mahout-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Sean Owen <sro...@gmail.com>
Subject Re: Mahout on Hadoop
Date Mon, 07 Dec 2009 16:25:19 GMT
Yes, use the GroupLens 100K data set.

I don't know what data file you are passing, but this does not appear
to be a valid preference file, what is this? Should be
userID,itemID,pref. You may need to transform the input.


On Mon, Dec 7, 2009 at 4:14 PM, Rajpal, Harjeet Kumar
<Harjeet.Kumar@honeywell.com> wrote:
>
> Ya you are right. I corrected that mistake. Sorry for inconvenience.
> I tried this with Grouplens data. It gave java.lang.OutOfMemoryError.
> can you suggest any smaller amount of data for this. I tried to use a
> part of grouplens data. Then it gave number format exception.
> "1,F,1,10,48067" .
>
> If you can provide some demo data. That will be helpful.
> Best Regards,
> Harjeet Kumar
>
> -----Original Message-----
> From: Sean Owen [mailto:srowen@gmail.com]
> Sent: Monday, December 07, 2009 3:17 PM
> To: mahout-user@lucene.apache.org
> Subject: Re: Mahout on Hadoop
>
> Yes, that's right, you want to use the .job file as your ".jar" file.
> Are you passing it both as the argument to "hadoop jar" and the
> "--jarFile" argument? I suspect you still have some deployment problem
> since this all works for me.
>
> On an unrelated note, I am changing this part of the code dramatically
> right now, improving it and cleaning it up. You may want to update at
> some point but it will likely break what you are doing a little bit.
>
> On Mon, Dec 7, 2009 at 9:18 AM, Rajpal, Harjeet Kumar
> <Harjeet.Kumar@honeywell.com> wrote:
>> I have extracted the mahout-core-snapshot.job and checked the lib. The
>> relevant uncommons-maths-1.2.jar is inside that.
>> I hope I am not making any stupid mistake here. If I am, please
> correct
>> me. Best Regards,
>> Harjeet Kumar Rajpal
>>
>> -----Original Message-----
>> From: Sean Owen [mailto:srowen@gmail.com]
>> Sent: Monday, December 07, 2009 1:02 PM
>> To: mahout-user@lucene.apache.org
>> Subject: Re: Mahout on Hadoop
>>
>> Caused by: java.lang.ClassNotFoundException:
>> org.uncommons.maths.random.MersenneTwisterRNG
>>
>> You aren't bundling the dependent class files with the .jar file used
>> with the Hadoop job.
>>
>> On Mon, Dec 7, 2009 at 7:01 AM, Rajpal, Harjeet Kumar
>> <Harjeet.Kumar@honeywell.com> wrote:
>>> Hi Sean,
>>>
>>> Sorry for Delay in Reply.
>>> I got the new code now. The error has changed.
>>>
>>
>

Mime
View raw message