mahout-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Bill Au <bill.w...@gmail.com>
Subject Re: Bayesian examples not working
Date Sat, 20 Jun 2009 01:23:04 GMT
Thanks for checking, Grant.  I am able to ran install with all the tests
passing on OS X too.  I check out the same revision, 786731, on my Linux box
(RHEL 4 64-bit) but the same test keep failing.
I checked the
POM.  It looks like both the trunk and the 0.1 release tar ball
requires Hadoop 0.19.1.  I will update the cluster to 0.19.1 next week
and try the example again.  Should we update the Wiki to point this
out?

Bill

On Fri, Jun 19, 2009 at 5:03 PM, Grant Ingersoll <gsingers@apache.org>wrote:

> I just did a clean and an update and ran "mvn install" and all tests pass
> for me.
>
> I'm running JDK 1.6 on OS X (latest) with Maven 2.0.10.
>
> You can skip the tests in Maven by passing in -DskipTests=true
>
> However, it makes me wonder what's going wrong and whether it will work.
>
> Note, also, we have been reworking some of the clustering stuff internals,
> so maybe you happened to get a copy of trunk at the wrong time.
>
> -Grant
>
>
> On Jun 19, 2009, at 4:23 PM, Grant Ingersoll wrote:
>
>  Hey Bill,
>>
>> I've been having problems w/ the Bayesian examples too, but not enough
>> time to look into them.  However, that is about to change and will be
>> digging into it a bit more early next week or this weekend.
>>
>> I will note a couple of things.  I've seen some reports of problems w/
>> Maven 2.1.0 not working on Mahout.  Also, I believe the trunk version
>> requires Hadoop 0.19.1, but check the POM to be sure.
>>
>> In the meantime, any insight you have is appreciated.
>>
>> -Grant
>>
>> On Jun 19, 2009, at 3:03 PM, Bill Au wrote:
>>
>>  I am using the 0.1 release tar ball and the Hadoop job files contained in
>>> it.  I am trying to run the twenty newsgroup example by following the
>>> instructions in the Wiki.  I am not getting any errors nor exception but
>>> the
>>> results are not correct.  Here is the output I got:
>>>
>>> http://people.apache.org/~billa/mahout0.1-output
>>>
>>> I am getting the same results for both Bayes and CBayes.  Any idea what I
>>> am
>>> doing wrong?
>>>
>>> I also want to run the example using the trunk so I checked it out and
>>> ran
>>> "mvn install".  I found that the Hadoop job files are not being created
>>> due
>>> to a test failure:
>>>
>>> Running org.apache.mahout.clustering.kmeans.TestKmeansClustering
>>> Tests run: 5, Failures: 1, Errors: 0, Skipped: 0, Time elapsed: 34.979
>>> sec
>>> <<< FAILURE!
>>>
>>>  <failure message="clusters[3] expected:&lt;4&gt; but was:&lt;2&gt;"
>>>
>>> type="junit.framework.AssertionFailedError">junit.framework.AssertionFailedError:
>>> clusters[3] expected:&lt;4&gt; but was:&lt;2&gt;
>>>      at junit.framework.Assert.fail(Assert.java:47)
>>>      at junit.framework.Assert.failNotEquals(Assert.java:280)
>>>      at junit.framework.Assert.assertEquals(Assert.java:64)
>>>      at junit.framework.Assert.assertEquals(Assert.java:198)
>>>      at
>>>
>>> org.apache.mahout.clustering.kmeans.TestKmeansClustering.testKMeansMRJob(TestKmeansClustering.java:432)
>>> </failure>
>>>
>>> I am not sure how to fix this.  Is there anyway to build the Hadoop job
>>> files without running the tests?
>>>
>>> By the way, I am using maven 2.1.0 and Hadoop 0.18.3.
>>>
>>> Bill
>>>
>>
>>
> --------------------------
> Grant Ingersoll
> http://www.lucidimagination.com/
>
> Search the Lucene ecosystem (Lucene/Solr/Nutch/Mahout/Tika/Droids) using
> Solr/Lucene:
> http://www.lucidimagination.com/search
>
>

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message