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 Tue, 08 Dec 2009 16:40:08 GMT
The good news is you are just about to get this running. I think it
will work now.

The bad news is I have already changed this code a little bit in the
last week. For example, the arguments have changed. You now pass the
ratings data as --input, there is no --dataModelFile flag, and there
is no more user ID file. It's simpler and faster though.

On Tue, Dec 8, 2009 at 11:32 AM, Rajpal, Harjeet Kumar
<Harjeet.Kumar@honeywell.com> wrote:
> Yes, I did that. Its working now. Sorry for inconvenience.
> Thanks for your help. Even though questions were stupid, still you were
> very helpful.
>
> -----Original Message-----
> From: Sean Owen [mailto:srowen@gmail.com]
> Sent: Tuesday, December 08, 2009 2:50 PM
> To: mahout-user@lucene.apache.org
> Subject: Re: Mahout on Hadoop
>
> Harjeet I think you need to read the documentation of the code, of the
> data you're using and of Hadoop. I think many of the questions you are
> asking are answerable from those resources.
>
> The userIdFile is not your ratings file. It is a file full of user IDs
> to recommend for.
>
> On Tue, Dec 8, 2009 at 7:05 AM, Rajpal, Harjeet Kumar
> <Harjeet.Kumar@honeywell.com> wrote:
>> I just read the code in Recommender and saw that userIdfile path is
>> given as input to mapred.input.dir. if I only give --userIdFile
> /input/
>> instead of --userIdFile /input/users.dat. then it still gives number
>> format error.
>>
>

Mime
View raw message