A Real-Time Clinical Decision Support System, for Mild Cognitive Impairment Detection, Based on a Hybrid Neural Architecture

Table 2

Results for the CPN systems using different input combinations from wrapper method. AUC: area under the ROC; Acc: accuracy; Sen: sensitivity; Spc: specificity; YOE: years of education.

	AUC (%)	Acc (%)	Sen (%)	Spc (%)
	95.11	86.84	90.00	84.78
	94.42	85.53	90.00	82.61
	92.97	88.16	86.67	89.13
	92.25	82.90	96.67	73.91
	90.98	81.58	86.67	78.26
	90.87	86.84	80.00	91.30
	90.22	86.84	70.00	97.83
	90.22	84.21	73.33	91.30
	88.44	81.58	86.67	78.26
	87.97	88.16	80.00	93.48
	87.75	82.90	70.00	91.30
	86.70	85.53	80.00	89.13
	86.49	80.26	73.33	84.78
	86.23	84.21	80.00	86.96
	86.23	80.26	70.00	86.96
	86.09	80.26	70.00	86.96
	85.58	78.95	76.67	80.44
	85.36	73.68	93.33	60.87
	84.31	75.00	93.33	63.04
	82.17	81.58	70.00	89.13
	80.58	73.68	83.33	67.39
	80.00	80.26	60.00	93.48
	71.92	65.79	76.67	58.70
	66.88	60.53	83.33	45.65
	65.69	65.79	56.67	71.74
	63.41	53.95	90.00	30.44