GitHub user holdenk opened a pull request:
https://github.com/apache/spark/pull/12943
[SPARK-15168][PySpark][ML] Add missing params to
MultilayerPerceptronClassifier
## What changes were proposed in this pull request?
MultilayerPerceptronClassifier is missing step size, solver, and weights.
Add these params. Also clarify the scaladoc a bit while we are updating these
params.
Eventually we should follow up and unify the HasSolver params (filed
https://issues.apache.org/jira/browse/SPARK-15169 )
## How was this patch tested?
Doc tests
You can merge this pull request into a Git repository by running:
$ git pull https://github.com/holdenk/spark
SPARK-15168-add-missing-params-to-MultilayerPerceptronClassifier
Alternatively you can review and apply these changes as the patch at:
https://github.com/apache/spark/pull/12943.patch
To close this pull request, make a commit to your master/trunk branch
with (at least) the following in the commit message:
This closes #12943
----
commit d9c0c6c9dd85620815ff5eb11facd3d9cb71865f
Author: Holden Karau <[email protected]>
Date: 2016-05-05T23:40:17Z
Start adding the missing params to mutli-layer perceptron, also investigate
how to handle the weights param being none
commit f763c3c3de12af5c4be996d0fe27fbec32331206
Author: Holden Karau <[email protected]>
Date: 2016-05-05T23:49:09Z
Or wait we just don't need to support None
commit 4b9e7b1f3d0a6359d50d10fc8bffb48844ead9ec
Author: Holden Karau <[email protected]>
Date: 2016-05-06T00:14:36Z
Update the doc string for weights param and add doctest that verifys layers
stay same but weights change
----
---
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]
