[jira] [Commented] (SPARK-9137) Unified label verification for Predictor

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

    [ https://issues.apache.org/jira/browse/SPARK-9137?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14631890#comment-14631890 ] 

Joseph K. Bradley commented on SPARK-9137:
------------------------------------------

More notes: Changing title to be for Classifier only since we don't really need to check it for regression.
Also, just noting that this should be as lightweight as possible, happening in a UDF or map so that it can be pipelined without causing an extra RDD action.

> Unified label verification for Predictor
> ----------------------------------------
>
>                 Key: SPARK-9137
>                 URL: https://issues.apache.org/jira/browse/SPARK-9137
>             Project: Spark
>          Issue Type: Improvement
>          Components: ML
>            Reporter: Yanbo Liang
>            Assignee: Yanbo Liang
>
> We should to check label valid before training model for ml.predictor such as LogisticRegression, NaiveBayes, etc. We can make this check at extractLabeledPoints. Some models do this check during training step at present and we need to unified them.



--
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