Table 5.

Comparisons of the average accuracies and standard deviations (%) of EEG-based SAD severity classification by using PDC and graph theory values.

Features	Classifiers	Evaluation metrics	Delta	Theta	Alpha	Beta
GT+PDC features	KNN	Sensitivity	99.52 ± 0.61	98.31 ± 0.52	92.85 ± 0.83	95.30 ± 2.13
		Specificity	97.82 ± 1.77	95.91 ± 4.01	94.32 ± 0.79	99.82 ± 0.79
		Accuracy	82.00 ± 4.40	87.83 ± 2.67	88.54 ± 2.76	84.37 ± 3.58
	LDA	Sensitivity	94.83 ± 1.75	99.66 ± 1.044	99.32 ± 1.38	99.83 ± 0.75
		Specificity	93.00 ± 4.44	95.62 ± 3.56	97.87 ± 3.46	98.61 ± 2.49
		Accuracy	77.00 ± 7.71	77.83 ± 7.26	78.08 ± 6.59	66.58 ± 7.39
	NB	Sensitivity	97.18 ± 1.87	97.12 ± 1.0	85.71 ± 16.07	96.98 ± 2.10
		Specificity	96.28 ± 2.17	96.33 ± 3.01	85.18 ± 1.51	99.23 ± 0.77
		Accuracy	82.70 ± 2.52	82.47 ± 2.07	83.75 ± 1.91	83.10 ± 2.11
	DT	Sensitivity	92.71 ± 7.10	89.30 ± 9.23	93.91 ± 11.12	99.80 ± 0.86
		Specificity	91.28 ± 6.20	81.84 ± 8.09	91.14 ± 4.51	96.34 ± 2.03
		Accuracy	85.95 ± 4.15	89.12 ± 3.68	89.62 ± 2.24	86.00 ± 2.74
	SVM	Sensitivity	91.27 ± 6.20	99.27 ± 0.13	95.25 ± 4.27	91.53 ± 6.47
		Specificity	90.13 ± 6.81	97.23 ± 1.81	94.12 ± 5.91	90.19 ± 9.28
		Accuracy	83.29 ± 5.93	91.66 ± 6.72	92.78 ± 7.86	86.08 ± 10.94