Table 5.

AUROC and Accuracy ($\pm$ Confidence Interval at 95%) for the Vascular and Structural Parameters before and after Feature Selection for Classifying Healthy Controls and Different Glaucoma Severity Levels (Mild, Moderate, and Severe)

			Before Feature Selection
Parameter	Model	No. of Features	AUROC	Accuracy	Sensitivity	Specificity
Vascular	S-IT SVM	1	0.72 $\pm$ 0.06	0.58 $\pm$ 0.09	0.58 $\pm$ 0.09	0.86 $\pm$ 0.04
	S-IT RF	1	0.70 $\pm$ 0.06	0.55 $\pm$ 0.09	0.55 $\pm$ 0.09	0.84 $\pm$ 0.04
	S-IT xGB	1	0.72 $\pm$ 0.06	0.58 $\pm$ 0.09	0.58 $\pm$ 0.09	0.85 $\pm$ 0.04
	MLS SVM	42	0.67 $\pm$ 0.06	0.51 $\pm$ 0.09	0.51 $\pm$ 0.09	0.83 $\pm$ 0.04
	MLS RF	42	0.72 $\pm$ 0.06	0.60 $\pm$ 0.09	0.60 $\pm$ 0.09	0.85 $\pm$ 0.04
	MLS xGB	42	0.72 $\pm$ 0.06	0.59 $\pm$ 0.09	0.59 $\pm$ 0.09	0.85 $\pm$ 0.04
Structural	RNFL SVM	1	0.60 $\pm$ 0.06	0.40 $\pm$ 0.09	0.40 $\pm$ 0.09	0.79 $\pm$ 0.04
	RNFL RF	1	0.59 $\pm$ 0.06	0.39 $\pm$ 0.09	0.39 $\pm$ 0.09	0.80 $\pm$ 0.03
	RNFL xGB	1	0.67 $\pm$ 0.06	0.53 $\pm$ 0.09	0.53 $\pm$ 0.09	0.82 $\pm$ 0.04
			After Feature Selection
Vascular	MLS SVM	21	0.74 $\pm$ 0.06	0.60 $\pm$ 0.09	0.60 $\pm$ 0.09	0.87 $\pm$ 0.04
	MLS RF	24	0.69 $\pm$ 0.06	0.55 $\pm$ 0.09	0.55 $\pm$ 0.09	0.84 $\pm$ 0.04
	MLS xGB	39	0.76$\pm$ 0.06	0.64 $\pm$ 0.08	0.64 $\pm$ 0.08	0.87 $\pm$ 0.04

The highest mean value per metric is highlighted in bold.