roc: Draw ROC curve.

In bootfs: Derive Robust Feature Sets for Two or Multiclass Classification Problems

Description

Draw a ROC curve and calculate corresponding AUC.

Usage

roc(pr, labels, measure = "tpr", x.measure = "fpr", colorize = TRUE, 
avg = "none", spread.estimate = "stddev", filter = NULL)

roc_binterval(fitted, labels)

Arguments

pr	Vector holding prediction values.
labels	Vector with the corresponding labels to pr.
measure	Argument from ROCR.
x.measure	Argument from ROCR.
colorize	Argument from ROCR.
avg	Argument from ROCR.
spread.estimate	Argument from ROCR.
filter	Cutoff, when to plot the roc.
fitted	List. Each element holds the prediction vector for one repeat, fold, etc.
labels	List. Each element holds the true class vector for one repeat, fold, etc.

Details

roc: Plotting of ROC curves using the ROCR package.
roc.binterval: Add a Monte Carlo interval for the roc curve, using bootstrap samples out of the prediction/label vectors. Note that roc has to be called before.

Value

The AUC value.

Author(s)

Christian Bender ([e-mail])

See Also

ROCR

Examples

## Not run: 
#Todo.

## End(Not run)

bootfs documentation built on May 2, 2019, 5:50 p.m.