selectFGR: Stepwise variable selection in the Fine & Gray regression...

In tagteam/pec: Prediction Error Curves for Risk Prediction Models in Survival Analysis

Stepwise variable selection in the Fine & Gray regression competing risk model

Description

This is a wrapper function which first selects variables in the Fine & Gray regression model using crrstep from the crrstep package and then returns a fitted Fine & Gray regression model with the selected variables.

Usage

selectFGR(formula, data, cause = 1, rule = "AIC", direction = "backward", ...)

Arguments

formula	A formula whose left hand side is a Hist object – see Hist. The right hand side specifies (a linear combination of) the covariates. See examples below.
data	A data.frame in which all the variables of formula can be interpreted.
cause	The failure type of interest. Defaults to 1.
rule	Rule to pass on to crrstep ("AIC", "BIC" or "BICcr"), also see crrstep
direction	see crrstep
...	Further arguments passed to crrstep.

Author(s)

Rob C.M. van Kruijsdijk [e-mail]

Thomas Alexander Gerds [e-mail]

Examples

## Not run: 
library(riskRegression)
library(prodlim)
library(lava)
if (!requireNamespace("cmprsk",quietly=TRUE)){
library(cmprsk)
library(pec)
m <- crModel()
m <- addvar(m,c('X1','X2','X3','X4','X5','X6','X7','X8','X9','X10'))
distribution(m,c("X2","X7","X9")) <- binomial.lvm()
regression(m,eventtime1~X1+X2+X5+X9) <- c(-1,1,0.5,0.8)
set.seed(100)
d <- sim(m,100)
## full formula
ff <- Hist(time, event) ~ X1 + X2 + X3 + X4 +X5 + X6 + X7+ X8 + X9 + X10

# Fit full model with FGR
fg <- FGR(ff,cause=1,data=d)

# Backward selection based on the AIC
sfgAIC <- selectFGR(ff, data=d, rule="AIC", direction="backward") 

sfgAIC$fit # Final FGR-model with selected variables

# Risk reclassification plot at time = 4
plot(predictEventProb(fg,times=4,newdata=d),
     predictEventProb(sfgAIC,times=4,newdata=d))

# Backward selection based on the BIC, while forcing 
# the last two variables (X9 and X10) in the model
sfgBIC <- selectFGR(ff, data=d, rule="BIC", direction="backward", 
                  scope.min=~X9+X10)

## apparent performance
pec(list(full.model=fg,selectedAIC=sfgAIC,selectedBIC=sfgBIC),
    formula=Hist(time, event)~1,
    data=d)


## bootstrap cross-validation performance
set.seed(7)
pec(list(full.model=fg,selectedAIC=sfgAIC,selectedBIC=sfgBIC),
    formula=Hist(time, event)~1,
    data=d,
    B=5,
    splitMethod="bootcv")
}

## End(Not run)

tagteam/pec documentation built on April 25, 2023, 12:03 a.m.