You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@mahout.apache.org by "Isabel Drost (JIRA)" <ji...@apache.org> on 2008/04/08 16:11:24 UTC

[jira] Created: (MAHOUT-28) EM implementation

EM implementation
-----------------

                 Key: MAHOUT-28
                 URL: https://issues.apache.org/jira/browse/MAHOUT-28
             Project: Mahout
          Issue Type: New Feature
          Components: Classification
            Reporter: Isabel Drost


One quarter of the work that could be done for Mahout-4. EM seems a bit to general a concept to be implemented only for the use case of PLSI, so the task of this issue is to find a way to implement EM general enough to be reused. As EM conceptually is related to k-Means it might be possible to make the implementation general enough to include k-Means as a special case.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


[jira] Commented: (MAHOUT-28) EM implementation

Posted by "Ted Dunning (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/MAHOUT-28?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12779269#action_12779269 ] 

Ted Dunning commented on MAHOUT-28:
-----------------------------------


Seems right to me.

> EM implementation
> -----------------
>
>                 Key: MAHOUT-28
>                 URL: https://issues.apache.org/jira/browse/MAHOUT-28
>             Project: Mahout
>          Issue Type: New Feature
>          Components: Classification
>            Reporter: Isabel Drost
>
> One quarter of the work that could be done for Mahout-4. EM seems a bit to general a concept to be implemented only for the use case of PLSI, so the task of this issue is to find a way to implement EM general enough to be reused. As EM conceptually is related to k-Means it might be possible to make the implementation general enough to include k-Means as a special case.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


[jira] Resolved: (MAHOUT-28) EM implementation

Posted by "Sean Owen (JIRA)" <ji...@apache.org>.
     [ https://issues.apache.org/jira/browse/MAHOUT-28?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Sean Owen resolved MAHOUT-28.
-----------------------------

    Resolution: Won't Fix

> EM implementation
> -----------------
>
>                 Key: MAHOUT-28
>                 URL: https://issues.apache.org/jira/browse/MAHOUT-28
>             Project: Mahout
>          Issue Type: New Feature
>          Components: Classification
>            Reporter: Isabel Drost
>
> One quarter of the work that could be done for Mahout-4. EM seems a bit to general a concept to be implemented only for the use case of PLSI, so the task of this issue is to find a way to implement EM general enough to be reused. As EM conceptually is related to k-Means it might be possible to make the implementation general enough to include k-Means as a special case.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


[jira] Commented: (MAHOUT-28) EM implementation

Posted by "Sean Owen (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/MAHOUT-28?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12779213#action_12779213 ] 

Sean Owen commented on MAHOUT-28:
---------------------------------

Am I understanding that since MAHOUT-4 is to be mothballed, its dependent issues like this should be too? 

> EM implementation
> -----------------
>
>                 Key: MAHOUT-28
>                 URL: https://issues.apache.org/jira/browse/MAHOUT-28
>             Project: Mahout
>          Issue Type: New Feature
>          Components: Classification
>            Reporter: Isabel Drost
>
> One quarter of the work that could be done for Mahout-4. EM seems a bit to general a concept to be implemented only for the use case of PLSI, so the task of this issue is to find a way to implement EM general enough to be reused. As EM conceptually is related to k-Means it might be possible to make the implementation general enough to include k-Means as a special case.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.