You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@mahout.apache.org by "Yoonmin Nam (JIRA)" <ji...@apache.org> on 2014/01/22 06:35:19 UTC

[jira] [Comment Edited] (MAHOUT-1401) Resurrect Frequent Pattern mining

    [ https://issues.apache.org/jira/browse/MAHOUT-1401?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13878282#comment-13878282 ] 

Yoonmin Nam edited comment on MAHOUT-1401 at 1/22/14 5:34 AM:
--------------------------------------------------------------

I mean that the very first, but crucial problem of current implementation of FPM algorithm is the explosion of intermediate data when generating frequent patterns.
Above explosion makes FPM not available in many cases even we use very small input.
That problem comes from the shortage of intermediate buffer of MapReduce, but as we consider the FPM in the algorithm-level, so we should find the another alternatives either to avoid or handle that problem I mentioned. 


was (Author: ronymin):
I mean that the very first, but crucial problem of current implementation of FPM algorithm is the explosion of intermediate data when generating frequent patterns.
Above explosion makes FPM not available in many cases even we use very small input.
That problem comes from the shortage of intermediate buffer of MapReduce, but as we consider the FPM in the algorithm-level, so we should find the another alternatives either avoid or handle that problem I mentioned. 

> Resurrect Frequent Pattern mining
> ---------------------------------
>
>                 Key: MAHOUT-1401
>                 URL: https://issues.apache.org/jira/browse/MAHOUT-1401
>             Project: Mahout
>          Issue Type: Bug
>            Reporter: Suneel Marthi
>            Assignee: Suneel Marthi
>            Priority: Critical
>




--
This message was sent by Atlassian JIRA
(v6.1.5#6160)