Table 3.

Best Classifying Machine-Learning Models Using the Variables MT1A, Na, and Mg, When Comparing AMD Patients and Control Subjects

Method	CA	AUC	Sensitivity	Specificity	Precision
Random forest	0.63	0.73	0.82	0.41	0.61
Naive Bayes	0.73	0.86	0.58	0.90	0.87
k-Nearest Neighbors	0.54	0.59	0.39	0.71	0.60
Classification tree	0.68	0.70	0.69	0.67	0.70
Support vector machine	0.64	0.71	0.77	0.50	0.73

AUC, area under the curve; CA, correct assignment.