Some facts about the EkP Prototype Selection System
Did you know that with the EkP System one can select exactly k most informative samples from every class?
Thus the size of a target prototype set obtained with running the EkP system can be controlled in advance (i.e. one may decide how many instances to retain).
The EkP system (similarly to the PM-M method) can use an arbitrary underlying classifier to select prototypes unlike most other methods which rely on the k-Nearest Neighbor rule.