srowen commented on issue #23751: [SPARK-26852][PYTHON] Add absolute transform to CrossValidator URL: https://github.com/apache/spark/pull/23751#issuecomment-462830575 Wouldn't you just want to return the absolute value of this metric, if your goal is to minimize the absolute value? These are error-like metrics, right -- those are inherently nonnegative. Or else what's an example of a metric that is properly negative sometimes but needs to be close to 0 ideally? i'm not sure I see the use case here.
---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [email protected] With regards, Apache Git Services --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
