[
https://issues.apache.org/jira/browse/SPARK-4887?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14252130#comment-14252130
]
Apache Spark commented on SPARK-4887:
-------------------------------------
User 'dbtsai' has created a pull request for this issue:
https://github.com/apache/spark/pull/3735
> Fix a bad unittest in LogisticRegressionSuite
> ---------------------------------------------
>
> Key: SPARK-4887
> URL: https://issues.apache.org/jira/browse/SPARK-4887
> Project: Spark
> Issue Type: Bug
> Components: MLlib
> Reporter: DB Tsai
>
> The original test doesn't make sense since if you step in, the lossSum
> already NaN, and the coefficients are diverging. That's because the step size
> is too large for SGD, so it doesn't work. The correct behavior is that you
> should get smaller coefficients than the one without regularization.
> Comparing the values using 20000.0 relative error doesn't make sense as well.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
