[
https://issues.apache.org/jira/browse/SPARK-5436?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14321028#comment-14321028
]
Chris T commented on SPARK-5436:
--------------------------------
What we normally observe is that the error rate assessed against the test data
decreases initially, and then starts to increase as the model begins to
overfit. I think the first step would be to just store/return the error on each
boosting iteration. After the model building is complete, this error rate can
be examined, and a submodel extracted based on where the error in minimum. I
don't think the level of error at the minimum can be known a priori.
> Validate GradientBoostedTrees during training
> ---------------------------------------------
>
> Key: SPARK-5436
> URL: https://issues.apache.org/jira/browse/SPARK-5436
> Project: Spark
> Issue Type: Improvement
> Components: MLlib
> Affects Versions: 1.3.0
> Reporter: Joseph K. Bradley
>
> For Gradient Boosting, it would be valuable to compute test error on a
> separate validation set during training. That way, training could stop early
> based on the test error (or some other metric specified by the user).
--
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
