Fig. 8.

AUPR scores obtained via 5-fold nested cross-validation on the DBHuman dataset for the S1 (random split), S2 (orphan proteins in test set), S3 (orphan molecules in test set), and S4 (double orphan) settings. The performances of $CN\text{-}feaMLP$ are compared to the reference shallow methods and to FNN and CN as reference methods for deep learning. For each setting, the order from left to right in which the results are displayed is indicated in legend