imbDis: imbDis - Imbalanced Discrimination

In ArjunChattoraj10/imbDis: Imbalanced Discrimination

Description

This function creates an object of type imbDis, an S3 Class. imbDis objects can be passed into the associated methods as arguments. Examples of associated methods are auc, brier and logLoss.

Usage

imbDis(labels, pred, case, bins = seq(0.05, 0.5, 0.05))

Arguments

labels	Vector of ground-truth labels.
pred	Vector of predicted probabilities.
case	A value present in the labels argument, that denotes the case class label.
bins	Vector of class imbalance frequencies. Values must be larger than 0 and smaller than 1. Default is seq(0.05, 0.5, 0.05).

Value

A simMetric object - S3 class that contains the input parameters, standardized labels, and sample size.

Examples

# using the mtcars dataset
LR = glm(vs ~ mpg + hp + wt, data = mtcars, family = "binomial")
preds = predict(LR, mtcars, type = "response")  # equiv. to LR$fitted.values

# define the imbDis object
imbD = imbDis(mtcars$vs, preds, 1, seq(0.1, 0.9, 0.1))

# metric results
auc(imbD)
brier(imbD)
logLoss(imbD)

ArjunChattoraj10/imbDis documentation built on Jan. 22, 2021, 3:50 p.m.