rsf2_fun: random survival forest model with nodesize 5

In SydneyBioX/SurvBenchmark_package: survival approaches benchmarking

random survival forest model with nodesize 5

Description

this is with 100 tress

Usage

rsf2_fun(
  r,
  data,
  cvK,
  fitform_ogl,
  formula1,
  formula2,
  formula3,
  formula4,
  timess
)

Arguments

r	a numeric value, a seed to run this method
data	a dataframe, the data used to performance this survival model
cvK	a numeric value, cross-validation fold
fitform_ogl	a Surv object from package survival, the survival function
formula1	a Surv object from package survival, to calculate a version of the brier score, details please check package pec
formula2	a Surv object from package survival, to calculate a version of the brier score, details please check package pec
formula3	a Surv object from package survival, to calculate a version of the brier score, details please check package pec
formula4	a Surv object from package survival, to calculate a version of the brier score, details please check package pec
timess	a numeric vector of length 15, contains time points to get the time-dependent AUC values

Value

a data.frame with allevaluation measurements in all columns and rows are each fold results from cross-validation

Examples

data("exampledt2", package = "SurvBenchmark")

xnam <- paste(colnames(veteran)[c(1,2,5,6,7,8)], sep="")
form=as.formula(paste("survival::Surv(time, status)~ ", paste(xnam, collapse= "+")))
fitform_ogl=form
formula1=fitform_ogl
formula2=fitform_ogl
formula3=survival::Surv(time,status)~1
formula4=survival::Surv(time,status)~1
timess=seq(as.numeric(summary(veteran$time)[2]),as.numeric(summary(veteran$time)[5]),(as.numeric(summary(veteran$time)[5])-as.numeric(summary(veteran$time)[2]))/14)
want=rsf2_fun(1,veteran,5,fitform_ogl,formula1,formula2,formula3, formula4, timess);

SydneyBioX/SurvBenchmark_package documentation built on June 4, 2022, 12:01 p.m.