Fast AUC computation in R

Install

The development version of the package can be installed via

if (!require(remotes)) install.packages("remotes")
remotes::install_github("cadam00/lightAUC")

Fast and lightweight computation of AUC metric for the binary case (1 positive and 0 negative) is offered by lightAUC package. The algorithm used is a fast implementation from algorithm of Fawcett (2006).

Example

# Create some data
probs   <- c(1, 0.4, 0.8)
actuals <- c(0, 0, 1)
lightAUC(probs, actuals)

## 0.5

For parallel calculations use:

# E.g. 2 cores (you can use cores = parallel::detectCores() for your case)
probs   <- c(1, 0.4, 0.8)
actuals <- c(0, 0, 1)
lightAUC(probs, actuals, parallel = TRUE, cores = 2)

## 0.5

References

Fawcett, T. (2006). An introduction to ROC analysis. Pattern Recognition Letters, 27(8), 861–874. doi:10.1016/j.patrec.2005.10.010