Table 3.

Average performance of 30 partitions for base classifiers (LDA, LR and KNN) and for the classifiers’ ensemble.

		KNN	LDA	LR	Ensemble
>Accuracy (%)	Train	92.86 ± 2.41	96.01 ± 1.50	100 ± 0.00	99.62 ± 0.59
	Validation	85.82 ± 3.82	90.03 ± 2.53	89.78 ± 3.29	92.61 ± 2.48
	Test	82.92 ± 2.90	88.77 ± 3.89	87.42 ± 4.00	91.64 ± 3.20
>F1-score (%)	Train	93.26 ± 2.20	96.14 ± 1.43	100.00 ± 0.00	99.63 ± 0.58
	Validation	87.26 ± 3.21	90.61 ± 2.23	89.55 ± 3.49	92.97 ± 2.27
	Test	84.63 ± 2.76	89.34 ± 3.50	86.87 ± 4.53	92.04 ± 2.97
>Sensitivity (%)	Train	98.43 ± 1.86	98.99 ± 1.29	100.00 ± 0.00	99.87 ± 0.48
	Validation	96.48 ± 2.88	95.79 ± 3.04	87.86 ± 5.07	97.36 ± 2.25
	Test	94.21 ± 5.00	93.58 ± 3.63	84.03 ± 6.73	96.23 ± 3.17
>Specificity (%)	Train	87.30 ± 4.05	93.02 ± 2.81	100.00 ± 0.00	99.37 ± 1.14
	Validation	75.16 ± 7.40	84.28 ± 5.48	91.70 ± 4.22	87.86 ± 4.34
	Test	71.64 ± 4.58	83.96 ± 6.59	90.82 ± 3.67	87.04 ± 5.47
>PPV (%)	Train	88.67 ± 3.28	93.48 ± 2.49	100.00 ± 0.00	99.39 ± 1.11
	Validation	79.83 ± 5.19	86.12 ± 3.95	91.51 ± 4.08	89.04±3.49
	Test	76.95 ± 2.76	85.63 ± 5.02	90.2 ± 3.58	88.33 ± 4.44
>NPV (%)	Train	98.25 ± 2.09	98.95 ± 1.33	100.00 ± 0.00	99.88 ± 0.47
	Validation	95.64 ± 3.32	95.41 ± 3.18	88.50 ± 4.28	97.13 ± 2.40
	Test	92.92 ± 5.42	92.99 ± 3.78	85.33 ± 5.28	95.94 ± 3.35
>AUC (%)	Train	98.49 ± 0.83	99.30 ± 0.56	100.00 ± 0.00	100.00 ± 0.00
	Validation	92.16 ± 2.37	94.72 ± 2.10	93.03 ± 2.74	98.63 ± 0.85
	Test	90.20 ± 2.41	94.72 ± 2.54	91.44 ± 2.63	98.13 ± 1.26