gputools (version 0.26)

getAucEstimate: Estimate the AUC of the ROC

Description

This function gives a quick estimate of the area under the curve (AUC) of the receiver operating characteristic (ROC). It is a quick way to estimate the quality of a binary classifier. The algorithm is based on a paper by David Hand and Robert Till (see references).

Usage

getAucEstimate(classes, scores)

Arguments

classes

a vector of floating point numbers. Each entry i corresponds to the real class of a point and should be either 0 or 1. The negative class is represented by 0 and the positive class by 1. These entries correspond both in number and order to the sam

scores

a vector of floating point numbers. Each entry i corresponds to the probability that a point is in the positive class of a binary classification. This will be the output of, for example, a binary classifier based on logistic regression. These ent

Value

• a single floating point number of double precision. This number represents an estimate of the auc score for the algorithm responsible for the scores vector. The estimation is according to the method of David Hand and Robert Till (see references).

References

Hand, David J. and Till, Robert J. (2001). A simple generalisation of the area under the ROC curve for multiple class classification problems. Machine Learning. 45, 171--186.

Examples

# generate some fake data
classes <- round(runif(20, min = 0, max = 1))

# fake probability that point i is in the positive class
scores <- runif(20, min = 0, max = 1)

b <- getAucEstimate(classes, scores)
print(b)