get.internal.valid: Internal Validation of Cox PH model

Description Usage Arguments Details Author(s) Examples

View source: R/get_internal_validation.R

Description

This function allows you to generate a Harrell's C-index of internal validation for a Cox PH model.

Usage

1
get.internal.valid(formula, TR.data, B.num = 200, seed = 2019)

Arguments

formula

a formula used to specify a Cox PH model, see cph function.

TR.data

the name of a data frame containing all needed variables, used as training dataset.

B.num

a number of repetitions for bootstrap. Defaults to 150

seed

a single value, interpreted as an integer, or NULL

Details

Returns C.index, a value of Harrell's C-index. C index is calculated from Dxy.

The relation between Dxy and C-index is Dxy=2*(C-0.5).

Author(s)

Xi "Ada" Wang

Ph.D. Student of Biostatistics

Penn State College of Medicine

xzw149@psu.edu

Examples

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
n <- 500
set.seed(2019)
var1 <- 30 + 5*rnorm(n)
var2 <- factor(sample(c("White","Black"),n,rep=TRUE,prob=c(.3,.7)))
hazard <- .1*exp(.2*(var1-30)+.5*(var2=="Black"))
time_to_event <- -log(runif(n))/hazard
cens <- 15*runif(n)
censor <- ifelse(time_to_event <= cens,1,0)
TR <- data.frame(time_to_event = time_to_event,censor=censor, var1 = var1, var2= var2)
get.internal.valid(formula = Surv(time_to_event,censor) ~ var1 + var2,TR,B.num=2)

ccwangxi/nomovalidation documentation built on Feb. 14, 2020, 9:21 a.m.