Skip to main content

Table 1 Prediction performance comparison with the two baseline methods, namely all-negatives and random-negatives

From: DDI-PULearn: a positive-unlabeled learning method for large-scale prediction of drug-drug interactions

MethodPrecisionRecallF1-score
DDI-PULearn0.9060.8280.865
Random-negatives0.7650.6760.718
All-negatives0.7090.4490.550