[
https://issues.apache.org/jira/browse/HAMA-660?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13485740#comment-13485740
]
Hudson commented on HAMA-660:
-----------------------------
Integrated in Hama-Nightly #720 (See
[https://builds.apache.org/job/Hama-Nightly/720/])
[HAMA-660] - total number of items to read is pre calculated and passed to
the CostFunction (Revision 1402944)
Result = SUCCESS
tommaso :
Files :
* /hama/trunk/ml/src/main/java/org/apache/hama/ml/regression/CostFunction.java
*
/hama/trunk/ml/src/main/java/org/apache/hama/ml/regression/GradientDescentBSP.java
*
/hama/trunk/ml/src/main/java/org/apache/hama/ml/regression/LinearRegressionModel.java
*
/hama/trunk/ml/src/main/java/org/apache/hama/ml/regression/LogisticRegressionModel.java
*
/hama/trunk/ml/src/main/java/org/apache/hama/ml/regression/RegressionModel.java
> Make item cost calculation consistent in GD BSP
> -----------------------------------------------
>
> Key: HAMA-660
> URL: https://issues.apache.org/jira/browse/HAMA-660
> Project: Hama
> Issue Type: Improvement
> Affects Versions: 0.5.0
> Reporter: Tommaso Teofili
> Assignee: Tommaso Teofili
> Fix For: 0.6.0
>
>
> Cost per item is calculated without knowing the total number of items which
> is ok from an efficiency point of view but makes plugging of new cost
> functions bug prone since the total cost is then always divided by the number
> of items.
> It'd be good to have a 0th superstep where total number of items is read and
> then it's used in CostFunction#calculateCostPerItem
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
