Risk factors for cardiovascular disease in patients with metabolic-associated fatty liver disease: a machine learning approach

Table 3 The performance of the multiple regression classifiers

Cut-point	Top-5		Top-10		Top-15			PCA (10 PCs)
Cut-point	i, ii	iii	i, iii	ii	i	ii	iii	i, ii	iii
Specificity	0.79	0.77	0.73	0.78	0.75	0.74	0.68	0.80	0.63
Sensitivity	0.85	0.87	0.85	0.79	0.87	0.89	0.96	0.77	1.00
CC low-risk [%]	79.17	77.08	72.92	78.47	75.00	74.31	68.06	79.86	62.50
CC high-risk [%]	85.11	87.23	85.11	78.72	87.23	89.36	95.74	76.60	100.0
CC All [%]	80.63	79.58	75.92	78.53	78.01	78.01	74.87	79.06	71.73
PPV [%]	57.14	55.41	50.63	54.41	53.25	53.16	49.45	55.38	46.53
NPV [%]	94.21	94.87	93.75	91.87	94.74	95.54	98.00	91.27	100.0

The best results are boldfaced, whereas the second best are underlined
CC correctly classified, PPV positive predictive value, NPV negative predictive value

ISSN: 1475-2840