Need help lanczos algorithm

[Yogalakshmi J <yo...@gmail.com>]

Hi,

I am a newbie to apache mahout. I have mahout installed in my system. I am
interested in the Recommender part. I have 2 questions:
1. I tried SVDRecommender with movielens(100k) data, with the parameters 10
for number of features, and 50 for number of iterations. I did not exactly
understand what is to be passed for these two parameters?
2. I want to make use of the Lanczos algorithm that is available. I want to
create a recommender, that will factorize the matrix using Lanczos
algorithm. Towards that I created a class extending AbstractRecommender.
Inside this we need to override the method recommend, estimatepreference
among others. I am not sure how to exactly go about. There are 2 lanczos
solvers that I could see: DistributedLanczosSolver and LanczosSolver. The
DistributedLanczossolver extends LanczosSolver, so should I make use of
DistributedLanczosSolver. It has something called LanczosState, which I
could not understand what it is. Any guidance is deeply thanked.

Thanks,
Yogalakshmi

Re: Need help lanczos algorithm

[Ted Dunning <te...@gmail.com>]

On Wed, Apr 24, 2013 at 2:41 AM, Yogalakshmi J <yo...@gmail.com>wrote:

> 2. I want to make use of the Lanczos algorithm that is available. I want to
> create a recommender, that will factorize the matrix using Lanczos
> algorithm.
>

Lanczos is deprecated.

Look for the SSVD implementation.  Or for recommendation, look for the ALS
decomposer.