On Wed, Jun 25, 2008 at 5:42 AM, Glynn Clements <[EMAIL PROTECTED]> wrote: [useful info skipped] > So, it may be necessary to pre-scale the matrix elements by some > factor, then scale the inverse (and determinant) to compensate.
Yes, maybe taking the data range into account, an a priori scale factor can be determined for the matrix elements. I'll try to generate some "bad" integer maps but I am not really sure on what it depends. One try could be to use too tiny training areas. Markus _______________________________________________ grass-dev mailing list [email protected] http://lists.osgeo.org/mailman/listinfo/grass-dev
