You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@mahout.apache.org by "Drew Farris (JIRA)" <ji...@apache.org> on 2010/06/14 05:25:13 UTC
[jira] Created: (MAHOUT-415) Lucene filter for Collocations
Lucene filter for Collocations
------------------------------
Key: MAHOUT-415
URL: https://issues.apache.org/jira/browse/MAHOUT-415
Project: Mahout
Issue Type: New Feature
Affects Versions: 0.3
Reporter: Drew Farris
Assignee: Drew Farris
Collocations generated using Mahout could be used to form a whitelist of terms to index into a Lucene index. This patch will provide a way to generate a serialized BloomFilter from CollocationsOutput and a Lucene filter that will take a BloomFilter and emit tokens that are members of that filter. This would allow a set of interesting collocations to be pre-computed for a corpus and then allow the documents to be indexed using only those collocations.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (MAHOUT-415) Lucene filter for Collocations
Posted by "Drew Farris (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/MAHOUT-415?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Drew Farris updated MAHOUT-415:
-------------------------------
Attachment: MAHOUT-415.patch
Here's a start: the Lucene TokenFilter implementation that takes a o.a.hadoop.util.bloom.Filter (BloomFilter) class and optionally keeps or tosses tokens that are members of the filter.
I doubt the package name is appropriate or even if something like this belongs in mahout, but I didn't spot it elsewhere.
> Lucene filter for Collocations
> ------------------------------
>
> Key: MAHOUT-415
> URL: https://issues.apache.org/jira/browse/MAHOUT-415
> Project: Mahout
> Issue Type: New Feature
> Affects Versions: 0.3
> Reporter: Drew Farris
> Assignee: Drew Farris
> Attachments: MAHOUT-415.patch
>
>
> Collocations generated using Mahout could be used to form a whitelist of terms to index into a Lucene index. This patch will provide a way to generate a serialized BloomFilter from CollocationsOutput and a Lucene filter that will take a BloomFilter and emit tokens that are members of that filter. This would allow a set of interesting collocations to be pre-computed for a corpus and then allow the documents to be indexed using only those collocations.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (MAHOUT-415) Lucene filter for Collocations
Posted by "Jake Mannix (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/MAHOUT-415?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12878465#action_12878465 ]
Jake Mannix commented on MAHOUT-415:
------------------------------------
Awesome, you rock, Drew! Very cool.
> Lucene filter for Collocations
> ------------------------------
>
> Key: MAHOUT-415
> URL: https://issues.apache.org/jira/browse/MAHOUT-415
> Project: Mahout
> Issue Type: New Feature
> Affects Versions: 0.3
> Reporter: Drew Farris
> Assignee: Drew Farris
> Attachments: MAHOUT-415.patch
>
>
> Collocations generated using Mahout could be used to form a whitelist of terms to index into a Lucene index. This patch will provide a way to generate a serialized BloomFilter from CollocationsOutput and a Lucene filter that will take a BloomFilter and emit tokens that are members of that filter. This would allow a set of interesting collocations to be pre-computed for a corpus and then allow the documents to be indexed using only those collocations.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.