A Novel Ensemble Credit Scoring Model Based on Extreme Learning Machine and Generalized Fuzzy Soft Sets

Table 2

Characteristics of data sets.

Data sets	Size	Attributes	Good/bad	Imbalance ratio	Number of classifiers
Credit scoring data sets
Germany	1000	25	700/300	2.33	3
Australia	690	14	307/383	0.80	1
Japan	690	15	296/357	0.83	1
Iran	1000	27	950/50	19	19
Bene 1	3,123	33	2,082/1,041	2	2
Bene 2	7,190	33	5,033/2,157	2.33	3
Additional imbalanced data sets
Shuttle	12,380	9	11,428/952	12.004	13
Skin_segment	117,728	3	114,039/3,679	30.99	31
MiniBooNE	201,355	10	196,555/4,800	40.95	41
LC2017Q1	95,633	72	94,414/1,219	77.45	78