spark-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Joseph K. Bradley (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (SPARK-3272) Calculate prediction for nodes separately from calculating information gain for splits in decision tree
Date Fri, 29 Aug 2014 01:49:09 GMT

    [ https://issues.apache.org/jira/browse/SPARK-3272?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14114742#comment-14114742
] 

Joseph K. Bradley commented on SPARK-3272:
------------------------------------------

Hi Qiping, you are right; I missed that!  I like your idea of storing the number of instances
in the InformationGainStats.  (That seems easier to understand than a special invalid gain
value.)  For now, I would recommend storing the number for the node, not for the left &
right child nodes.  That would allow you to decide if the node being considered is a leaf
(not its children).

I agree that, eventually, we should identify if the children are leafs at the same time. 
That should be part of [SPARK-3158], which could modify findBestSplits to return ImpurityCalculators
(a new class from my PR [https://github.com/apache/spark/pull/2125]) for the left and right
child nodes.  Does that sound reasonable?

> Calculate prediction for nodes separately from calculating information gain for splits
in decision tree
> -------------------------------------------------------------------------------------------------------
>
>                 Key: SPARK-3272
>                 URL: https://issues.apache.org/jira/browse/SPARK-3272
>             Project: Spark
>          Issue Type: Improvement
>          Components: MLlib
>    Affects Versions: 1.0.2
>            Reporter: Qiping Li
>             Fix For: 1.1.0
>
>
> In current implementation, prediction for a node is calculated along with calculation
of information gain stats for each possible splits. The value to predict for a specific node
is determined, no matter what the splits are.
> To save computation, we can first calculate prediction first and then calculate information
gain stats for each split.
> This is also necessary if we want to support minimum instances per node parameters([SPARK-2207|https://issues.apache.org/jira/browse/SPARK-2207])
because when all splits don't satisfy minimum instances requirement , we don't use information
gain of any splits. There should be a way to get the prediction value.  



--
This message was sent by Atlassian JIRA
(v6.2#6252)

---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@spark.apache.org
For additional commands, e-mail: issues-help@spark.apache.org


Mime
View raw message