[
https://issues.apache.org/jira/browse/MAHOUT-911?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Robin Anil resolved MAHOUT-911.
-------------------------------
Resolution: Fixed
Fix Version/s: 0.7
Assignee: Robin Anil
There is a new naive bayes implementation which has compact models based on
mahout vectors, should be able to scale to wikipedia. The old version is
deleted in the new release. I am marking this as resolved. Please reopen if you
find any issues with that
> Naive Bayes trains models that are too large to apply
> -----------------------------------------------------
>
> Key: MAHOUT-911
> URL: https://issues.apache.org/jira/browse/MAHOUT-911
> Project: Mahout
> Issue Type: Bug
> Components: Classification
> Affects Versions: 0.6
> Reporter: tom pierce
> Assignee: Robin Anil
> Fix For: 0.7
>
> Attachments: example.wiki.categories.txt
>
>
> I'm seeing the same issue that Lyall Morrison mentioned on the user list not
> too long ago; I can train a model that apparently has too many classes (or is
> otherwise too large) to read back in and apply to new documents.
> I was able to duplicate this issue using the Wikipedia classification
> example. I used an expanded set of categories (125, which is well over the
> 30-some that caused trouble for Lyall).
> I'll attach the list of categories I used.
--
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
