s-news
[Top] [All Lists]

k-means cluster identification

To: s-news@wubios.wustl.edu
Subject: k-means cluster identification
From: Stafford Phillip-APS095 <phillip.stafford@motorola.com>
Date: Mon, 31 Dec 2001 11:57:41 -0600
Hello S group;

I would like to use kmeans, pam, clara, or another of the clustering algorithms 
in S+ 6 to analyze a 640x640 matrix.  I have to shrink my dataset to 213x213 in 
order to avoid this message:

Problem in replace.(formula, data): Object to copy has depth 257, only 256 
allowed (option "expressions") 


Is there a way to make any of the algorithms use all my data (clara allows 
select data to be used but I'd like to include all data) instead of only a 
portion?  I have enough memory but I'm not sure what S+ is telling me about 
it's limit.  Is there a workaround?  


Thanks for any suggestions,

Phillip Stafford
Motorola Life Sciences

<Prev in Thread] Current Thread [Next in Thread>
  • k-means cluster identification, Stafford Phillip-APS095 <=