Skip to main content

Table 1 Performance comparison between negative-undersampling and positive-oversampling

From: Utilizing random Forest QSAR models with optimized parameters for target identification and its application to target-fishing server

Sampling method

Negative-undersampling

Positive-oversampling

Overall ROC AUC

0.975

0.956

Top 1 recall

0.534

0.549

Top 4 recall

0.81

0.822

Top 8 recall

0.849

0.855

Top 11 recall

0.86

0.865