Table 7.

DNN model performance in all folds for all test datasets.

Dataset	Folds	Accuracy (%)	Precision (%)	Recall or Sensitivity (%)	Specificity (%)	F1-Score (%)	AUC (%)	MCC (%)
First dataset [2]	Fold 1	88.57	93.75	90.00	85.00	91.84	85.50	72.98
	Fold 2	87.14	83.33	94.59	78.79	88.61	88.10	74.77
	Fold 3	95.65	94.23	100	85.00	97.03	97.10	89.50
	Fold 4	97.10	96.15	100	89.47	98.04	98.10	92.75
	Average	92.11	91.86	96.14	84.56	93.88	92.20	82.50
Second dataset (OSR Dataset) [6]	Fold 1	91.63	89.00	93.68	89.81	91.28	91.6	83.34
	Fold 2	92.61	89.50	95.21	90.37	92.27	92.60	85.36
	Fold 3	94.09	94.38	92.31	95.54	93.33	94.10	88.04
	Fold 4	94.33	95.51	91.89	96.38	93.66	94.50	88.60
	Average	93.16	92.09	93.27	93.02	92.63	93.20	86.33
Third dataset [9]	Fold 1	92.67	52.94	75.00	94.20	62.07	84.60	59.22
	Fold 2	92.00	63.16	70.59	94.74	66.67	82.30	62.26
	Fold 3	95.33	84.21	80.00	97.69	82.05	90.20	79.41
	Fold 4	93.33	76.00	82.61	95.28	79.17	86.80	75.30
	Average	93.33	69.07	77.05	95.47	72.49	85.97	69.04