Severely imbalanced Big Data challenges: investigating data sampling approaches

Journal of Big Data

Table 6 Case 1: Medicare-Tukey’s HSD test

Learner	Sampling	AUC	std	r	g	Min	Max	Q25	Q50	Q75
(a) AUC
GBT	RUS	0.79833	0.02599	250	a	0.72815	0.87018	0.78092	0.80045	0.81537
	None	0.79047	0.02386	50	a	0.72580	0.83013	0.78059	0.79586	0.80595
	ROS	0.73363	0.06754	250	b	0.51519	0.84819	0.70903	0.74192	0.77947
	SMOTE	0.73031	0.02584	250	b	0.64410	0.81724	0.71385	0.72880	0.74982
	ADASYN	0.69918	0.02609	250	c	0.61985	0.76370	0.68276	0.69946	0.71667
	SMOTEb2	0.67189	0.03213	250	d	0.57265	0.74786	0.65170	0.67248	0.69508
	SMOTEb1	0.66769	0.03720	250	d	0.48250	0.76948	0.64574	0.66988	0.69252
LR	SMOTE	0.82279	0.02125	250	a	0.75783	0.87290	0.81044	0.82237	0.83636
	None	0.81554	0.02227	50	ab	0.75532	0.84700	0.80752	0.81924	0.82659
	ADASYN	0.81509	0.02287	250	ab	0.74781	0.88334	0.80130	0.81666	0.83065
	RUS	0.81169	0.02040	250	b	0.73199	0.86455	0.80016	0.81220	0.82536
	ROS	0.74079	0.06836	250	c	0.55630	0.85671	0.69202	0.75347	0.79658
	SMOTEb1	0.73868	0.02748	250	c	0.66533	0.81191	0.71970	0.73888	0.75844
	SMOTEb2	0.72293	0.03287	250	d	0.61406	0.80044	0.70363	0.72765	0.74389
RF	RUS	0.81195	0.02373	250	a	0.74285	0.86547	0.79696	0.81221	0.82930
	None	0.79383	0.02306	50	b	0.74416	0.83161	0.77569	0.79317	0.81477
	SMOTE	0.77240	0.02304	250	c	0.70450	0.84333	0.75649	0.77252	0.78692
	ROS	0.77042	0.02790	250	c	0.70014	0.85378	0.75188	0.77028	0.78984
	SMOTEb1	0.75732	0.02536	250	d	0.66211	0.81080	0.74231	0.76021	0.77356
	SMOTEb2	0.75383	0.02794	250	d	0.68869	0.82191	0.73425	0.75200	0.77434
	ADASYN	0.73559	0.02654	250	e	0.66474	0.80357	0.71806	0.73933	0.75122

Learner	Sampling	GM	std	r	g	Min	Max	Q25	Q50	Q75
(b) GM
GBT	RUS	0.48872	0.23777	250	a	0	0.78014	0.33953	0.60566	0.68945
	ROS	0.34109	0.25087	250	b	0	0.77439	0.10295	0.35164	0.52541
	SMOTEb1	0.25244	0.13924	250	c	0	0.50945	0.17726	0.27133	0.36454
	SMOTEb2	0.24145	0.13200	250	cd	0	0.49887	0.14570	0.25059	0.33908
	SMOTE	0.22259	0.17815	250	d	0	0.56455	0	0.22840	0.37532
	ADASYN	0.09793	0.12322	250	e	0	0.39311	0	0	0.17759
	None	0.00907	0.03150	50	f	0	0.14509	0	0	0
LR	RUS	0.54028	0.23037	250	a	0	0.77315	0.41864	0.66278	0.71993
	SMOTE	0.53900	0.23596	250	a	0	0.80154	0.42166	0.64136	0.72653
	ADASYN	0.49105	0.25417	250	b	0	0.80124	0.32327	0.59041	0.70910
	ROS	0.48725	0.25459	250	b	0	0.79442	0.33676	0.57923	0.69986
	SMOTEb1	0.43478	0.17227	250	c	0	0.67841	0.35173	0.49188	0.56532
	SMOTEb2	0.42099	0.18294	250	c	0	0.70354	0.32119	0.48474	0.56333
	None	0	0	50	d	0	0	0	0	0
RF	RUS	0.46657	0.24978	250	a	0	0.77101	0.25077	0.57335	0.69443
	SMOTE	0.20508	0.10625	250	b	0	0.43130	0.14498	0.22834	0.28849
	ADASYN	0.14943	0.09303	250	c	0	0.35404	0.10257	0.14575	0.22886
	SMOTEb1	0.11672	0.07686	250	d	0	0.30758	0.10241	0.14469	0.17743
	SMOTEb2	0.09289	0.07056	250	de	0	0.23024	0	0.10260	0.14509
	ROS	0.08824	0.12095	250	e	0	0.39744	0	0	0.14505
	None	0.00823	0.02819	50	f	0	0.10314	0	0	0