[jira] [Commented] (SPARK-17744) Parity check between the ml and mllib test suites for NB

["Joseph K. Bradley (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/SPARK-17744?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15544557#comment-15544557 ] 

Joseph K. Bradley commented on SPARK-17744:
-------------------------------------------

I support copying the unit tests over from mllib to ml.  I'd also be happy to see some of them improved where needed since the mllib unit tests vary in quality.

I don't think we need to make mllib be a wrapper for ml.  It's OK with me if there's a good reason for it, but I don't see a great need to do that extra work.  It seems fine for further improvements to happen in ml only and not mllib, especially since we have limited development resources and lots to work on.

> Parity check between the ml and mllib test suites for NB
> --------------------------------------------------------
>
>                 Key: SPARK-17744
>                 URL: https://issues.apache.org/jira/browse/SPARK-17744
>             Project: Spark
>          Issue Type: Improvement
>          Components: ML
>            Reporter: zhengruifeng
>            Priority: Minor
>
> We have moving {{NaiveBayes}} implementation from mllib to ml package in SPARK-14077. After that, we should do parity check between the ml and mllib test suites, and complement missing test cases for ml, since we may delete spark.mllib packages in the future.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

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