[jira] [Commented] (SPARK-21340) Bring PySpark MLLib evaluation metrics to parity with Scala API
[ https://issues.apache.org/jira/browse/SPARK-21340?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16089487#comment-16089487 ] Marco Gaido commented on SPARK-21340: - [~jake.charland] I submitted a PR but I am not sure it will be merged. If not, please use the new ml package. > Bring PySpark MLLib evaluation metrics to parity with Scala API > --- > > Key: SPARK-21340 > URL: https://issues.apache.org/jira/browse/SPARK-21340 > Project: Spark > Issue Type: Improvement > Components: MLlib >Affects Versions: 2.1.1 >Reporter: Jake Charland > > This JIRA is a request to bring in PySparks MLLib evaluation metrics to > parity with the Scala API. For example in BinaryClassificationMetrics there > are only two eval metrics exposed to pyspark, areaUnderROC and areaUnderPR > while scala has support for a much wider set of eval metrics including > precision recall curves and the ability to set thresholds for recall and > precision values. These evaluation metrics are critical for understanding and > seeing the performance of trained models and should be available to those > using the pyspak api's. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Commented] (SPARK-21340) Bring PySpark MLLib evaluation metrics to parity with Scala API
[ https://issues.apache.org/jira/browse/SPARK-21340?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16085729#comment-16085729 ] Apache Spark commented on SPARK-21340: -- User 'mgaido91' has created a pull request for this issue: https://github.com/apache/spark/pull/18622 > Bring PySpark MLLib evaluation metrics to parity with Scala API > --- > > Key: SPARK-21340 > URL: https://issues.apache.org/jira/browse/SPARK-21340 > Project: Spark > Issue Type: Improvement > Components: MLlib >Affects Versions: 2.1.1 >Reporter: Jake Charland > > This JIRA is a request to bring in PySparks MLLib evaluation metrics to > parity with the Scala API. For example in BinaryClassificationMetrics there > are only two eval metrics exposed to pyspark, areaUnderROC and areaUnderPR > while scala has support for a much wider set of eval metrics including > precision recall curves and the ability to set thresholds for recall and > precision values. These evaluation metrics are critical for understanding and > seeing the performance of trained models and should be available to those > using the pyspak api's. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org