You are viewing a plain text version of this content. The canonical link for it is here.

Posted to issues@spark.apache.org by "Joseph Peng (JIRA)" <ji...@apache.org> on 2017/10/26 04:56:00 UTC

[jira] [Commented] (SPARK-9579) Improve Word2Vec unit tests

    [ https://issues.apache.org/jira/browse/SPARK-9579?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16219968#comment-16219968 ] 

Joseph Peng commented on SPARK-9579:
------------------------------------

Hi, I am interested in working on this.

> Improve Word2Vec unit tests
> ---------------------------
>
>                 Key: SPARK-9579
>                 URL: https://issues.apache.org/jira/browse/SPARK-9579
>             Project: Spark
>          Issue Type: Test
>          Components: MLlib
>            Reporter: Joseph K. Bradley
>            Priority: Minor
>
> Word2Vec unit tests should be improved in a few ways:
> * Test individual components of the algorithm.  This may mean breaking the code into smaller methods which can be tested individually.
> * Test vs another library, if possible.  Following the example of unit tests for LogisticRegression, create robust unit tests making sure the two implementations produce similar results.  This may be too hard to do robustly (and deterministically).  In that case, the first improvement will suffice.



--
This message was sent by Atlassian JIRA
(v6.4.14#64029)

---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@spark.apache.org
For additional commands, e-mail: issues-help@spark.apache.org