roc.train: Caret ROC

In decisionpatterns/caret.tools: Utitilies and extensions for the caret package

Description

Create ROC object directly from a caret model

Usage

## S3 method for class 'train'
roc(fit, outcome, ...)

Arguments

fit	caret model fit
outcome	string; outcome label
...	additional arguments passed to roc

Details

This returns an 'roc' object. Which you probably want to then pass to plot.

Value

An object of class 'roc'

See Also

roc
plot

Examples

data(aSAH)
  ctrl <- trainControl( classProb=TRUE, savePrediction=TRUE )
  fit <- train(
               outcome ~ gender + age + s100b + ndka, data=aSAH, method="rf"
               , trControl = ctrl
              )

  roc.train(fit) %>% plot( print.auc=TRUE, grid=TRUE)

decisionpatterns/caret.tools documentation built on May 15, 2019, 1:55 a.m.