[
https://issues.apache.org/jira/browse/SPARK-8845?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14615759#comment-14615759
]
Apache Spark commented on SPARK-8845:
-------------------------------------
User 'dbtsai' has created a pull request for this issue:
https://github.com/apache/spark/pull/7245
> ML use of Breeze optimization: use adjustedValue not value?
> -----------------------------------------------------------
>
> Key: SPARK-8845
> URL: https://issues.apache.org/jira/browse/SPARK-8845
> Project: Spark
> Issue Type: Bug
> Components: ML
> Reporter: Joseph K. Bradley
> Priority: Minor
>
> In LinearRegression and LogisticRegression, we use Breeze's optimizers (LBFGS
> and OWLQN). We check the {{State.value}} to see the current objective.
> However, Breeze's documentation makes it sound like {{value}} and
> {{adjustedValue}} differ for some optimizers, possibly including OWLQN:
> [https://github.com/scalanlp/breeze/blob/26faf622862e8d7a42a401aef601347aac655f2b/math/src/main/scala/breeze/optimize/FirstOrderMinimizer.scala#L36]
> If that is the case, then we should use adjustedValue instead of value. This
> is relevant to [SPARK-8538] and [SPARK-8539], where we will provide the
> objective trace to the user.
> CC: [~dbtsai] Could you please take a look? Thanks!
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
