traineR (version 2.2.11)

ROC.area: ROC.area

Description

Function that calculates the area of the ROC curve of a prediction with only 2 categories.

Usage

ROC.area(prediction, real)

Value

The value of the area(numeric).

Arguments

prediction

A vector of real numbers representing the prediction score of a category.

real

A vector with the real categories of the individuals in the prediction.

See Also

prediction and performance

Examples

iris2 <- dplyr::filter(iris,(Species == "setosa") | (Species == "virginica"))
iris2$Species <- factor(iris2$Species,levels = c("setosa","virginica"))
sam <- sample(1:100,20)
ttesting <- iris2[sam,]
ttraining <- iris2[-sam,]
model <- train.rpart(Species~.,ttraining)
prediction.prob <- predict(model,ttesting, type = "prob")
ROC.area(prediction.prob$prediction[,2],ttesting$Species)