A Cost-Sensitive Ensemble Method for Class-Imbalanced Datasets

Table 2

Summary of data sets.

Data set	Total samples	no of attributes	no of positive	no of negative	Ratio (majority/minority)
Haberman	306	3	81	225	2.8
Transfusion	926	4	178	748	4.2
Pima	768	8	268	500	1.9
Phoneme	5404	5	1586	3818	2.4
Letter	20000	16	789	19211	24.3