Learning self-supervised molecular representations for drug–drug interaction prediction

BMC Bioinformatics

Table 1 Performance of different sampling strategies

	AUPRC	AUROC	ACC	F1	Precision	Recall
Task 1
BALANCED_BATCH	0.897	0.991	0.912	0.912	0.913	0.912
CLASS_WEIGHT	0.710	0.971	0.305	0.214	0.398	0.305
PyTorch WEIGTHED RANDOM SAMPLER	0.868	0.990	0.862	0.862	0.866	0.862
Task 2
BALANCED_BATCH	*0.420*	0.900	0.537	0.508	0.525	0.537
CLASS_WEIGHT	0.390	0.886	0.237	0.180	0.433	0.237
PyTorch WEIGHTED RANDOM SAMPLER	0.417	0.878	0.530	0.500	0.519	0.530
Task 3
BALANCED_BATCH	0.170	0.731	0.299	0.264	0.283	0.299
CLASS_WEIGHT	0.205	0.756	0.174	0.161	0.250	0.174
PyTorch WEIGHTED RANDOM SAMPLER	0.175	0.730	0.299	0.255	0.269	0.299

ISSN: 1471-2105