ROC plot generic function for a gg_roc object.

ROC plot generic function for a gg_roc object.

Usage

# S3 method for class 'gg_roc'
plot(x, which_outcome = NULL, ...)

Arguments

x

gg_roc object created from a classification forest

which_outcome

for multiclass problems, choose the class for plotting

...

arguments passed to the gg_roc function

Value

ggplot object of the ROC curve

References

Breiman L. (2001). Random forests, Machine Learning, 45:5-32.

Ishwaran H. and Kogalur U.B. (2007). Random survival forests for R, Rnews, 7(2):25-31.

Ishwaran H. and Kogalur U.B. (2013). Random Forests for Survival, Regression and Classification (RF-SRC), R package version 1.4.

See also

gg_roc rfsrc

Examples

## ------------------------------------------------------------
## classification example
## ------------------------------------------------------------
## -------- iris data
# Build a small classification forest (ntree=50 keeps example fast)
set.seed(42)
rfsrc_iris <- rfsrc(Species ~ ., data = iris, ntree = 50)

# ROC for setosa (outcome index 1)
gg_dta <- gg_roc(rfsrc_iris, which_outcome = 1)
plot.gg_roc(gg_dta)


# ROC for versicolor (outcome index 2)
gg_dta <- gg_roc(rfsrc_iris, which_outcome = 2)
plot.gg_roc(gg_dta)


# ROC for virginica (outcome index 3)
gg_dta <- gg_roc(rfsrc_iris, which_outcome = 3)
plot.gg_roc(gg_dta)


# Alternatively, pass the forest directly to plot all three ROC curves
plot.gg_roc(rfsrc_iris)