DiagTest3Grp (version 1.6)

NonParametric.VUS.var: Boostrap variance of the nonparametric estimate on volumn under ROC surface (VUS)

Description

Calculate the variance of the nonparametric estimate of volumn under ROC surface (VUS) through bootstrapping.

Usage

NonParametric.VUS.var(x, y, z, alpha = 0.05, NBOOT = 100,FisherZ=FALSE)

Arguments

x

A numeric vector, a diagnostic test's measurements in the $D^-$ (usually healthy subjects).

y

A numeric vector, a diagnostic test's measurements in the $D^0$ (usually mildly diseased subjects).

z

A numeric vector, a diagnostic test's measurements in the $D^+$ (usually severely diseased subjects).

alpha

A numeric value, confidence level for confidence interval, will result in (1-alpha)*100%; Default alpha=0.05, i.e., 95% confidence interval.

NBOOT

A numeric value, total number of bootstrap samples to draw for estimate of variance of the VUS nonparametric estimate.

FisherZ

A logic value. Default=FALSE. If TRUE, will transform the nonparametric estimate through Fisher's Z transformation: $\theta^*=1/2log(\frac{1+\theta}{1-\theta})$

Value

variance

A numeric value, the estimate of variance of nonparametric VUS through bootstrapping.

lower

A numeric value, the lower limit of bootstrap basic quantile confidence interval for the estimate of variance of nonparametric VUS, i.e., (alpha/2)*100% quantile. Under default alpha=0.05, lower= basic 0.25% quantile of bootstrap variance estimates of the nonparametric VUS estimator.

upper

A numeric value, the upper limit of bootstrap basic quantile confidence interval for the estimate of variance of nonparametric VUS, i.e., (1-alpha/2)*100% quantile. Under default alpha=0.05, upper= basic 0.975% quantile of bootstrap variance estimates of the nonparametric VUS estimator.

Warning

The bootstrap variance estimate may take a minute.

Details

Return a numeric value, the variance estimation on the nonparametric VUS estimate obtained through bootstrapping.

References

Bradley Efron, Robert Tibshirani, Robert J. Tibshirani (1993) An introduction to the bootstrap CRC Press.

See Also

VUS Normal.VUS Youden3Grp.Variance.Bootstrap

Examples

 data(AL)
 group <- AL$group
 table(group)

 ##take the negated FACTOR1 marker measurements
 factor1 <- -AL$FACTOR1

 x <- factor1[group=="D-"]
 y <- factor1[group=="D0"]
 z <- factor1[group=="D+"]

 nonpar.vus <- NonParametric.VUS.var (x=x, y=y, z=z, alpha=0.05, NBOOT=10)