confIntAUCbinorm: Function to compute ROC curve and (bootstrap) confidence...

In biostatUZH: Misc Tools of the Department of Biostatistics, EBPI, University of Zurich

Description

This function computes for values of a continuous variable of a group of cases and a group of controls the ROC curve in the binormal model. Additionally, the AUC including a confidence interval is provided, with bootstrap or with Wald, assuming the variances are equal (the assumption of unequal variances has yet to be implemented).

Usage

confIntAUCbinorm(cases, controls, conf.level = 0.95, replicates =
  1000, grid = 100, ci.method = c("boot", "wald"), var.equal = TRUE)

Arguments

cases	Values of the continuous variable for the cases.
controls	Values of the continuous variable for the controls.
ci.method	Method of calculating the confidence interval. Implemented are boot for bootstrap and wald for the Wald based confidence interval.
conf.level	Confidence level for confidence interval.
replicates	Number of boostrap replicates. Only needed if ci.method = 'boot'.
grid	The interval [0, 1] is split in grid intervals to compute the ROC curve on. Only needed if ci.method = 'boot'.
var.equal	Logical, are the variances assumed to be equal or not (only TRUE implemented). Only needed if ci.method = 'wald'.

Value

The results for ci.method = 'boot' are

a	The values of a.
b	The values of b.
x.val	The values on the x-axis of a ROC plot.
y.val	The values on the y-axis of a ROC plot, i.e. the binormal ROC curve.
auc	Area under the ROC curve.
lowBootCI	Lower limit of bootstrap confidence interval.
upBootCI	Upper limit of bootstrap confidence interval.

The results for ci.method = 'wald' are

a	The values of a.
b	The values of b.
auc	Area under the ROC curve.
lowCI	Lower limit of confidence interval.
upCI	Upper limit of confidence interval.

Note

The Wald approach is documented in a vignette: see vignette("aucbinormal")

Author(s)

Kaspar Rufibach (ci.method = 'boot') and Leonhard Held (ci.method = 'wald')

References

Pepe, M.S. (2003) The statistical evaluation of medical tests for classification and prediction. Oxford: Oxford University Press.

Examples

set.seed(1977)
ns <- c(50, 40)
truth <- c(rep(0, ns[1]), rep(1, ns[2]))
estimates <- c(rnorm(ns[1]), rnorm(ns[2], mean = 0.5, sd = 1.5))
cases <- estimates[truth == 1]
controls <- estimates[truth == 0]
res <- summaryROC(cases, controls, conf.level = 0.95)

## ci.method = 'boot'
## --------------
resBinorm <- confIntAUCbinorm(cases, controls, 
    conf.level = 0.95, replicates = 1000, grid = 100)

# display results
res
resBinorm

## plot ROC curve
## --------------
plot(0, 0, xlim = c(0, 1), ylim = c(0, 1), type = 'l', 
    xlab = "1 - specificity", ylab = "sensitivity", pty = 's')
segments(0, 0, 1, 1, lty = 2)
lines(res$x.val, res$y.val, type = 'l', col = 2, lwd = 2, lty = 2)
lines(resBinorm$x.val, resBinorm$y.val, type = 'l', col = 4, lwd = 2, lty = 2)

## ci.method = 'wald'
## --------------
resBinorm <- confIntAUCbinorm(cases, controls, conf.level = 0.95, ci.method = "wald")

biostatUZH documentation built on May 2, 2019, 6:06 p.m.