[
https://issues.apache.org/jira/browse/SPARK-11730?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15021301#comment-15021301
]
Joseph K. Bradley commented on SPARK-11730:
-------------------------------------------
I wrote that note since I did not have time to research what people do for
GBTs. I'd be Ok with matching sklearn's implementation, though it would be
great if we could find academic work indicating a "right" way to handle GBTs.
In particular, I am not sure if trees' contributions should be weighted
differently (based on the learning process) or if they should just use the tree
weights (resembling how prediction works).
> Feature Importance for GBT
> --------------------------
>
> Key: SPARK-11730
> URL: https://issues.apache.org/jira/browse/SPARK-11730
> Project: Spark
> Issue Type: New Feature
> Components: ML, MLlib
> Reporter: Brian Webb
>
> Random Forests have feature importance, but GBT do not. It would be great if
> we can add feature importance to GBT as well. Perhaps the code in Random
> Forests can be refactored to apply to both types of ensembles.
> See https://issues.apache.org/jira/browse/SPARK-5133
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
