Github user coderxiang commented on the pull request:
https://github.com/apache/spark/pull/458#issuecomment-41178782
@yinxusen thanks for the testing.
- For lasso, the label actually does not matter since it is a regression
model. I'll take care of this in Logistic regression.
- The normalization is a good point. However, maybe it is better to give a
separate procedure for this, since normalization may be required by more than
just Lasso.
For the convergence issue of SGD, I guess more complicated SGD algorithm
(like dual averaging) may solve this problem.
---
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.
---
