[
https://issues.apache.org/jira/browse/MAHOUT-586?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13055383#comment-13055383
]
Lance Norskog commented on MAHOUT-586:
--------------------------------------
This is a much less ambitious rewrite.
* The existing RecommenderEvaluator is slightly refactored and can now run one
recommender against two different data models.
* OrderBasedRecommenderEvaluator does not implement RecommenderEvaluator.
* All of the order-based algorithms are subclasses and OBRE is an abstract
class.
* Evaluations are collected in a RunningAverage parameter.
* An RMSRunningAverage class is included to allow RMSE evaluation of
order-based scoring. Unit test included.
* A program to run the Yahoo KDD 2011 data sets with separate training,
validation and test data sets, as this is apparently the point of the contest.
> Redo RecommenderEvaluator for modularity
> ----------------------------------------
>
> Key: MAHOUT-586
> URL: https://issues.apache.org/jira/browse/MAHOUT-586
> Project: Mahout
> Issue Type: Improvement
> Components: Collaborative Filtering
> Reporter: Lance Norskog
> Assignee: Sean Owen
> Fix For: 0.5
>
> Attachments: MAHOUT-559.patch, MAHOUT-586.patch, MAHOUT-586.patch,
> MAHOUT-586.patch, MAHOUT-586.patch
>
>
> The RecommenderEvaluator implementation is hard-coded around one algorithm.
> This is a more flexible, modular rewrite.
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
