TABLE 7.

Classification performance corresponding to different methods.

Method	Accuracy	Sensitivity	Specificity	AUC
Connection (C)	73.44	64.06	82.81	0.9058
Global (G)	56.25	57.81	54.69	0.7242
Nodal (N)	78.91	75.00	82.81	0.8354
C + G + N	83.59	75.00	92.19	0.9571

C + G + N methods are significantly superior to Connection, Global, and Nodal under 95% confidence interval with p-value equals to 2 $\times {10}^{-5}$ , 6 $\times {10}^{-8}$ and 0.045 respectively.