Calculate the statistics for receiver operating characteristic curve

This function was adapted from Raffel (https://github.com/joyofdata): https://github.com/joyofdata/joyofdata-articles/blob/master/roc-auc/calculate_roc.R

Usage

calculate_roc(predictions, observed, cost_of_fp = 1, cost_of_fn = 1, n = 100)

# S3 method for class 'roc'
summary(object, digits = 3, ...)

# S3 method for class 'roc'
plot(x, curve = "accuracy", ...)

Arguments

predictions

predicted values.

observed

actual observed outcomes.

cost_of_fp

cost of a false positive.

cost_of_fn

cost of a false negative.

n

the number of steps to estimate.

object

result of [calculate_roc()].

digits

number of digits to print.

...

currently unused.

x

result of [calculate_roc()].

curve

values can be cost, accuracy, or NULL.

Value

a ggplot2 expression.