Model selection for high-dimensional Cox models with SCAD penalty

Automatic model selection for high-dimensional Cox models with SCAD penalty, evaluated by penalized partial-likelihood.

Usage

fit_scad(
  x,
  y,
  nfolds = 5L,
  gammas = c(2.01, 2.3, 3.7, 200),
  eps = 1e-04,
  max.iter = 10000L,
  seed = 1001,
  trace = FALSE,
  parallel = FALSE
)

Arguments

x: Data matrix.
y: Response matrix made by Surv.
nfolds: Fold numbers of cross-validation.
gammas: Gammas to tune in cv.ncvsurv.
eps: Convergence threshhold.
max.iter: Maximum number of iterations.
seed: A random seed for cross-validation fold division.
trace: Output the cross-validation parameter tuning progress or not. Default is FALSE.
parallel: Logical. Enable parallel parameter tuning or not, default is FALSE. To enable parallel tuning, load the doParallel package and run registerDoParallel() with the number of CPU cores before calling this function.

Examples

# \donttest{
data("smart")
x <- as.matrix(smart[, -c(1, 2)])
time <- smart$TEVENT
event <- smart$EVENT
y <- survival::Surv(time, event)

fit <- fit_scad(
  x, y,
  nfolds = 3, gammas = c(3.7, 5),
  max.iter = 15000, seed = 1010
)
#> Warning: ncvsurv() is intended for pathwise optimization, not for single values of lambda.
#>   1. You are strongly encouraged to fit a path and extract the solution at the lambda value of interest, rather than use ncvsurv() in this way.
#>   2. In particular, if you are using the MCP or SCAD penalties, be aware that you greatly increase your risk of converging to an inferior local maximum if you do not fit an entire path.
#>   3. You may wish to look at the ncvfit() function, which is intended for non-path (i.e., single-lambda) optimization and allows the user to supply initial values.

nom <- as_nomogram(
  fit, x, time, event,
  pred.at = 365 * 2,
  funlabel = "2-Year Overall Survival Probability"
)

plot(nom)

# }