ROCR: performance-class – R documentation

Become an expert in R — Interactive courses, Cheat Sheets, certificates and more!

performance-class

Class performance

Object to capture the result of a performance evaluation, optionally collecting evaluations from several cross-validation or bootstrapping runs.

A performance object can capture information from four different evaluation scenarios:

The behaviour of a cutoff-dependent performance measure across the range of all cutoffs (e.g. performance( predObj, 'acc' ) ). Here, x.values contains the cutoffs, y.values the corresponding values of the performance measure, and alpha.values is empty.
The trade-off between two performance measures across the range of all cutoffs (e.g. performance( predObj, 'tpr', 'fpr' ) ). In this case, the cutoffs are stored in alpha.values, while x.values and y.values contain the corresponding values of the two performance measures.
A performance measure that comes along with an obligatory second axis (e.g. performance( predObj, 'ecost' ) ). Here, the measure values are stored in y.values, while the corresponding values of the obligatory axis are stored in x.values, and alpha.values is empty.
A performance measure whose value is just a scalar (e.g. performance( predObj, 'auc' ) ). The value is then stored in y.values, while x.values and alpha.values are empty.

x.name: Performance measure used for the x axis.
y.name: Performance measure used for the y axis.
alpha.name: Name of the unit that is used to create the parametrized curve. Currently, curves can only be parametrized by cutoff, so alpha.name is either none or cutoff.
x.values: A list in which each entry contains the x values of the curve of this particular cross-validation run. x.values[[i]], y.values[[i]], and alpha.values[[i]] correspond to each other.
y.values: A list in which each entry contains the y values of the curve of this particular cross-validation run.
alpha.values: A list in which each entry contains the cutoff values of the curve of this particular cross-validation run.

Objects can be created by using the performance function.

A detailed list of references can be found on the ROCR homepage at http://rocr.bioinf.mpi-sb.mpg.de.