Table 3

ROC

true positives

correctly identified true edges

tp

false positives

spurious edges

fp

true negatives

correctly identified zero edges

tn

false negatives

unrecognized true edges

fn

positives

all true edges

p = tp + fn

negatives

all zero edges

n = tn + fp

false positive rate

part of negatives set positive

fpr = fp/n

true positive rate

part of positives set positive

tpr = tp/p

false negative rate

part of positives set negative

fnr = fn/p

true negative rate

part of negatives set negative

tnr = tn/n

recall (sensitivity)

true positive rate

tpr

specificity

true negative rate

tnr

precision

positive predictive value

tp/(tp + fp)


Summary of important quantities in ROC analysis.

Hempel et al. BMC Bioinformatics 2011 12:292   doi:10.1186/1471-2105-12-292

Open Data