[
https://issues.apache.org/jira/browse/MAHOUT-1464?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14029315#comment-14029315
]
ASF GitHub Bot commented on MAHOUT-1464:
----------------------------------------
Github user tdunning commented on the pull request:
https://github.com/apache/mahout/pull/12#issuecomment-45913357
This discussion isn't getting echoed to the mailing list. I didn't even
know it was happening.
I think that a non-zero counter is nice, but it would be better to have a
more general general aggregator of somethings. We have two instances already
of this pattern and there will be more (sum of the abs values is common).
Why not implement a general aggregator? THis is different from our current
aggregateColumns because that function is not parallelizable.
Something like def columnAggregator(combiner, mapper) is what I am aiming
for. Positive counter would be m.columnAggregator(_ + _, _ > 0)
> Cooccurrence Analysis on Spark
> ------------------------------
>
> Key: MAHOUT-1464
> URL: https://issues.apache.org/jira/browse/MAHOUT-1464
> Project: Mahout
> Issue Type: Improvement
> Components: Collaborative Filtering
> Environment: hadoop, spark
> Reporter: Pat Ferrel
> Assignee: Pat Ferrel
> Fix For: 1.0
>
> Attachments: MAHOUT-1464.patch, MAHOUT-1464.patch, MAHOUT-1464.patch,
> MAHOUT-1464.patch, MAHOUT-1464.patch, MAHOUT-1464.patch, run-spark-xrsj.sh
>
>
> Create a version of Cooccurrence Analysis (RowSimilarityJob with LLR) that
> runs on Spark. This should be compatible with Mahout Spark DRM DSL so a DRM
> can be used as input.
> Ideally this would extend to cover MAHOUT-1422. This cross-cooccurrence has
> several applications including cross-action recommendations.
--
This message was sent by Atlassian JIRA
(v6.2#6252)
