From: A graphical heuristic for reduction and partitioning of large datasets for scalable supervised training
Procedure
# class 1 points
# class 2 points
SD
None
1743
10,256
4256.5
GSH
1661
2001
170.0