mahout-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Julian Ortega <jorte...@gmail.com>
Subject Re: ALSWRFactorizerTest.java
Date Thu, 22 Nov 2012 17:51:37 GMT
You should have a look at the Mahout in Action book.

Nonetheless, this
example<https://github.com/tdunning/MiA/blob/mahout-0.7/src/main/java/mia/recommender/ch02/RecommenderIntro.java>from
the book shows the basics of making a recommender. In the example they
are using GenericUserBasedRecommender, in your case you would need a
SVDRecommender and you tell that recommender the type of Factorizer you
want to use, e.g. ALSWRFactorizer. That said, for a SVDRecommender you
don't need to specify a sense of similarity nor of neighborhood. However,
you can specify sampling strategies and/or whether the generated model
should be persisted (to avoid recalculating if you don't need to. Hope that
helps


On Thu, Nov 22, 2012 at 6:37 PM, Abhijith CHandraprabhu <abhijithc@gmail.com
> wrote:

>  I am a first time java user. After struggling with MAHOUT+MAVEN+ECLISE for
> a whole month, I have been able to get the system working. Could ony of you
> kindly save me sometime by letting me know how can I use a particular
> function like ALSWRFactorizer.java. My main intention is to study the
> ALS-WR factorization, hence I would like to see the way recommendation is
> done on the test data in ALSWRFactorizerTest.java. Where should I write my
> code if any. How can I check the code in Debug mode i.e with breakpoints
> like in MATLAB.  I am willing to learn things myself, but i just need some
> right direction. I spent a whole month is just knowing what is MAVEN and
> ECLIPSE and setting up MAHOUT.
>
> I have no problems with the mathematical aspects, but I am not familiar
> with the implementation in java, however I am hellbent on learning to
> program with java but I am so lost.
>
> --
> Best regards,
> Abhijith
>

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