Special: Drawing bootstrapped cross-validation curves and the .632 or...
In pec: Prediction Error Curves for Risk Prediction Models in Survival Analysis

Special

R Documentation

Drawing bootstrapped cross-validation curves and the .632 or .632plus error of models. The prediction error for an optional benchmark model can be added together with bootstrapped cross-validation error and apparent errors.

This function is invoked and controlled by plot.pec.

Special(
  x,
  y,
  addprederr,
  models,
  bench,
  benchcol,
  times,
  maxboot,
  bootcol,
  col,
  lty,
  lwd
)

`x`	an object of class 'pec' as returned by the `pec` function.
`y`	Prediction error values.
`addprederr`	Additional prediction errors. The options are bootstrap cross-validation errors or apparent errors.
`models`	One model also specified in `pec` for which the `predErr` in `plot.pec` is to be drawn.
`bench`	A benchmark model (also specified in `pec`) for which the `predErr` in `plot.pec` is to be drawn.
`benchcol`	Color of the benchmark curve.
`times`	Time points at which the curves must be plotted.
`maxboot`	Maximum number of bootstrap curves to be added. Default is all.
`bootcol`	Color of the bootstrapped curves. Default is 'gray77'.
`col`	Color of the different error curves for `models`.
`lty`	Line type of the different error curves for `models`.
`lwd`	Line width of the different error curves for `models`.

This function should not be called directly. The arguments can be specified as Special.arg in the call to plot.pec.

Invisible object.