[Octave-bug-tracker] [bug #47871] optimized kmeans.m for speed and modified convergence criterion

[Lachlan Andrew]

Follow-up Comment #1, bug #47871 (project octave):

Thanks, cuantiar.

Note that kmeans produces a decreasing sequence of costs (ignoring numerical
errors).  Have you observed cases where  err  was a large negative value?  I
think that the  abs()  is just an unnecessary function call.

The patch at bug #43959 already avoids the double call to obj_cost.

On a quick reading, I think it also loops over centroids instead of over data
sets.


This is the second patch within a few weeks that tries to improve on kmeans.m,
with changes that seem to be already in my patch from about a year ago.  Could
the package maintainer please apply my patch (or point out remaining flaws),
to save all of this duplicated effort?  Thanks!

    _______________________________________________________

Reply to this item at:

  <http://savannah.gnu.org/bugs/?47871>

_______________________________________________
  Message sent via/by Savannah
  http://savannah.gnu.org/