[
https://issues.apache.org/jira/browse/MADLIB-996?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15655698#comment-15655698
]
Frank McQuillan commented on MADLIB-996:
----------------------------------------
[~iyerr3] I just spoke with Woo Jae Jung from the Pivotal Data Science team.
His recommendation is as follows:
1. Lambda would be much higher priority than alpha. Please have a look at:
https://cran.r-project.org/web/packages/glmnet/glmnet.pdf
The R cv.glmnet() function allows the user to not need to supply a list of
potential lambda values. It would be great to investigate how R does this, and
whether it'd make sense to also have this auto-lambda generation functionality
available in MADlib.
User would also specify nfold as an input parameter, but it does not require a
grid search.
2. If we limit ourself to lambda only, then would grid search still be required?
> Add cross validation support to elastic net
> -------------------------------------------
>
> Key: MADLIB-996
> URL: https://issues.apache.org/jira/browse/MADLIB-996
> Project: Apache MADlib
> Issue Type: New Feature
> Components: Module: Regularized Regression
> Reporter: Frank McQuillan
> Fix For: v1.10
>
>
> http://doc.madlib.net/latest/group__grp__elasticnet.html
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
