[jira] [Commented] (SPARK-11223) PySpark CrossValidatorModel does not output metrics for every param in paramGrid
[ https://issues.apache.org/jira/browse/SPARK-11223?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15579923#comment-15579923 ] holdenk commented on SPARK-11223: - Oh wait it looks like we've already done this and I was looking at the wrong tuning model. Going to go ahead and resolve :) > PySpark CrossValidatorModel does not output metrics for every param in > paramGrid > > > Key: SPARK-11223 > URL: https://issues.apache.org/jira/browse/SPARK-11223 > Project: Spark > Issue Type: Improvement > Components: ML, PySpark >Reporter: Raela Wang >Priority: Minor > Labels: starter > -- This message was sent by Atlassian JIRA (v6.3.4#6332) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Commented] (SPARK-11223) PySpark CrossValidatorModel does not output metrics for every param in paramGrid
[ https://issues.apache.org/jira/browse/SPARK-11223?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15579921#comment-15579921 ] holdenk commented on SPARK-11223: - This could be a good starter issue for someone interested in ML or PySpark. You will want to expose some additional getters on the Python CrossValidator model so users can get at the avgMetrics. Feel free to ping me for help if you want :) If no one takes this by December I'll probably give it a shot myself :) > PySpark CrossValidatorModel does not output metrics for every param in > paramGrid > > > Key: SPARK-11223 > URL: https://issues.apache.org/jira/browse/SPARK-11223 > Project: Spark > Issue Type: Improvement > Components: ML, PySpark >Reporter: Raela Wang >Priority: Minor > Labels: starter > -- This message was sent by Atlassian JIRA (v6.3.4#6332) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Commented] (SPARK-11223) PySpark CrossValidatorModel does not output metrics for every param in paramGrid
[ https://issues.apache.org/jira/browse/SPARK-11223?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=14983112#comment-14983112 ] Raela Wang commented on SPARK-11223: It wasn't exactly for debugging purposes. It will be convenient for users to detect overfitting if they are able to access errors for every model created. > PySpark CrossValidatorModel does not output metrics for every param in > paramGrid > > > Key: SPARK-11223 > URL: https://issues.apache.org/jira/browse/SPARK-11223 > Project: Spark > Issue Type: Improvement > Components: ML, PySpark >Reporter: Raela Wang >Priority: Minor > -- This message was sent by Atlassian JIRA (v6.3.4#6332) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Commented] (SPARK-11223) PySpark CrossValidatorModel does not output metrics for every param in paramGrid
[ https://issues.apache.org/jira/browse/SPARK-11223?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=14981720#comment-14981720 ] Kai Sasaki commented on SPARK-11223: Yes. I think so too. But I wonder the purpose of this is only for debugging. If so only printing parameters might be sufficient. > PySpark CrossValidatorModel does not output metrics for every param in > paramGrid > > > Key: SPARK-11223 > URL: https://issues.apache.org/jira/browse/SPARK-11223 > Project: Spark > Issue Type: Improvement > Components: ML, PySpark >Reporter: Raela Wang >Priority: Minor > -- This message was sent by Atlassian JIRA (v6.3.4#6332) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Commented] (SPARK-11223) PySpark CrossValidatorModel does not output metrics for every param in paramGrid
[ https://issues.apache.org/jira/browse/SPARK-11223?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=14975623#comment-14975623 ] Joseph K. Bradley commented on SPARK-11223: --- Do we need to do that? I think we should just provide it in the API as in Scala. > PySpark CrossValidatorModel does not output metrics for every param in > paramGrid > > > Key: SPARK-11223 > URL: https://issues.apache.org/jira/browse/SPARK-11223 > Project: Spark > Issue Type: Improvement > Components: ML, PySpark >Reporter: Raela Wang >Priority: Minor > -- This message was sent by Atlassian JIRA (v6.3.4#6332) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
