SAS Enterprise Miner is a large product (including the SAS statistical
package), and sometimes cumbersome and quaintly old fashioned. It is
being migrated into a modern world, slowly, and its functionality is
improving regularly. Below we highlight some issues and tricks that
may be helpful.
- Log Transform of a dataset seems somewhat of a trap. A
log transform essentially transforms the data items with
. The and
are determined from the current dataset. However, if you log
transform your training dataset, and separately log transform your
test dataset, then the transforms will end up being different!
Hence, the model you build will not be applicable to any data unless
it is transformed with the same values of and !
You may want to avoid such transforms.
Copyright © 2004-2006 Graham.Williams@togaware.com
Support further development through the purchase of the PDF version of the book.
Brought to you by Togaware.