commons-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Phil Steitz (JIRA)" <>
Subject [jira] [Commented] (MATH-917) More distance measurements are needed in o.a.c.m.stat.clustering.
Date Sat, 23 Mar 2013 17:51:15 GMT


Phil Steitz commented on MATH-917:

In general, I like this. It may be best to do this as a replacement in 4.0, though, instead
of yet another new package, moved for convenience.  My opinion is that this belongs where
it is - in .stat. I agree with the rationale for the changes though, especially the second
bullet. I have a couple of small comments:
* I think I agree, but want to make sure there is no loss in the first bullet.  Doing this
effectively turns a Clusterable into a vector.  I don't have immediate use cases where this
will not be true; but the existing setup is a bit more general and more logical to me - the
distance measure is a property of the space of objects that are being clustered. I get your
point about it making it a bit easier to implement the interface; but I want to make sure
it does not end up cutting out some applications.
* I am curious why you chose to make CentroidBasedCluster and ClusterCenter inner classes
of KMeansPlusPlusClusterer.
* Don't forget the javadoc of the modified Cluster::Cluster() should omit reference to the
> More distance measurements are needed in o.a.c.m.stat.clustering.
> -----------------------------------------------------------------
>                 Key: MATH-917
>                 URL:
>             Project: Commons Math
>          Issue Type: Improvement
>            Reporter: Reid Hochstedler
>             Fix For: 4.0
>         Attachments:
> 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:

View raw message