netben: Calculate netbenefit curves of risk predictions

In Atools: Atools

Description

Calculate netbenefit curves of risk predictions

Usage

netben(y.hat, y, Ngrid = NULL)

Arguments

y.hat	numeric. risk between 0 and 1
y	status yes=1, no=0 or dead=1, alive=0
Ngrid	Number of grid points. Defaults to NULL where all distint y.hat are used.

Value

data.frame with columns netben, all and threshold

Author(s)

Andi Boeck

Examples

require(ggplot2)
set.seed(123)
risk <- runif(30, .1, .9);
outcome <- rbinom(30, size=1, prob=risk);
bla <- netben(risk, outcome)

bla2 <- ggplot(bla, aes(x=threshold*100)) +
 geom_hline(aes(yintercept=0, linetype="No biopsies")) +
 geom_line(aes(y=all, linetype="Biopsy all"),alpha=.7, size=.4) +
 geom_line(aes(y=netben, linetype="Biopsy with risk threshold"),alpha=.7, size=.4) +
 xlab("Risk (%)") + ylab("Net benefit") +
 scale_linetype_discrete(guide = guide_legend(title=""))
print(bla2)

Atools documentation built on May 2, 2019, 6:30 p.m.