Statalist


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

st: RE: kmeans clustering -initial starting points


From   "Nick Cox" <n.j.cox@durham.ac.uk>
To   <statalist@hsphsun2.harvard.edu>
Subject   st: RE: kmeans clustering -initial starting points
Date   Mon, 8 Jun 2009 14:22:01 +0100

The references you give are just names and dates. The Statalist FAQ
spells out that posters are asked to give full references. 

On your main point, it is impossible to comment on syntax errors if you
do not indicate exactly what you typed. Again, the Statalist FAQ does
ask you to do that. 

More positively:

1. It would seem that -start(group(varname))- is the syntax you're
looking for. 

2. I've not seen this advice in the literature I've read, but I'd find
it more convincing if kmeans and Ward's were applied independently and
shown to produce similar results (or not, as the case may be). 

Nick 
n.j.cox@durham.ac.uk 

Alp Eren Yurtseven

I am quite new to STATA and I have a question about kmeans clustering. A

combination of hierarchical and non-hierarchical techniques is advised 
in the literature (Punj and Stewart, 1983; de Jong and Marsili, 2006). 
This literature suggests generating initial solutions using the Ward's 
method and then employ the centroids in the kmeans clustering as  
starting points. I checked out the start option but keep getting syntax 
errors. I would be most glad if I could get an insight on how to combine

these two methods.


*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/



© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index