[jira] [Commented] (SPARK-13785) Deprecate model field in ML model summary classes

["Yanbo Liang (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/SPARK-13785?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15196971#comment-15196971 ] 

Yanbo Liang commented on SPARK-13785:
-------------------------------------

[~josephkb] Can I work on this? 
I vote to make model field private from 2.0. Because all summary classes are experimental, users can get model directly and we do not have Python API for summary classes currently. 
Meanwhile, I think we also should not expose the name of the columns in the public API such as featureCol, labelCol. We can remove other unnecessary fields for model summary classes.

> Deprecate model field in ML model summary classes
> -------------------------------------------------
>
>                 Key: SPARK-13785
>                 URL: https://issues.apache.org/jira/browse/SPARK-13785
>             Project: Spark
>          Issue Type: Improvement
>          Components: ML
>            Reporter: Joseph K. Bradley
>            Priority: Minor
>
> ML model summary classes (e.g., LinearRegressionSummary) currently expose a field "model" containing the parent model.  It's weird to have this circular reference, and I don't see a good reason why the summary should expose it (unless I'm forgetting some decision we made before...).
> I'd propose to deprecate that field in 2.0 and to remove it in 2.1.



--
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