Thank you Andrew for the nice suggestion. I found out the mistake I made in the
code. Here is the complete function with the problem of optimization.
trun.mean<- function(x)
{
mu=x[1];
sigma=x[2];
thre=x[3];
f <- function(x) (1/(sigma*sqrt(2*pi)))*exp(-(x-mu)^2/(2*sigma^2));
pdf.fun <- function(x) x*f(x);
sd.fun <- function(x) x^2*f(x);
a=integrate(pdf.fun,thre,upper=Inf)$value;
a1=integrate(f,thre,upper=Inf)$value;
st=integrate(sd.fun,thre,upper=Inf)$value;
mut <- a/a1;
sdt <- sqrt(st/a1-(a/a1)^2)
(mut-mu)^2+(sdt-sd)^2
}
I need to minimize trun.mean ( ), so I tried
> trun.mean(c(100,20,13))
[1] 399200454
> optim(c(200,100,112),trun.mean);
Error in integrate(f, thre, upper = Inf) :
roundoff error is detected in the extrapolation table
Thanks!
Jenny
-----Original Message-----
From:Andrew Robinson , [EMAIL PROTECTED]
Sent: 2006-11-09, 15:14:58
To: Jenny Stadt
CC:[email protected]
Subject: Re: [R] optimize function with integral form ?
Hi Jenny,
whenever a function that calls a function complains about the function
it is calling, it helps to try to run that function yourself. So, I
declared your function trun.mean as below and tried to call it:
> trun.mean(c(200,100,113))
Error in integrate(f, thre, upper = Inf) :
object "thre" not found
This might help you identify the current, and possibly future, errors.
Cheers
Andrew
On Thu, Nov 09, 2006 at 02:32:18PM -0700, Jenny Stadt wrote:
> Hi all,
>
> Does anybody have the experience of using optim to estimate variables with
> integral forms?
>
> here the code:
>
>
> trun.mean <- function(x) # t is the threshold
> {
>
> mu=x[1];
>
> sigma=x[2];
>
> t=x[3];
>
> f <- function(x) (1/(sigma*sqrt(2*pi)))*exp(-(x-mu)^2/(2*sigma^2));
>
> pdf.fun <- function(x) x*f(x);
>
> integrate(f,thre,upper=Inf)$value/integrate(pdf.fun,thre,upper=Inf)$value ;
> }
>
> when I run :
>
> > optim(c(200,100,113),trun.mean)
>
> Error in optim(c(2, 3, 3), trun.mean) : function cannot be evaluated at
> initial parameters
>
>
>
> I need your help, thanks!
>
> Jen.
>
> [[alternative HTML version deleted]]
>
> ______________________________________________
> [email protected] mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
--
Andrew Robinson
Department of Mathematics and Statistics Tel: +61-3-8344-9763
University of Melbourne, VIC 3010 Australia Fax: +61-3-8344-4599
http://www.ms.unimelb.edu.au/~andrewpr
http://blogs.mbs.edu/fishing-in-the-bay/
[[alternative HTML version deleted]]
______________________________________________
[email protected] mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
