[
https://issues.apache.org/jira/browse/MAHOUT-418?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12880902#action_12880902
]
Sean Owen commented on MAHOUT-418:
----------------------------------
Let's see how far apart these two implementations are. It would be great to
spend some time unifying them a bit, now, if there is no hurry to get a second
implementation in.
Yes, the recommender-specific job needs an additional phase at the start and
end, to map from longs to ints and back. It does do this. This can remain. But
once data is converted into vectors, the general code you are creating should
be able to take over?
Both implementations can write the whole matrix, and take the same approach to
self-similarity. That is I think you are welcome to make them both assume the
same thing. Just compute and store everything for good measure.
If those are the only differences, it really seems like they are doing the same
thing and this can be a move of code rather than copy. I think you should feel
free to go this way, even if it requires change in other code. I can help
adjust other code if it means some assumptions have changed.
That way you are not burdened with maintaining two implementations. I think
that makes MAHOUT-420 easier.
What do you think, are you keen to commit this, or open to pushing towards one
implementation?
> Computing the pairwise similarities of the rows of a matrix
> -----------------------------------------------------------
>
> Key: MAHOUT-418
> URL: https://issues.apache.org/jira/browse/MAHOUT-418
> Project: Mahout
> Issue Type: New Feature
> Components: Math
> Reporter: Sebastian Schelter
> Attachments: MAHOUT-418-2.patch, MAHOUT-418.patch
>
>
> In response to the wish from MAHOUT-362 and the latest discussion on the
> mailing list started by Kris Jack about computing a document similarity
> matrix, I tried to generalize the approach we're already using to compute the
> item-item-similarities for collaborative filtering.
> The job in the patch computes the pairwise similarity of the rows of a matrix
> in a distributed manner, is uses a SequenceFile<IntWritable,VectorWritable>
> as input and outputs such a file too. Custom similarity implementations can
> be supplied, I've already implemented tanimoto and cosine for demo and
> testing purposes. The algorithm is based on the one presented here:
> http://www.umiacs.umd.edu/~jimmylin/publications/Elsayed_etal_ACL2008_short.pdf
> I'd be glad if someone could verify the applicability of this approach by
> running it with a reasonably large input, I'm also worried that it might
> buffer to much data in certain steps.
> If you decide to include it in mahout, some more efforts and decisions (like
> more tests, more similarity measures, integration with DistributedRowMatrix)
> would need to be made, I guess.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
