[
https://issues.apache.org/jira/browse/MAHOUT-825?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13119425#comment-13119425
]
Paritosh Ranjan commented on MAHOUT-825:
----------------------------------------
I don't see findClosestCanopy() or emitPointToClosestCanopy() being called in
buildClusters() phase as you have mentioned.
(d<T1) will be called while clustering phase. So, if there is a record which is
not inside t1 is assigned to the canopy.
So, when the runClustering variable is true, then the clustered created are
meaning less.
The other place from where it is called is CanopyDriver's clusterDataSeq()
which also demonstrates the same problem. If runClustering is set to true, the
clusters created are incorrect ( I have seen the problem in both sequential and
mapreduce ). I experience this while using CanopyDriver's run method with
clusterFilter > 0 and runClustering=true.
The problem you have explained above will not happen because this method is not
even used in map reduce phase of buildClusters.
You can try CanopyDriver's run method with clusterFilter > 1 and runClustering
= true to reproduce this issue.
Please don't' confuse this issue with "Sequential and MapReduce give different
results". That's a separate problem.
Thanks and Regards,
Paritosh Ranjan
> 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-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
