survSuperLearner.control: Control parameters for the survival Super Learner

In tedwestling/survSuperLearner: Super Learning for Conditional Survival Functions with Right-Censored Data

Control parameters for the survival Super Learner

Description

This function initiates control parameters for the survSuperLearner function.

Usage

survSuperLearner.control(
  initWeightAlg = "survSL.rfsrc",
  initWeight = "censoring",
  max.SL.iter = 20,
  event.t.grid,
  cens.t.grid,
  saveFitLibrary = TRUE
)

Arguments

initWeightAlg	Algorithm to use for the first iteration of the iterative SuperLearner algorithm. Defaults to survSL.rfsrc
initWeight	Whether to start the iterative SuperLearner by fitting censoring weights ("censoring", default) or by fitting event weights ("event").
max.SL.iter	Maximum iterations of the iterative SuperLearner algorithm. Defaults to 20.
event.t.grid	Grid of times to use to approximate the integral in the risk function for the conditional survival function of the event. Defaults to 250 points equally spaced between 0 and the last uncensored follow-up time.
cens.t.grid	Grid of times to use to approximate the integral in the risk function for the conditional censoring survival function. Defaults to 250 points equally spaced between 0 and the last censored follow-up time, minus a small constant in order to approximate left-continuous survivals.
saveFitLibrary	Logical indicating whether to save the fit library on the full data. Defaults to TRUE. If FALSE, cannot obtain predicted values on new data later.

Value

Returns a named list with control parameters.

tedwestling/survSuperLearner documentation built on Dec. 12, 2024, 4:16 p.m.