rowpAUCs {genefilter}R Documentation

Rowwise ROC and pAUC

Description

Fast rowwise calculation of ROC curves and pAUC.

Usage

rowpAUCs(x, fac, cutpts, p=0.1)

Arguments

x exprSet or numeric matrix. The matrix must not contain NA values.
fac Factor; if x is an exprSet, this may also be a character vector of length 1 with the name of a covariate variable in x. fac must have exactly 2 levels.
cutpts Matrix with same number of rows as x or vector of length greater than 1. It specifies the thresholds for the calculation of the ROC curves. Smaller numbers may further speed up computations. If missing, the ROC curves are calculated between data points.
p Numeric vector of length 1. Limit in (0,1) to integrate pAUC to.

Details

Rowwise calculation of Receiver Operating Characteristic (ROC) curves and the corresponding partial area under the curve (pAUC) for a given data matrix or exprSet. The function is implemented in C and thus reasonably fast and memory efficient.

The definition of the pAUC uses a naive trapezoidal rule which, although less accurate, is faster than more elaborate integrators.

Value

A list with the calculated specificities and sensitivities for each row as matrices, and the corresponding pAUCs.

Author(s)

Florian Hahne <f.hahne@dkfz.de>

References

Pepe MS, Longton G, Anderson GL, Schummer M.: Selecting differentially expressed genes from microarray experiments. Biometrics. 2003 Mar;59(1):133-42.

See Also

rocdemo.sca, pAUC

Examples

  
   data(eset)

   r1 = rowttests(eset, "cov2")
   r2 = rowpAUCs(eset, "cov2")

 if(interactive()) {
   plot(r2$pAUC, r1$statistic, pch=16)
   plot(1-r2$spec[1,], r2$sens[1,], pch=16)
 }

[Package genefilter version 1.8.0 Index]