Table 5.

Five-fold cross-validation performance on original and on balanced BTD-Combo dataset

Tool	Imbalanced			Balanced via Majority Class Undersampling			Balanced via Minority Class Oversampling
	Sensitivity	Specificity	MCC	Sensitivity	Specificity	MCC	Sensitivity	Specificity	MCC
Local-DPP	0.4541	0.9119	0.4234	0.7148	0.7364	0.4319	0.5283	0.8756	0.4345
DNABP	0.1034	0.7637	−0.1591	0.3136	0.5017	−0.1757	0.1716	0.6631	−0.1733
LSTM-CNN_Fusion	0.4809	0.8766	0.3935	0.6626	0.7659	0.4209	0.7409	0.6956	0.415