[
https://issues.apache.org/jira/browse/MAHOUT-825?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13120773#comment-13120773
]
Sean Owen commented on MAHOUT-825:
----------------------------------
I am not an expert on this code, so take my comments as informed questions
rather than something authoritative.
But I think the question of whether every point participates in, or generates
its own canopy, is separate from whether every point is assigned to a cluster.
Am I right about this? You're right that not every point will contribute to a
canopy, or even be in a canopy. But, those points may still become clustered,
to the canopy to which they are nearest?
I also understand what you're doing with the new patch, and it does seem that
at best this is a selectable feature. But does it make sense to overload the
clusterFilter setting to control this as well? I do get that you think they go
hand-in-hand, and maybe they do, though a separate flag probably still makes
more sense IMHO.
I'm trying to describe the down-side in clustering a distant point. The cost of
emitting the point is trivial. Providing an answer, instead of no answer, can't
be worse, I think. What does it slow down or degrade? I can only come up with
the possibility that it degrades accuracy in iterative algorithms like k-means.
> Canopies grouping records outside t1
> ------------------------------------
>
> Key: MAHOUT-825
> URL: https://issues.apache.org/jira/browse/MAHOUT-825
> Project: Mahout
> Issue Type: Bug
> Components: Clustering
> Affects Versions: 0.6
> Environment: windows, linux
> Reporter: Paritosh Ranjan
> Labels: features, newbie, patch
> Fix For: 0.6
>
> Attachments: canopy-clusterFilter-t1, canopy-outside-t1-points-patch-1
>
>
> While finding closest canopy, there is no check to ensure that it returns
> canopies which are within distance t1 from the point. This results in
> incorrect result i.e. Points outside t1 are grouped in canopies.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators:
https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
