[bonsciak]

K means cluster algorithm try to partition data in an r-dimensional feature space into K groups minimizing the average distance in that space. Using the Euclidean Distance as metric the r-dimensional space will be divided by r-1 dimensional hyperplanes. So if I am in a 3-Dimensional Space, the algorithm partition my space with planes. If instead of using the minimization of the Euclidean Distance, I use the Harmonic Mean of the Euclidean Distance, my space is partitioned the same with planes or not?

[sourgrapes11]

You what?

[DFranklin]

Not.

[bonsciak]

ok. but why not?