Table 3.

Submodel building with ensemble data

	Optimal algorithm	Optimal feature	Accuracy	F1-score
Model_E0	GBDT	18	0.87	0.87
Model_E1	GBDT	19	0.85	0.85
Model_E2	GBDT	17	0.84	0.84
Model_E3	GBDT	20	0.85	0.85
Model_E4	GBDT	20	0.84	0.84
Model_E5	GBDT	14	0.85	0.85
Model_E6	GBDT	17	0.85	0.85
Model_E7	GBDT	17	0.85	0.85
Model_E8	GBDT	14	0.85	0.85
Model_E9	GBDT	10	0.85	0.85
Model_E10	GBDT	20	0.85	0.85
Model_E11	GBDT	21	0.87	0.87
Model_E12	GBDT	16	0.89	0.89
Model_E13	GBDT	16	0.89	0.89
Model_E14	GBDT	16	0.88	0.88