[
https://issues.apache.org/jira/browse/SPARK-14661?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Piotr Suszyński updated SPARK-14661:
------------------------------------
Description: PCA is often used to trim the number of features by using the
minimal variance retained by those features. It would be very convenient to
have a method in PCAModel that would return a trimmed model, especially given
that the PCAModel constructor is private so doing this somewhere in the user
code is not possible. (was: PCA is often used to trim the number of features
by using the minimal variance retained by those features. It would be very
convenient to have a method in PCAModel that would return a trimmed model,
especially given that the PCAModel constructor is private so doing this
somewhere in the user code is not possible.
I already have almost ready code here:
https://github.com/psuszyns/spark/commit/072e9b50c29047700c6482e2c718969d9f6d9753)
> Trim PCAModel by required explained variance
> --------------------------------------------
>
> Key: SPARK-14661
> URL: https://issues.apache.org/jira/browse/SPARK-14661
> Project: Spark
> Issue Type: Improvement
> Components: MLlib
> Affects Versions: 2.0.0
> Reporter: Piotr Suszyński
> Priority: Trivial
>
> PCA is often used to trim the number of features by using the minimal
> variance retained by those features. It would be very convenient to have a
> method in PCAModel that would return a trimmed model, especially given that
> the PCAModel constructor is private so doing this somewhere in the user code
> is not possible.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
