates as weights. The number of boosting steps of the interval (0, boost.steps), for which the minimal IPEC is obtained, is returned.
peperr, CoxBoost