Github user pralabhkumar commented on the issue:
https://github.com/apache/spark/pull/18118
@sethah
Thanks for reviewing the pull request .
- Change the title to obey the proper format [SPARK-20199][ML] ...
- Response : Done
- Change title to reflect that both GBTClassifier and GBTRegressor are
changed
- Response : Done
- Please remove all the text you did not write from the PR description
- Response : Done
- Add a test to check that the default values are correct for
GBTClassifier/Regressor. See the test in logistic regression titled: "logistic
regression: default params" for reference
- Response : Done
- I'd like to test that this change takes effect. One way might be to
construct a small dataset where one feature is highly predictive and other
features are less so, train with featureSubsetStrategy = "all" and with
featureSubsetStrategy = "1" and they should not produce the same tree. I'm open
to other, simpler ways to test it if you can think of some.
- Response : Added test case to check for featureSubsetStrategy parameter.
Creating two GBT trees ,one with subset strategy "all" and other with "1" .
Comparing their most important feature and important features vector to make
sure tree are different
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
