Skip to content

Instantly share code, notes, and snippets.

@KrisKusano
Created June 24, 2014 16:26
Show Gist options
  • Save KrisKusano/3e5608324b58a5ca21a0 to your computer and use it in GitHub Desktop.
Save KrisKusano/3e5608324b58a5ca21a0 to your computer and use it in GitHub Desktop.
compute AUC for Kaggle Competition results
function auc = AUC(pred, act)
%% compute the AUC for Kaggle competitions
% auc = AUC(pred, act) returns the AUC value on [0, 1]. pred is the predicted
% class membership, {0, 1}, and act is the actual class membership, also {0, 1}
%
% Sources:
% http://link.springer.com/article/10.1023%2FA%3A1010920819831
% http://www.kaggle.com/c/SemiSupervisedFeatureLearning/forums/t/919/auc-implementation/6130
%
% Kristofer D. Kusano - 6/24/14
% check
assert(all(size(pred) == size(act)),...
'size of pred and act must be equal')
assert(range(pred) == 1,...
'pred must be 0 or 1')
assert(range(act) == 1,...
'act must be 0 or 1')
% compute AUC
r = tiedrank(pred);
auc = (sum(r(act == 1)) - sum(act == 1)*(sum(act == 1) + 1) / 2) / ....
(sum(act == 1) * sum(act == 0));
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment