I think Classification and Regression Tree (CART) analysis might be the
way to go on that. I think there is a good demo on Salford Systems
website. Also, your math dept may have a license.
Clay
Clay E. Corbin, PhD
Biology and Allied Health
Bloomsburg University
Bloomsburg, PA 17815
-------------------------
email: [EMAIL PROTECTED]
phone: 570-389-4134
-----Original Message-----
From: morphmet [mailto:[EMAIL PROTECTED]
Sent: Monday, January 16, 2006 11:28 AM
To: morphmet
Subject: Re: Variable selection in clustering
Hi Tim,
the best software that I know for cluster analysis is NTSYSpc.
About how to select variables, I guess it depends on what you want to
do. With NTSYSpc you can have an idea of how well relationships in your
distance matrix are summarized by the tree topology by computing
cophenetic correlations. Also, you can do resampling statistics
(jacknife, permutations etc.) and write batch files for running series
of analyses.
Good luck
Andrea
>Hello all,
>
>I wonder if someone might help with advice on approaches to selecting
variables
>when clustering cases (I'm using several methods - Ward,
>bagged k-means, etc) - and am working with a large number of apparently
>relevant variables. I fear that "noisy" or irrelevant variables may be
>weakening my analysis and I would like to refine the input space by
>identifying and then eliminating any nuisance variables.
>
>My concern is to locate procedures (hopefully software) to select a
best
>sub-set of variables as input; and then refine the input space
following
>my initial exploratory clustering. I can use discriminant analysis or
>simply examine univariate F ratios - but would would seem to simply
bias
>any subsequent runs towards the classification structure produced by
the
>first analysis?
>
>Are there any other procedures for estimating the relative power of the
input
>variables? and then refining the input space?
>
> Regards
Tim
>Brennan
>-- Replies will be sent to the list.
>For more information visit http://www.morphometrics.org
>
>
>
Dr. Andrea Cardini
Hull York Medical School
The University of York, Heslington, York YO10 5DD, UK
&
The University of Hull,Cottingham Road, Hull HU6 7RX, UK
tel. 01904 321752
fax 01904 321696
E-mail: [EMAIL PROTECTED]
http://www.york.ac.uk/res/fme/people/andrea.htm
------------------------------------------------------------------------
-
Visita http://domini.interfree.it, il sito di Interfree dove trovare
soluzioni semplici e complete che soddisfano le tue esigenze in
Internet,
ecco due esempi di offerte:
- Registrazione Dominio: un dominio con 1 MB di spazio disco + 2
caselle
email a soli 18,59 euro
- MioDominio: un dominio con 20 MB di spazio disco + 5 caselle email
a soli 51,13 euro
Vieni a trovarci!
Lo Staff di Interfree
------------------------------------------------------------------------
-
--
Replies will be sent to the list.
For more information visit http://www.morphometrics.org
