Re: Clustering / Classification

hi nadav,
i think this program should do

http://www.mathworks.com/matlabcentral/fileexchange/loadFile.do?objectId=71&objectType=file

particle reddy



On Mar 19, 11:13 am, "Nadav" <nadav...@xxxxxxxxx> wrote:
Hi,

I have a huge NxN ( thousands of elements ) correlation matrix, I need
to cluster this data ( of N elements ) to X groups of the most similar
items, this can be done using K-Means & friends though it would take
allot of time, I wonder, taking in mind I already have the correlation
matrix ( hence the similarity of each element to the other ) is there
a way, more efficient then K-means ( and friends ) to cluster the data
to X groups ?

Any help would be appreciated.

Thanks,
Nadav Rubinstein.


.

Relevant Pages

  • Clustering / Classification
    ... I have a huge NxN correlation matrix, ... to cluster this data to X groups of the most similar ... this can be done using K-Means & friends though it would take ...
    (comp.dsp)
  • Re: Clustering von kategoriellen Daten
    ... > Florian Laws wrote: ... wieviele Cluster man haben will. ... Ich hatte k-Means nur mal so als Beispiel genommen, ... Next by Date: ...
    (de.sci.informatik.ki)
  • Re: Write the cluster in dataset
    ... Ok, I know whats is k-means, I'm graduated and master in Statistics, and Phd ... "Roy Harvey (SQL Server MVP)" wrote: ... So I just want to create a field with the cluster group in the table that I ... The k-means algorithm is an algorithm to cluster n objects based on ...
    (microsoft.public.sqlserver.clustering)
  • Calculating centroids in k-means clustering (for different distance metrics)
    ... I have successfully implemented k-means using Euclidean distance ... Normalized Cluster Centers with KMedians) ... A central part of k-means clustering is that a centroid is calculated ...
    (sci.math)
  • Re: Setting cluster seeds for K-Means analysis in SPSS
    ... First do a K-means, 2 cluster, select Write final as (select Options ... Open the previously saved file, replace the data in SPSS and save. ... > I've done a heirarchical cluster analysis and decided on a 2 cluster ...
    (sci.stat.edu)