Github user mengxr commented on the pull request:
https://github.com/apache/spark/pull/1484#issuecomment-51438840
1) https://github.com/apache/spark/pull/1814 separates model from algorithm.
2) It is just an idea such that a feature transformer doesn't need to worry
about other data in the instance.
3) It won't work because both transform methods have the same type
signature. This is also related to 2). We want to apply the filter on the
vector and we don't care other data, labeled or not. I'm now working on the API
design to make feature transformation easier. Let's discuss more after v1.1.
Btw, I will re-visit the discretization PR after v1.1 to make sure it
doesn't have performance issues.
---
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]
