[ai-geostats] Re: facies indiccator modeling
The whole point (as I understand it) of using an
indicator approach is to remove all question of what
the original distribution was. All your sample values
become 0 or 1, at each 'cutoff'.
Not knowing the algorithms you are using, I cannot
comment on what other needs it might have but SIS
should be 'distribution free'.
I am not clear why you feel your sill needs to be 1.
Is this a requirement of the software package?