Re: voigt errors

From: Dan Dewey <dd_at_email.domain.hidden>
Date: Fri, 16 Nov 2007 15:27:06 -0500
Hi Rob,

   I also would suggest that having parameter limits like this,
while philosophically suggesting "exploring all of parameter space",
in reality-of-fitting-algorithms is not a good idea.
   At best, a fitting algorithm would essentially ignore these
values because they are so far from actual parameter values.
   At worst if the fitting routine really did explore all of
allowed parameter space, then these large limits would make it
a vanishingly small chance that anything close to the 'right' values
would ever be found.
   Even though I have just a little fitting experience, my feeling is
that fitting, especially with complex models, has to be done with
more, rather than less, human guidance and input.

happy fitting,


On Nov 16, 2007, at 1:45 PM, rgibson_at_email.domain.hidden      min         max
>  -1e+100      1e+100

     Daniel Dewey                           dd_at_email.domain.hidden     MIT Kavli Institute              Office:(617) 253-7244
     70 Vassar St., NE80-6085            Fax:(617) 253-8084
     Cambridge, MA 02139

You received this message because you are
subscribed to the isis-users list.
To unsubscribe, send a message to
isis-users-request_at_email.domain.hiddenwith the first line of the message as:
Received on Fri Nov 16 2007 - 15:28:15 EST

This archive was generated by hypermail 2.2.0 : Fri Nov 16 2007 - 15:54:19 EST