You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@spark.apache.org by "Oumar Nour (Jira)" <ji...@apache.org> on 2023/10/03 11:41:00 UTC
[jira] [Commented] (SPARK-44848) MLlib GBTClassifier has wrong impurity method 'variance' instead of 'gini' or 'entropy'.
[ https://issues.apache.org/jira/browse/SPARK-44848?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17771433#comment-17771433 ]
Oumar Nour commented on SPARK-44848:
------------------------------------
Hello,
I have the same issue. I want to know if that issue is solved ?
Thanks
> MLlib GBTClassifier has wrong impurity method 'variance' instead of 'gini' or 'entropy'.
> -----------------------------------------------------------------------------------------
>
> Key: SPARK-44848
> URL: https://issues.apache.org/jira/browse/SPARK-44848
> Project: Spark
> Issue Type: Bug
> Components: MLlib
> Affects Versions: 3.4.1
> Reporter: Elisabeth Niederbacher
> Priority: Major
>
> Impurity method 'variance' should only be used for regressors, *not* classifiers. For classifiers gini and entropy should be available as it is already the case for the RandomForestClassifier [https://spark.apache.org/docs/3.1.3/api/python/reference/api/pyspark.ml.classification.RandomForestClassifier.html] .
> Because of this bug 'minInfoGain' hyperparameter cannot be tuned to combat overfitting.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@spark.apache.org
For additional commands, e-mail: issues-help@spark.apache.org