TABLE IV.

Performance Metrics (Weighted Average) Computed Using Only All RNFL Features for Different Classifiers

Classifier	TPR	FPR	F- Measure	MCC	AUPRC	Sensitivity (80% specificity)	Sensitivity (90% specificity)	AUROC (95% CI)
Bayesian Net	0.77	0.21	0.78	0.54	0.82	0.76	0.68	0.84 [0.81,0.87]
Lazy K Star	0.80	0.21	0.80	0.58	0.88	0.80	0.73	0.88 [0.86,0.91]
Meta Classification Regression	0.80	0.22	0.80	0.56	0.85	0.78	0.65	0.86 [0.83,0.89]
Meta Ensemble Selection	0.80	0.23	0.80	0.57	0.85	0.80	0.68	0.86 [0.83,0.89]
Alternating Decision Tree	0.79	0.24	0.79	0.55	0.85	0.77	0.71	0.87 [0.84,0.90]
Random Forest Tree	0.79	0.26	0.79	0.54	0.85	0.82	0.74	0.87 [0.84,0.90]
Classification And Regression Tree (CART)	0.79	0.23	0.79	0.55	0.81	0.77	0.59	0.84 [0.81,0.87]