[
https://issues.apache.org/jira/browse/MATH-917?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13611801#comment-13611801
]
Thomas Neidhart commented on MATH-917:
--------------------------------------
Thanks for all your feedback, I only tried to focus on the interface and kept
all other things aside (but will be completely cleaned up before committing).
ad Luc@Clusterable interface: totally agree, this was literally my first idea
without much thinking, the only constraint that we have is to avoid any
conflict with other typical interfaces, imho.
ad Phil@bullet1: well, it's a trade-off, I think the idea to only implement a
distance method in the Clusterable interface has some benefits too, but it also
means that you lose flexibility. Typically, (and I had some experience with
clustering algorithms in the past), you want to experiment with different
distance methods, so this should be a parameter to the clustering algorithm and
not defined by the data objects.
ad Phil@bullet2: just a coincidence, I played around with different solutions,
hiding completely the implementation details of a KMeans clusterer at first,
but later on I thought it might be interesting for some use-cases. That's why I
made it public and also return it directly in the cluster method, but kept it
in the class itself for the moment. If the idea is accepted, this should be a
stand-alone class imho.
ad Phil@bullet3: see above, did not touch the javadoc yet, but wanted to get
feedback about the direction
> More distance measurements are needed in o.a.c.m.stat.clustering.
> -----------------------------------------------------------------
>
> Key: MATH-917
> URL: https://issues.apache.org/jira/browse/MATH-917
> Project: Commons Math
> Issue Type: Improvement
> Reporter: Reid Hochstedler
> Fix For: 4.0
>
> Attachments: clustering.zip
>
>
> Currently only Euclidean distance is used for distance measurement, it would
> be easy to quickly add Manhattan and Chebyshev distance among others.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
