[
https://issues.apache.org/jira/browse/MADLIB-996?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15649317#comment-15649317
]
Rahul Iyer commented on MADLIB-996:
-----------------------------------
There are some open questions on how to add the interface to cross validation.
1. Which parameters should we allow for CV?
The obvious parameters seem to be alpha and lambda. Are there other optimizer
parameters that need to be validated?
2. How should we accept the grid of points for the CV params?
a. We override the existing lambda and alpha parameter inputs to allow
float8[]. This is a clunky solution since we'll have to implement multiple
versions of the training function.
b. We add an additional 'cv_params' string that includes the values for lambda
and alpha as a list. The problem with this solution is that the float8 values
provided as part of the mandatory parameters will become useless.
c. We add the above params but club them with optimizer_params (similar to SVM)
> 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)
