Improving chemical similarity ensemble approach in target prediction

Table 4 At significance level 0.05, the test result of different SEA models. The numbers after “Multi-voting” denote each voting scheme, e.g. Mult-voting (3) is a 3-vote scheme

	Accuracy	Precision	Sensitivity	Specificity	\(F_{0.5}{\text{-}}Measure\)	\(F_{0.25}{\text{-}}Measure\)
Atom pair	0.692	0.817	0.432	0.916	0.694	0.777
MACCS	0.682	0.802	0.417	0.911	0.677	0.76
Morgan	0.676	0.826	0.382	0.93	0.67	0.773
Topological	0.682	0.837	0.39	0.934	0.681	0.784
Pharmacophore	0.667	0.752	0.42	0.88	0.65	0.719
Multi-voting (1)	0.681	0.71	0.529	0.813	0.664	0.696
Multi-voting (2)	0.688	0.797	0.44	0.903	0.686	0.761
Multi-voting (3)	0.684	0.837	0.396	0.933	0.684	0.786
Multi-voting (4)	0.675	0.864	0.356	0.952	0.672	0.797
Multi-voting (5)	0.669	0.906	0.32	0.971	0.663	0.817

ISSN: 1758-2946