Github user sethah commented on the pull request:
https://github.com/apache/spark/pull/12577#issuecomment-213137397
From a high level, one concern is that this seems to be a bit of band-aid
fix. If the only scenario where this is a problem is using ALS in cross
validation then it would seem better to address the problem at its root. I am
trying to think of other scenarios where a predictor would output `NaN` and
whether this behavior is desirable in that case. In the
[Jira](https://issues.apache.org/jira/browse/SPARK-14489), @jkbradley mentioned
linear models with too little regularization. Are there any other scenarios
that come to mind? I'll give it some thought.
---
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]
