A Novel Selective Ensemble Algorithm for Imbalanced Data Classification Based on Exploratory Undersampling

Table 7

Pairwise comparisons of all algorithms based on the AUC criterion.

Algorithms		CART	RUSB	SMOB	UNBag	SMBag	AdaC	RAMO	RotF	Easy	RotE-un	RotEasy
Mean		0.7912	0.8751	0.8870	0.8885	0.8726	0.7234	0.8833	0.8862	0.8958	0.8955	0.9241
CART	𝑟̇		1.1053	1.1191	1.1248	1.1032	0.7894	1.1133	1.1204	1.1327	1.1329	1.1749
	𝑠		17-0-3	19-0-1	19-0-1	18-0-2	13-0-7	17-0-3	20-0-0	20-0-0	20-0-0	20-0-0
	𝑞		0.0015	0.0000	0.0000	0.0002	0.1892	0.0015	0.0000	0.0000	0.0000	0.0000
RUSB	𝑟̇			1.0124	1.0176	0.9981	0.7142	1.0072	1.0136	1.0248	1.0250	1.0629
	𝑠			12-0-8	13-0-7	6-0-14	5-0-15	9-0-11	13-0-7	15-0-5	16-1-3	17-0-3
	𝑞			0.3833	0.1892	0.0784	0.0266	0.6636	0.1892	0.0266	0.0026	0.0015
SMOB	𝑟̇				1.0051	0.9858	0.7054	0.9949	1.0012	1.0122	1.0124	1.0499
	𝑠				11-0-9	5-0-15	2-0-18	9-0-11	9-0-11	13-0-7	12-0-8	16-0-4
	𝑞				0.6636	0.0266	0.0002	0.6636	0.6636	0.1892	0.3833	0.0072
UNBag	𝑟̇					0.9808	0.7019	0.9898	0.9961	1.0071	1.0072	1.0445
	𝑠					6-0-14	8-0-12	10-0-10	10-0-10	15-0-5	14-0-6	19-0-1
	𝑞					0.0784	0.3833	1.0000	1.0000	0.0266	0.0784	0.0000
SMBag	𝑟̇						0.8633	1.0000	1.0108	1.0146	1.0166	1.0497
	𝑠						12-0-8	13-0-7	18-0-2	19-0-1	19-0-1	20-0-0
	𝑞						0.3833	0.1892	0.0002	0.0000	0.0000	0.0000
AdaC	𝑟̇							1.4103	1.4192	1.4349	1.4351	1.4883
	𝑠							17-0-3	16-0-4	17-0-3	18-0-2	19-0-1
	𝑞							0.0015	0.0072	0.0015	0.0002	0.0000
RAMO	𝑟̇								1.0063	1.0174	1.0176	1.0553
	𝑠								10-0-10	15-0-5	13-0-7	15-0-5
	𝑞								1.0000	0.0266	0.1892	0.0266
RotF	𝑟̇									1.0110	1.0112	1.0486
	𝑠									13-0-7	15-0-5	16-1-3
	𝑞									0.1892	0.0266	0.0026
Easy	𝑟̇										1.0002	1.0372
	𝑠										12-0-8	15-1-4
	𝑞										0.3833	0.0118
RotE-un	𝑟̇											1.0370
	𝑠											14-0-6
	𝑞											0.0784