Edit on GitHub

ROCKS.jl

Two of the commonly used metrics to evaluate the performance of a binary classifier are:

  • C-statistic (concordance statistic) or Area Under Curve or just ROC

(Receiver Operating Characteristic)

  • KS-Statistic (Kolmogorov-Smirnov 2 sample CDF max separation)

This package makes it easy to calculate ROC and KS, and produce graphs and tables for documentation purposes.

Installation

This package is installable via its URL:

] add https://github.com/DaymondLing/ROCKS.jl