FindMinimum
- To: mathgroup at smc.vnet.net
- Subject: [mg13968] FindMinimum
- From: Christian Keysers <ck6 at st-andrews.ac.uk>
- Date: Fri, 11 Sep 1998 15:06:57 -0400
- Sender: owner-wri-mathgroup at wolfram.com
Dear All,
I am trying to fit gaussians to data using the FindMinimum function.
The problem is that in some case, the error function that I try to
minimize as a function of the mean and standart deviation of the normal
distribution keeps going down with m and s increasing. The error only
becomes marginally smaller, but FindMinimum keeps trying larger and
larger values. How can I tell it that if the error only changes in
steps smaller than 0.001, it shouldn't bother going on? I know there
are two paramaters AccuracyGoal and PrecisionGoal, but they are not
explained in the book. Does anybody know more?
--
________________________________________________________
Christian Keysers
School of Psychology
Univ. of St Andrews
Scotland KY16 9JU
UK
Tel.: +44.1334.462083 or 463045
Fax.: +44.1334.463042
Email: ck6 at st-and.ac.uk
________________________________________________________