Github user jkbradley commented on the pull request:
https://github.com/apache/spark/pull/2607#issuecomment-60820308
@manishamde Thanks in advance for the API simplification!
Also, I'm realizing that this code should be correct for SquaredError but
might not be quite right for the other losses. Looking at Friedman's paper, I
believe that the weak hypothesis weight needs to be adjusted according to the
loss. That calculation is simple for squared error, but it could get
complicated for absolute error and logistic loss (requiring median calculations
and general convex optimization, respectively, I'd guess). I'm OK with leaving
those other losses as long as they are marked with warnings. I believe the
code will still do something reasonable, although not quite ideal.
---
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 infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
