Clustering without hadoop

[Florents Tselai <fl...@gmail.com>]

Hello,

I'm working on market basket data clustering and I'd like to know the
fastest(quick and dirty) way to use mahout.

Specifically, Is it possible to run Kmeans or EM without having a HDFS
configured?

If I'm not wrong Mahout 0.5 Kmeans implementation didn't require HDFS but
the latest version does.

Re: Clustering without hadoop

[Johannes Schulte <jo...@gmail.com>]

Hi Florents,

it just became different but still works without hdfs, i also had trouble
getting the right classes together but here is something that will
hopefully work correctly:

  DistanceMeasure measure = new CosineDistanceMeasure();

  // ClusterUtils is no mahout class

  List<Cluster> initialClusters = ClusterUtils.getInitialClusters(points,
k, measure);


   System.out.println("going ma d!");


        ClusterClassifier prior =

                new ClusterClassifier(initialClusters,
newKMeansClusteringPolicy(0.01));


        ClusterClassifier clustered = ClusterIterator.iterate(points,
prior, 10);

        System.out.println(clustered.getModels());

        List<Cluster> finalClusters = clustered.getModels();


Cheers,


Johannes


On Mon, Nov 12, 2012 at 11:55 PM, Florents Tselai <florents.tselai@gmail.com
> wrote:

> Hello,
>
> I'm working on market basket data clustering and I'd like to know the
> fastest(quick and dirty) way to use mahout.
>
> Specifically, Is it possible to run Kmeans or EM without having a HDFS
> configured?
>
> If I'm not wrong Mahout 0.5 Kmeans implementation didn't require HDFS but
> the latest version does.
>