R help - Mar 2010 - Question regarding to maxNR

Hi R-users,

Recently, I use maxNR function to find maximizer. I have error appears as
follows
Error in maxNRCompute(fn = fn, grad = grad, hess = hess, start = start,  : 
  NA in the initial gradient

My code is 

mu=2
s=1
n=300
library(maxLik)
set.seed(1004)
x<-rcauchy(n,mu,s)
loglik<-function(mu)
{
log(prod(dcauchy(x,mu,s)))
}
maxNR(loglik,start=median(x))$estimate 


Does anyone know how to solve this problem?

Thanks,

Kate


	[[alternative HTML version deleted]]

Hi Kate!

On Fri, Mar 12, 2010 at 6:20 AM, kate <yhsu6 at illinois.edu>
wrote:
> Hi R-users,
>
> Recently, I use maxNR function to find maximizer. I have error appears as
follows
> Error in maxNRCompute(fn = fn, grad = grad, hess = hess, start = start, ?:
> ?NA in the initial gradient
>
> My code is
>
> mu=2
> s=1
> n=300
> library(maxLik)
> set.seed(1004)
> x<-rcauchy(n,mu,s)
> loglik<-function(mu)
> {
> log(prod(dcauchy(x,mu,s)))
> }
> maxNR(loglik,start=median(x))$estimate
>
>
> Does anyone know how to solve this problem?
Yes :-)

Algebraically, "log(prod(z))" is equal to "sum(log(z))" but
these two
expressions might return different numbers on digital computers:

R> ll<-function(mu) { sum(log(dcauchy(x,mu,s))) }
R> ll(2)
[1] -754.4928
R> loglik(2)
[1] -Inf

R> maxNR(ll,start=median(x))$estimate
[1] 2.075059

Best wishes from Copenhagen,
Arne

-- 
Arne Henningsen
http://www.arne-henningsen.name

Your problem is numerical.

Try replacing

log(prod(dcauchy(x,mu,s)))

by

sum(log(dcauchy(x,mu,s)))

and see the difference.  Here's what I get:
> mu <- 2
> s <- 1
> n <- 300
> library(maxLik)
> set.seed(1004)
> x <- rcauchy(n,mu,s)
> loglik <- function(mu) {
+   sum(log(dcauchy(x,mu,s)))
+ }
> maxNR(loglik,start=median(x))$estimate 
[1] 2.075059

which looks about right to me.

Bill Venables
CSIRO/CMIS Cleveland Laboratories


-----Original Message-----
From: r-help-bounces at r-project.org [mailto:r-help-bounces at r-project.org]
On Behalf Of kate
Sent: Friday, 12 March 2010 3:21 PM
To: r-help at r-project.org
Subject: [R] Question regarding to maxNR

Hi R-users,

Recently, I use maxNR function to find maximizer. I have error appears as
follows
Error in maxNRCompute(fn = fn, grad = grad, hess = hess, start = start,  : 
  NA in the initial gradient

My code is 

mu=2
s=1
n=300
library(maxLik)
set.seed(1004)
x<-rcauchy(n,mu,s)
loglik<-function(mu)
{
log(prod(dcauchy(x,mu,s)))
}
maxNR(loglik,start=median(x))$estimate 


Does anyone know how to solve this problem?

Thanks,

Kate


	[[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.