Table 3.

Statistical analysis for the performance of the best 9 SSMs on the target Small dataset.

Query Set *	Nodes	% Id	Ac	Correct Class	Incorrect Class	κ	Sn	Sp	Ppos	Pneg
H + sing	467	40	0.917	860	78	0.834	0.838	0.996	0.995	0.86
		50	0.916	859	79	0.832	0.836	0.996	0.995	0.858
		60	0.914	857	81	0.827	0.832	0.996	0.995	0.855
W + sing	469	40	0.92	863	75	0.84	0.844	0.996	0.995	0.865
		50	0.92	863	75	0.84	0.844	0.996	0.995	0.865
		60	0.919	862	76	0.838	0.842	0.996	0.995	0.863
H + W + sing	479	40	0.928	870	68	0.855	0.859	0.996	0.995	0.876
		50	0.928	870	68	0.855	0.859	0.996	0.995	0.876
		60	0.926	869	69	0.853	0.857	0.996	0.995	0.875

* H is the set obtained when harmonic centrality was calculated, W is the set obtained when the weighted degree was calculated, and sing is the set of 99 outliers.