Handling Imbalance Classification Virtual Screening Big Data Using Machine Learning Algorithms

Table 3

Sensitivity and specificity results of the three datasets in numeric and fingerprint descriptors.

Algorithm		PaDEL numeric descriptor						PaDEL fingerprint
		No-sample		SMOTE		KSMOTE		No-sample		SMOTE		KSMOTE
		Sensitivity	Specificity	Sensitivity	Specificity	Sensitivity	Specificity	Sensitivity	Specificity	Sensitivity	Specificity	Sensitivity	Specificity
AID 440	RF	0.028	0.998	0.32	0.997	0.91	0.997	0.09	0.985	0.2	0.998	0.948	0.99
	DT	0.25	0.992	0.36	0.986	0.9	0.98	0.257	0.99	0.214	0.985	0.93	0.98
	MLP	0.37	0.996	0.26	0.993	0.94	0.993	0.22	0.99	0.25	0.994	0.951	0.99
	LG	0.314	0.998	0.46	0.98	0.94	0.98	0.17	0.998	0.267	0.976	0.94	0.98
	GBT	0.05	0.997	0.32	0.993	0.94	0.991	0.228	0.99	0.17	0.995	0.95	0.98
AID 624202	RF	0.2	0.993	0.39	0.965	0.905	0.993	0.25	0.997	0.4	0.987	0.93	0.99
	DT	0.351	0.958	0.4	0.957	0.92	0.956	0.305	0.946	0.33	0.934	0.924	0.958
	MLP	0.57	0.967	0.53	0.993	0.928	0.96	0.418	0.961	0.24	0.964	0.956	0.971
	LG	0.4	0.807	0.8	0.806	0.921	0.81	0.775	0.987	0.76	0.86	0.825	0.984
	GBT	0.25	0.985	0.38	0.957	0.91	0.985	0.249	0.994	0.23	0.9848	0.924	0.995
AID 651820	RF	0.529	0.983	0.648	0.944	0.94	0.94	0.56	0.98	0.67	0.966	0.9	0.95
	DT	0.57	0.923	0.579	0.876	0.91	0.9	0.65	0.9	0.63	0.896	0.88	0.95
	MLP	0.728	0.95	0.716	0.943	0.9	0.913	0.58	0.93	0.673	0.933	0.9	0.93
	LG	0.62	0.964	0.79	0.856	0.94	0.9	0.59	0.97	0.69	0.876	0.88	0.95
	GBT	0.52	0.97	0.56	0.93	0.89	0.914	0.63	0.972	0.63	0.967	0.91	0.91