Mailing-List: contact user-help@mahout.apache.org; run by ezmlm
Precedence: bulk
Reply-To: user@mahout.apache.org
Received-SPF: pass (nike.apache.org: domain of recsysuser@gmail.com designates
 209.85.212.52 as permitted sender)
MIME-Version: 1.0
In-Reply-To: <51D0CFAA.9020804@googlemail.com>
References: 
 <CACvDQoLjNETUvyF4uPtH0QnUuKgmj1daFoMEp79bSMzBpZgT2A@mail.gmail.com>
	<51D0CFAA.9020804@googlemail.com>
Date: Sun, 30 Jun 2013 20:47:14 -0400
Message-ID: 
 <CACvDQoJ-QZ+B0GPXzi1NegDd=6Eb3apXQNy8fTv7+djzrMgjFg@mail.gmail.com>
Subject: Re: Vectorize the movielens 100K dataset for Mahout k-means
 clustering
From: Carlos Seminario <recsysuser@gmail.com>
To: user@mahout.apache.org
Content-Type: multipart/alternative; boundary=20cf3079babacc605304e06891b6

--20cf3079babacc605304e06891b6
Content-Type: text/plain; charset=ISO-8859-1

Ok. Has anyone already done this (per user) and are willing to share either
the vectors or the code? Thanks .. Carlos

On Sun, Jun 30, 2013 at 8:39 PM, Sebastian Schelter <ssc.open@googlemail.com
> wrote:

> Simply write a Java program, create the vectors per user or item (don't
> know how you want to cluster) and write them out via SequenceFileWriter.
>
> On 01.07.2013 02:29, Carlos Seminario wrote:
> > Hi: I want to vectorize the movielens 100K dataset as a
> > RandomAccessSparseVector and use it to run Mahout k-means clustering. Has
> > anyone done this before? If not, any ideas on a how this can be done?
> (BTW,
> > movielens dataset contains ~100K records/lines with this format: userid,
> > itemid, rating, unix time.)
> >
> > Thanks .. Carlos
> >
>
>

--20cf3079babacc605304e06891b6--