rocc: ROC based classification

Functions for a classification method based on receiver operating characteristics (ROC). Briefly, features are selected according to their ranked AUC value in the training set. The selected features are merged by the mean value to form a metagene. The samples are ranked by their metagene value and the metagene threshold that has the highest accuracy in splitting the training samples is determined. A new sample is classified by its metagene value relative to the threshold. In the first place, the package is aimed at two class problems in gene expression data, but might also apply to other problems.

Version: 1.2
Depends: ROCR
Published: 2010-10-04
Author: Martin Lauss
Maintainer: Martin Lauss <martin.lauss at med.lu.se>
License: GPL-2 | GPL-3 [expanded from: GPL (≥ 2)]
NeedsCompilation: no
CRAN checks: rocc results

Downloads:

Reference manual: rocc.pdf
Package source: rocc_1.2.tar.gz
Windows binaries: r-devel: rocc_1.2.zip, r-release: rocc_1.2.zip, r-oldrel: rocc_1.2.zip
OS X Snow Leopard binaries: r-release: rocc_1.2.tgz, r-oldrel: rocc_1.2.tgz
OS X Mavericks binaries: r-release: rocc_1.2.tgz
Old sources: rocc archive

Reverse dependencies:

Reverse suggests: fscaret