R help - Sep 2007 - problem with 'integrate'

Hello -

I am having a problem with the function 'integrate'.  I am running R  
on OSX (R 2.5.1).

I am trying to suppress the error message when 'integrate' attempts  
to integrate across a parameter set giving a non-finite function  
value.  I'm using it in a MCMC / simulated annealing algorithm, and  
it is entirely possible that some parameter sets will give non- 
integrable functions.  I thought that the argument

stop.on.error = FALSE

would cause the function to return NA or at least not return an error  
(I can work around the situation where integrate returns garbage, but  
if it generates an error, it terminates my chain and all appears lost).

This may not be necessary, but here is a simplification of an  
instance that generates the error.  If you send any B > 500 to  
"questionableFunction", the integration fails and returns an error  
message - even with stop.on.error = FALSE.  Is there any way I can  
work around this?  Perhaps there is something similar to  
"suppressWarnings()" that I might be able to use?

questionableFunction <- function(B = 10)
{
	fx4 <- function(B, y, z){
    		1 + exp(B*y - B*z);
    	 }

	fx1 <- function(x, r0 = 0.12, k=-0.10, a=0.5, tshift=8.36){
		(r0*exp(r0*t0 + (k/B)*log(fx4(B, t0, tshift)) - r0*x - (k/B)*log(fx4 
(B, x, tshift))))
	}

	integrate(fx1, 8.36, 10, stop.on.error = FALSE);
}



Thanks in advance for your help...

Dan Rabosky
Department of Ecology and Evolutionary Biology &
Fuller Evolutionary Biology Program
Cornell Lab of Ornithology
Cornell University
Ithaca, NY 14853-2701




	[[alternative HTML version deleted]]

?try

This will catch the error and let you take alternative action.

On 9/21/07, Dan Rabosky <dlr32 at cornell.edu>
wrote:
>
> Hello -
>
> I am having a problem with the function 'integrate'.  I am running
R
> on OSX (R 2.5.1).
>
> I am trying to suppress the error message when 'integrate' attempts
> to integrate across a parameter set giving a non-finite function
> value.  I'm using it in a MCMC / simulated annealing algorithm, and
> it is entirely possible that some parameter sets will give non-
> integrable functions.  I thought that the argument
>
> stop.on.error = FALSE
>
> would cause the function to return NA or at least not return an error
> (I can work around the situation where integrate returns garbage, but
> if it generates an error, it terminates my chain and all appears lost).
>
> This may not be necessary, but here is a simplification of an
> instance that generates the error.  If you send any B > 500 to
> "questionableFunction", the integration fails and returns an
error
> message - even with stop.on.error = FALSE.  Is there any way I can
> work around this?  Perhaps there is something similar to
> "suppressWarnings()" that I might be able to use?
>
> questionableFunction <- function(B = 10)
> {
>        fx4 <- function(B, y, z){
>                1 + exp(B*y - B*z);
>         }
>
>        fx1 <- function(x, r0 = 0.12, k=-0.10, a=0.5, tshift=8.36){
>                (r0*exp(r0*t0 + (k/B)*log(fx4(B, t0, tshift)) - r0*x -
(k/B)*log(fx4
> (B, x, tshift))))
>        }
>
>        integrate(fx1, 8.36, 10, stop.on.error = FALSE);
> }
>
>
>
> Thanks in advance for your help...
>
> Dan Rabosky
> Department of Ecology and Evolutionary Biology &
> Fuller Evolutionary Biology Program
> Cornell Lab of Ornithology
> Cornell University
> Ithaca, NY 14853-2701
>
>
>
>
>        [[alternative HTML version deleted]]
>
> ______________________________________________
> R-help at r-project.org 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.
>

-- 
Jim Holtman
Cincinnati, OH
+1 513 646 9390

What is the problem you are trying to solve?