Fig. 17
From: Graph clustering-based discretization of splitting and merging methods (GraphS and GraphM)
The average rankings of: a average number of intervals and b average running times for imbalanced datasets. These rankings obtain from the non-parametric Friedman test and the critical difference (CD) of the Nemenyi post hoc test. Note that, the ranking of number of intervals in a is tested with 100 samples (20 imbalanced datasets and fivefolds), the ranking of running times in b is tested with 20 samples (20 imbalanced datasets)