| clustering- yet again [message #56243] |
Tue, 09 October 2007 14:05 |
nivedita.raghunath
Messages: 15
Registered: July 2006
|
Junior Member |
|
|
Hi All,
The clustering saga never ended!
I have a large dataset with 5 visibly distinct clusters. Based on
advice from y'all, I tried standardize() and it worked to some extent.
It finds 4 clusters. I have a good idea about my input variables-
included the variable_wts option and tried, yet with no success.
This is the number of samples it returns in each cluster:
IDL> print,clust_num
6.00000
28903.0
25205.0
37694.0
42458.0
It put only 6 samples into cluster 0, whereas actually I have samples
in the range 25000 or so. The remaining samples have been put into
cluster 4 (and I can visually see that 0 and 4 are NOT close).
Any ideas on this?
My dataset is very huge to post here, but if anyone is willing to
help, I can mail it.
Sincerely,
Nivedita
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
