rang_roc_cut: ROC curve optimal cut for a ranger object

In jedwards24/jemodel: Personal package to help with modelling

ROC curve optimal cut for a ranger object

Description

Calls roc_cut() for a fitted ranger model. See roc_cut() for details. Requires probability = TRUE to be used when the ranger model was fitted.

Usage

rang_roc_cut(rf, actual, class_name = NULL, plot = TRUE)

Arguments

rf	A ranger fitted model.
actual	A binary class target vector matching rf.
class_name	String giving name of the class predicted in actual.
plot	Produce a plot. Defaults to TRUE.

Details

The rf$predictions is a matrix with columns for each class. The class corresponding to the predictions in actual can be named in class_name. Default is to select the "maximum" column name i.e. "TRUE" if classes are TRUE/FALSE and the largest number if numeric.

jedwards24/jemodel documentation built on July 1, 2023, 3:59 p.m.