R : how to apply isotonic regression calibration to posterior probability

Question

I want to perform an isotonic regression calibration of posterior probabilities of records. Is there any R package/function for it?

Answer 1

probability.calibration() function of rfUtilities package can be used to perform an isotonic regression calibration of posterior probabilities.

If rfUtilities is not installed, you can install it before running the function.

library(rfUtilities)

p_calibrated = probability.calibration(y, p, regularization = FALSE)

where y = vector of actual labels

and p = vector of posterior probabilities from the model