Github user MLnick commented on the pull request:
https://github.com/apache/spark/pull/12577#issuecomment-213289938
@sethah you're correct, this is a a bit of a band-aid fix. However, the
real fix is getting CrossValidator to handle cases like this in a principled
and generic way (and/or to change the behavior of ALS in predicting missing
users). But even fixing ALS for missing users, I think the issue will still
arise for missing items.
As for the "ideal" fix in CrossValidator, it seems to be from your JIRA
comment that this will be fairly complex. So until we can fix that, users can
not use ALS in cross-validation in many cases.
I've kept it an expertParam and tried to highlight one should only use it
when you know what you're doing. I think we could also deprecate this option
once the "real" fix comes in CrossValidator...
@srowen @jkbradley ?
---
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]
