mahout-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Sean Owen" <sro...@gmail.com>
Subject Re: help for item based recommender
Date Tue, 27 May 2008 14:39:27 GMT
This sounds like a different question. You have preference data and
you want to provide it to a Recommender? you need to use a DataModel
implementation for this, and FileDataModel should read the file as you
have it. Just pass it to the FileDataModel, and FileDataModel to your
Recommender.

If you have some file of *item-item correlations* that you want to
feed to GenericItemCorrelation, then simply read them one by one,
create ItemItemCorrelation objects in a List, and feed that to  the
GenericItemCorrelation constructor.

If you want to use PearsonCorrelation to automatically figure the
correlations between items, then feed FileDataModel to
PearsonCorrelation and use PearsonCorrelation in the constructor to
GenericItemBasedRecommender.


On Tue, May 27, 2008 at 4:36 AM, Cam Bazz <cambazz@gmail.com> wrote:
> Hello Sean,
>
> Thanks a lot for the detailed explanation. But I still have not understood -
> how.
>
> lets say I have a file that has lines of userid, itemid, score
>
> I could found in the documentation a way of feeding it to the recomender.
>
> Best Regards,
> -C.B.

Mime
View raw message