Mathematica 9 is now available
Services & Resources / Wolfram Forums
-----
 /
MathGroup Archive
2002
*January
*February
*March
*April
*May
*June
*July
*August
*September
*October
*November
*December
*Archive Index
*Ask about this page
*Print this page
*Give us feedback
*Sign up for the Wolfram Insider

MathGroup Archive 2002

[Date Index] [Thread Index] [Author Index]

Search the Archive

Re: Is it possible to access internal variables?

  • To: mathgroup at smc.vnet.net
  • Subject: [mg34825] Re: Is it possible to access internal variables?
  • From: "Arny" <someone at somewhere.sometime>
  • Date: Sat, 8 Jun 2002 05:21:47 -0400 (EDT)
  • Organization: University of California, Riverside
  • References: <adkg3s$a09$1@smc.vnet.net>
  • Sender: owner-wri-mathgroup at wolfram.com

Dear Mssrs Hanlon, Wulf,

Thanks again for this help.  I believe Mr. Wulf is correct, I was confused
as to what was happening with my function.  I had believed that due to some
minor perturbations that the function, which was supposed to be bounded
between -1 and 1, was expanding outside of this range and that FindMinimum
was discovering the new Min.  The function is a log-likelihood function
which is being maximized (minned) over this range.  Outside of this range it
is complex.  It is a smooth function, but perhaps not smooth enough, it can
have very sharp bends near the bounds.  So what was probably happening was
that FindMinimum was jumping outside of the bound, hitting the non-real
area, and returning the error, which I misread to be that it had found some
value.  In fact, as you point out, it has only returned it's last-tried
value.

I ended up implementing a solution that intercepts the errors and just
starts the routine over again with different starting values.

I have attached a picture of the potential function.  
[Contact the author to get this attachment - moderator]

Ok, thanks again!

Regards,

B
___________________________________
  Bernard Gress
  Department of Economics
  University of California, Riverside
  1150 University Ave.
  Riverside, CA 92521-0247
  Fax: (909) 787-5685
  Phone: (909) 778 9813
  BGRESS at MAIL.UCR.EDU
  http://csep.ucr.edu/staff/bernard/index.htm
  ICQ: 9083461

>BobHanlon at aol.com> wrote in message news:adkg3s$a09$1 at smc.vnet.net...
> My response failed to extract the final (failed) value with the
constraints.
> My bad.
>
> However, the definition of g is not behaving differently from that for h
due
> to anything other than different definitions.  The definition of h does
not
> assign v a new value.  Copied from below:
>
> h[x_] := Module[{t = f[x]}, Append[v, {x, t}]; t];
>
> With the assignment included it will behave similarly, i.e., still won't
do
> what was requested.
>
> Bob
>
> In a message dated 6/4/02 11:59:59 AM, majort at cox-internet.com writes:


  • Prev by Date: Re: Is it possible to access internal variables? [CompoundExpression]
  • Next by Date: Re: PlotVectorField
  • Previous by thread: RE: RE: Re: Is it possible to access internal variables?
  • Next by thread: Programming an algebraic experiment