This is an artificial data set with 2 classes (the minority and the majority class) and 600 examples randomly and uniformly distributed in the two-dimensional space (both attributes were real-valued). The minority class is decomposed into 3 elliptic subregions of varying cardinalities, where two subregions are located close to each other, and the remaining smaller sub-region is separated. The borders of sub-regions in the minority class are disturbed increasing the ratio of borderline examples from the minority class subregions (the disturbance ratio) in 30%.