Consensus Clustering-Based Undersampling Approach to Imbalanced Learning

Table 1

Descriptive information for the datasets [12, 24].

Dataset	Number of data samples	Number of features	Imbalance ratio
Small-scale datasets
Abalone9-18	731	8	16.68
Abalone19	4174	8	128.87
Ecoli-0_vs_1	220	7	1.86
Ecoli-0-1-3-7_vs_2-6	281	7	39.15
Ecoli1	336	7	3.36
Ecoli2	336	7	5.46
Ecoli3	336	7	8.19
Ecoli4	336	7	13.84
Glass0	214	9	3.19
Glass0123vs456	192	9	10.29
Glass016vs2	184	9	19.44
Glass016vs5	214	9	1.82
Glass1	214	9	10.39
Glass2	214	9	15.47
Glass4	214	9	22.81
Glass5	214	9	22.81
Glass6	214	9	6.38
Haberman	306	3	2.68
Iris0	150	4	2
New-thyroid1	215	5	5.14
New-thyroid2	215	5	4.92
Page-blocks0	5472	10	8.77
Page-blocks13vs2	472	10	15.85
Pima	768	8	1.9
Segment	2308	19	6.01
Shuttle0vs4	1829	9	13.87
Shuttle2vs4	129	9	20.5
Vehicle0	846	18	3.23
Vehicle1	846	18	2.52
Vehicle2	846	18	2.52
Vehicle3	846	18	2.52
Vowel0	988	13	10.1
Wisconsin	683	9	1.86
Yeast05679vs4	528	8	9.35
Yeast1	1484	8	2.46
Yeast1vs7	459	8	13.87
Yeast1289vs7	947	8	30.56
Yeast1458vs7	693	8	22.1
Yeast2vs4	514	8	9.08
Yeast2vs8	482	8	23.1
Yeast3	1484	8	8.11
Yeast4	1484	8	28.41
Yeast5	1484	8	32.78
Yeast6	1484	8	39.15
Large-scale datasets
Breast cancer	102294	117	163.19
Protein homology prediction	145751	74	111.46