You are viewing a plain text version of this content. The canonical link for it is here.

Posted to user@mahout.apache.org by Jay Vyas <ja...@gmail.com> on 2014/03/28 14:04:08 UTC

The 3 distributed recommenders

Hi mahout:

Looking through the source code there are 3 distributed recommenders...

the als recommender
the item recommender
the pseudo recommender

Any docs differentiating these?

-- 
Jay Vyas
http://jayunit100.blogspot.com

Re: The 3 distributed recommenders

Posted by Jay Vyas <ja...@gmail.com>.

Well, anyways:  To clarify....  I have a pretty simple implementation im
trying to run:

       int ret = recommenderJob.run(new String[] {
             "--input",args[0],
             "--output",args[1],
             "--usersFile","/tmp/users.txt",
             "--tempDir", "/tmp",
             "--similarityClassname", "SIMILARITY_PEARSON_CORRELATION",

And it gets the error described here.... (numUsers.bin not found).

http://web.archiveorange.com/archive/v/y0uRZUrY8BQGOkwWg34i

If i manually create numUsers.bin , then it returns a new exception:

 /tmp/preparePreferenceMatrix/ratingMatrix not Found.

So... Im thinking i must be doing something horribly wrong in my
recommender, but cant figure out exactly what?



On Fri, Mar 28, 2014 at 2:29 PM, Jay Vyas <ja...@gmail.com> wrote:

> Thanks sebastian.     I guess im looking more for some hints on how to use
> the mahout API for this.
>
> 1)  Does mahout support pure distributed collaborative filtering ?
>
> 2) If so , is there any available example of how to use the Mahout API for
> this?
>
>
>
>
>
> On Fri, Mar 28, 2014 at 11:56 AM, Sebastian Schelter <ss...@apache.org>wrote:
>
>> Hi Jay,
>>
>> there's not much documentation unfortunately. We're in the process of
>> creating that however. We removed the pseudo-distributed recommender,
>> mainly because nobody ever used it. There are two research papers that
>> could help you with understanding the other two distributed recommenders:
>>
>> For ALS:
>>
>> Distributed Matrix Factorization with MapReduce using a series of
>> Broadcast-Joins, RecSys'13
>>
>> http://ssc.io/wp-content/uploads/2011/12/sys024-schelter.pdf
>>
>> For item-based:
>>
>> Scalable Similarity-Based Neighborhood Methods with MapReduce, RecSys'12
>>
>> http://ssc.io/wp-content/uploads/2012/06/rec11-schelter.pdf
>>
>>
>>
>> On 03/28/2014 02:04 PM, Jay Vyas wrote:
>>
>>> Hi mahout:
>>>
>>> Looking through the source code there are 3 distributed recommenders...
>>>
>>> the als recommender
>>> the item recommender
>>> the pseudo recommender
>>>
>>> Any docs differentiating these?
>>>
>>>
>>
>
>
> --
> Jay Vyas
> http://jayunit100.blogspot.com
>



-- 
Jay Vyas
http://jayunit100.blogspot.com

Re: The 3 distributed recommenders

Posted by Jay Vyas <ja...@gmail.com>.

Thanks sebastian.     I guess im looking more for some hints on how to use
the mahout API for this.

1)  Does mahout support pure distributed collaborative filtering ?

2) If so , is there any available example of how to use the Mahout API for
this?





On Fri, Mar 28, 2014 at 11:56 AM, Sebastian Schelter <ss...@apache.org> wrote:

> Hi Jay,
>
> there's not much documentation unfortunately. We're in the process of
> creating that however. We removed the pseudo-distributed recommender,
> mainly because nobody ever used it. There are two research papers that
> could help you with understanding the other two distributed recommenders:
>
> For ALS:
>
> Distributed Matrix Factorization with MapReduce using a series of
> Broadcast-Joins, RecSys'13
>
> http://ssc.io/wp-content/uploads/2011/12/sys024-schelter.pdf
>
> For item-based:
>
> Scalable Similarity-Based Neighborhood Methods with MapReduce, RecSys'12
>
> http://ssc.io/wp-content/uploads/2012/06/rec11-schelter.pdf
>
>
>
> On 03/28/2014 02:04 PM, Jay Vyas wrote:
>
>> Hi mahout:
>>
>> Looking through the source code there are 3 distributed recommenders...
>>
>> the als recommender
>> the item recommender
>> the pseudo recommender
>>
>> Any docs differentiating these?
>>
>>
>


-- 
Jay Vyas
http://jayunit100.blogspot.com

Re: The 3 distributed recommenders

Posted by Sebastian Schelter <ss...@apache.org>.

Hi Jay,

there's not much documentation unfortunately. We're in the process of 
creating that however. We removed the pseudo-distributed recommender, 
mainly because nobody ever used it. There are two research papers that 
could help you with understanding the other two distributed recommenders:

For ALS:

Distributed Matrix Factorization with MapReduce using a series of 
Broadcast-Joins, RecSys'13

http://ssc.io/wp-content/uploads/2011/12/sys024-schelter.pdf

For item-based:

Scalable Similarity-Based Neighborhood Methods with MapReduce, RecSys'12

http://ssc.io/wp-content/uploads/2012/06/rec11-schelter.pdf

On 03/28/2014 02:04 PM, Jay Vyas wrote:
> Hi mahout:
>
> Looking through the source code there are 3 distributed recommenders...
>
> the als recommender
> the item recommender
> the pseudo recommender
>
> Any docs differentiating these?
>