Warning: This documentation is for scikits.learn version 0.7.1. — Latest stable version

Contents

6.9.3. scikits.learn.metrics.roc_curve

scikits.learn.metrics.roc_curve(y, probas_)

compute Receiver operating characteristic (ROC)

Parameters :

y : array, shape = [n_samples]

true targets

probas_ : array, shape = [n_samples]

estimated probabilities

Returns :

fpr : array, shape = [n]

False Positive Rates

tpr : array, shape = [n]

True Positive Rates

thresholds : array, shape = [n]

Thresholds on proba_ used to compute fpr and tpr

References

http://en.wikipedia.org/wiki/Receiver_operating_characteristic