Figure 6.

The receiver operating characteristic (ROC) curve for the drug-target prediction performance of the DNN-based DTI model. True positive rate (y axis) is also known as sensitivity, which refers to the probability of accuracy in the positive dataset. False-positive rate (x axis) refers to the probability of failure in the negative dataset. Area under curve (AUC) value indicates the performance of separability. The higher AUC value means the model has better separability, i.e., the better drug-target prediction for drug targets (biomarkers).