CBribiescas opened a new pull request #32813:
URL: https://github.com/apache/spark/pull/32813
### What changes were proposed in this pull request?
This PR disables a feature created in SPARK-3159 where LearningNodes are
merged after a RF model is trained.
### Why are the changes needed?
2 Reasons:
1. In addition to basic classification, another use case for decision trees
are the probabilities associated with predictions.
Once pruned, these predictions are lost and it makes the trees/predictions
challenging to work with if not unusable.
2. It is not in line with the default behavior in sklearn. In sklearn, the
trees are left unpruned by default.
Please see Jira ticket for more explanation.
### Does this PR introduce _any_ user-facing change?
No, it's dev-only.
### How was this patch tested?
I modified the two tests introduced with this change to verify
postive/negative use of feature.
Locally ran `./build/mvn -pl mllib package` and verified tests passed
Additionally, running through git workflow as described here:
https://spark.apache.org/developer-tools.html#github-workflow-tests
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
