mahout-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Manuel Blechschmidt (Created) (JIRA)" <j...@apache.org>
Subject [jira] [Created] (MAHOUT-905) CachingUserSimilarity and CachingItemSimilarity have wrong (far to small) default maxSizes
Date Wed, 30 Nov 2011 22:13:40 GMT
CachingUserSimilarity and CachingItemSimilarity have wrong (far to small) default maxSizes
------------------------------------------------------------------------------------------

                 Key: MAHOUT-905
                 URL: https://issues.apache.org/jira/browse/MAHOUT-905
             Project: Mahout
          Issue Type: Bug
          Components: Collaborative Filtering
    Affects Versions: 0.5
         Environment: Mac OS X 10.6.8
java version "1.6.0_29"
Java(TM) SE Runtime Environment (build 1.6.0_29-b11-402-10M3527)
Java HotSpot(TM) 64-Bit Server VM (build 20.4-b02-402, mixed mode)

This does not matter should be reproducible on every system.
            Reporter: Manuel Blechschmidt
            Assignee: Sean Owen
         Attachments: CachingSimilariyAdjustedDefaultSize.patch

I am currently tuning my recommender discussed here: http://thread.gmane.org/gmane.comp.apache.mahout.user/10433.

As a first step I wrapped my LogLikelihoodSimilarity with an CachingUserSimilarity. I used
Java Visual VM to profile the calls. I recognized that I didn't get any performance benefits.
So I had a look into the code.

Actually line 47 this(similarity, dataModel.getNumItems()); in CachingUserSimilarity.java
is wrong. If we want to cache all item similarities we need a cache with (dataModel.getNumItems()*(dataModel.getNumItems()-1))/2
possible entries.

I am now doing this in the constructor. I attached a patch to adjust this in the trunk build.

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira

        

Mime
View raw message