[
https://issues.apache.org/jira/browse/MAHOUT-1045?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13414785#comment-13414785
]
Jeff Eastman commented on MAHOUT-1045:
--------------------------------------
Well, the dot product is correct. There are 43 common elements between the two
vectors: Their distance is indeed 0. What is happening is that, since there are
representative points that are not *equal to* the cluster center, the
invalidCluster filter is letting them into the intra-cluster calculation. The
fact that they are all 0-distance from the center is not taken into account by
the filter and thus the max & min values are - correctly - zero too. This
results in the divide by zero and NaN is a correct intra-density.
Once a NaN intra-cluster density is computed, it forces the rest of the average
density calculation to return NaN too. I don't know what to do at this point.
I'm betting this kind of pattern is at work in the CDbw calculation too.
> Cluster evaluators returning bad results
> ----------------------------------------
>
> Key: MAHOUT-1045
> URL: https://issues.apache.org/jira/browse/MAHOUT-1045
> Project: Mahout
> Issue Type: Bug
> Components: Clustering
> Affects Versions: 0.6, 0.7, 0.8
> Environment: Several environments and data sets
> Reporter: Pat Ferrel
> Fix For: 0.8
>
> Attachments: MAHOUT-1045.patch, first-time-density-nan.txt
>
>
> With real world crawl data the Intra-cluster density from ClusterEvaluator is
> almost always NaN. The CDbw inter-cluster density is almost always 0. I have
> also seen several cases where CDbw fails to return any results but have not
> tracked down why yet.
> I have sent a link to an 8G data set that reproduces these errors to Jeff
> Eastman.
--
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
