[
https://issues.apache.org/jira/browse/SPARK-5099?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14265828#comment-14265828
]
Apache Spark commented on SPARK-5099:
-------------------------------------
User 'viirya' has created a pull request for this issue:
https://github.com/apache/spark/pull/3899
> Simplify logistic loss function and fix deviance loss function
> --------------------------------------------------------------
>
> Key: SPARK-5099
> URL: https://issues.apache.org/jira/browse/SPARK-5099
> Project: Spark
> Issue Type: Bug
> Reporter: Liang-Chi Hsieh
> Priority: Minor
>
> This is a minor pr where I think that we can simply take minus of margin,
> instead of subtracting margin, in the LogisticGradient.
> Mathematically, they are equal. But the modified equation is the common form
> of logistic loss function and so more readable. It also computes more
> accurate value as some quick tests show.
> Besides, there is a bug in computing the loss in LogLoss. This pr fixes it
> too.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
