Handling Imbalance Classification Virtual Screening Big Data Using Machine Learning Algorithms

Table 2

Complete set of experiments and average G-mean results.

Algorithm		PaDEL numeric descriptor				PaDEL fingerprint
		No-sample	SMOTE	KSMOTE	Time	No-sample	SMOTE	KSMOTE	Time
AID 440	RF	0.167	0.565	0.954	23	0.29	0.442	0.96	12
	DT	0.5	0.59	0.937	9.3	0.51	0.459	0.958	4.9
	MLP	0.6	0.5	0.963	20	0.477	0.498	0.964	9.6
	LG	0.56	0.67	0.963	11	0.413	0.512	0.96	5.6
	GBT	0.23	0.56	0.963	33	0.477	0.421	0.963	17.1
AID624202	RF	0.445	0.625	0.952	29.7	0.5	0.628	0.96	15.3
	DT	0.576	0.614	0.94	10	0.54	0.564	0.94	5
	MLP	0.74	0.715	0.95	25.2	0.636	0.497	0.958	13.5
	LG	0.628	0.83	0.94	26.8	0.791	0.78	0.837	13.25
	GBT	0.489	0.61	0.95	45	0.495	0.482	0.954	22.36
AID 651820	RF	0.722	0.792	0.956	41	0.741	0.798	0.92	19.25
	DT	0.725	0.72	0.932	8.78	0.765	0.743	0.89	4.44
	MLP	0.82	0.817	0.915	35	0.788	0.8	0.91	17.3
	LG	0.779	0.8357	0.962	19	0.75	0.768	0.89	9.36
	GBT	0.714	0.742	0.9	60.5	0.762	0.766	0.905	29.9