I checked the source code for usage of ClusterWritables to write centriods to a sequence file i found out this
SequenceFile.Writer writer = new SequenceFile.Writer(fs, conf, path, Text.class, ClusterWritable.class); ClusterWritable clusterWritable = new ClusterWritable(); clusterWritable.setValue(canopy); writer.append(new Text(canopy.getIdentifier()), clusterWritable); heresetValue expects Cluster, what is the way i can convert a raw vector(SequentialAccessSparseVector) to Cluster type ? ________________________________ From: Jeff Eastman <[email protected]> To: [email protected] Sent: Saturday, 2 February 2013 12:18 AM Subject: Re: intial centriods for fuzzy k means algorithm If you don't specify a -k value but specify a -ci directory that contains clusters you want to use for the prior then the ClusterIterator will use them for kmeans and fuzzyk. You will need to create one or more sequence files containing ClusterWritables to do this. On 2/1/13 9:08 AM, sri krishna wrote: > my question was more like how can i generate new centriods based on the > predefined points i give, as in my case i know the number of clusters and > also few points in each of the cluster. > > > > > > > ________________________________ > From: Rajesh Nikam <[email protected]> > To: [email protected]; sri krishna <[email protected]> > Sent: Friday, 1 February 2013 5:07 PM > Subject: Re: intial centriods for fuzzy k means algorithm > > you could use canopy clustering from mahout to initialize centroids. > > Thanks > Rajesh > > > On Fri, Feb 1, 2013 at 4:43 PM, sri krishna <[email protected]> wrote: > >> Hi, >> >> >> I have sample set of few documents of each cluster(no of clusters known >> and also few documents in each cluster are known in advance). How to >> initialize the centriods with known documents, so that algorithm runs using >> the given data points as centriods in mahout ?
