Github user trevorcai commented on the pull request:
https://github.com/apache/spark/pull/8609#issuecomment-138480624
Took a brief look - I have a quick high level thought.
The motivation behind the patch is to replace the redundant functionality
of `predictNodeIndex` with the existing `Node.predictImpl` function. Therefore,
I'd suggest overloading the `Node.predictImpl` function with the
`binnedFeatures` and `splits` arguments and roughly mirroring its current
implementation and semantics (just with different arguments provided to
`shouldGoLeft`).
That creates consistent semantics of simulating prediction, instead of the
split implementations that are currently in the master branch.
---
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]
