Github user srowen commented on the pull request:
https://github.com/apache/spark/pull/11322#issuecomment-188171986
The code looks OK... I think two passes is too much complexity.
I guess the thing is that I've never seen log error used before. That could
just be my ignorance. What's the case where it's the theoretically right thing
to use? squaring the log is throwing me here. I only see it mentioned in the
context of Kaggle, where the motivation is just to penalize predictions that
are low more than high, but this doesn't make it a generally useful metric
beyond ranking submissions.
---
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]
