R help - Mar 2009 - error statement: missing value where TRUE/FALSE needed

Hi list,

I want to try Gibbs sampling as a method of estimating a markov-switching 
model of a mean-deviating, pth-order autoregressive process with time 
varying transition probabilities via R and am using a code originally 
written by another person; I attach the useful pdf document explaining the 
code. When I run the code, I get an error message:

Error in if (r < vQ[i]) { : missing value where TRUE/FALSE needed

I am using R.2.6.0 on Windows XP. Can anyone tell me what the error message 
means?

The specific line in the code is:

if (r<vQ[i]){ vS[i]=0 } else vS[i]=1 

The code is available at: 
http://www.michael-curran.com/gibbs.html
and the two data sets are available at:
http://www.geocities.jp/atsmatsumoto/ci.txt
and
http://www.geocities.jp/atsmatsumoto/callrate.txt

Note: I saved the first data set as ci.txt and the second as boj.txt and so 
these are the file names that the code loads. If anyone uses the code and 
the data and manages to get it to work, I would gratefully appreciate if 
they could tell me what adjustments they made to the code in order to do 
so.

Kind regards,

Michael
-- 
Michael Curran
Candidate for the MPhil in Economics
Cambridge University
http://www.michael-curran.com/



-------------- next part --------------
A non-text attachment was scrubbed...
Name: gibbsms-1.pdf
Type: application/pdf
Size: 175239 bytes
Desc: gibbsms-1.pdf
URL:
<https://stat.ethz.ch/pipermail/r-help/attachments/20090323/5cc81676/attachment-0003.pdf>

I would assume the expression in the 'if' is giving an NA as a result:
> if (NA) 1 else 2
Error in if (NA) 1 else 2 : missing value where TRUE/FALSE
needed
>

On Mon, Mar 23, 2009 at 10:38 AM, Michael Curran <mpc46 at cam.ac.uk>
wrote:
> Hi list,
>
> I want to try Gibbs sampling as a method of estimating a markov-switching
> model of a mean-deviating, pth-order autoregressive process with time
> varying transition probabilities via R and am using a code originally
> written by another person; I attach the useful pdf document explaining the
> code. When I run the code, I get an error message:
>
> Error in if (r < vQ[i]) { : missing value where TRUE/FALSE needed
>
> I am using R.2.6.0 on Windows XP. Can anyone tell me what the error message
> means?
>
> The specific line in the code is:
>
> if (r<vQ[i]){ vS[i]=0 } else vS[i]=1
> The code is available at: http://www.michael-curran.com/gibbs.html
> and the two data sets are available at:
> http://www.geocities.jp/atsmatsumoto/ci.txt
> and
> http://www.geocities.jp/atsmatsumoto/callrate.txt
>
> Note: I saved the first data set as ci.txt and the second as boj.txt and so
> these are the file names that the code loads. If anyone uses the code and
> the data and manages to get it to work, I would gratefully appreciate if
> they could tell me what adjustments they made to the code in order to do
so.
>
> Kind regards,
>
> Michael
> --
> Michael Curran
> Candidate for the MPhil in Economics
> Cambridge University
> http://www.michael-curran.com/
>
>
>
>
> ______________________________________________
> 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 that you are trying to solve?

Hi


r-help-bounces at r-project.org napsal dne 23.03.2009 15:38:49:
> Hi list,
> 
> I want to try Gibbs sampling as a method of estimating a 
markov-switching 
> model of a mean-deviating, pth-order autoregressive process with time 
> varying transition probabilities via R and am using a code originally 
> written by another person; I attach the useful pdf document explaining 
the 
> code. When I run the code, I get an error message:
> 
> Error in if (r < vQ[i]) { : missing value where TRUE/FALSE needed
> 
> I am using R.2.6.0 on Windows XP. Can anyone tell me what the error 
message 
> means?
The error message means that there is missing value instead of logical 
when you are calling if
> a<-NA
> b<-1
> a<b
[1] NA
> if(a<b) print("No.way")
Error in if (a < b) print("No.way") : 
  missing value where TRUE/FALSE needed
> a<-0
> if(a<b) print("No.way")
[1] "No.way"

So test r and vQ if they contain missing values or change your code

Besides, it looks like code written in Basic not in R

Few other comments

* read.table() already produce data frame

* This is really strange, I did not see such construction yet
> mG0 <-rbind(cbind(100,0,0,0), cbind(0,100,0,0), cbind(0,0,100,0), 
cbind(0,0,0,100))

what about
> x<-matrix(0, 4,4)
> diag(x)<-100
> x
     [,1] [,2] [,3] [,4]
[1,]  100    0    0    0
[2,]    0  100    0    0
[3,]    0    0  100    0
[4,]    0    0    0  100

* Better upgrade to 2.8.1, or maybe wait about a month for 2.9.0

* I do not think that somebody will go through whole your code, it is 
terribly complicated.

Regards
Petr

>
> 
> The specific line in the code is:
> 
> if (r<vQ[i]){ vS[i]=0 } else vS[i]=1 
> 
> The code is available at: 
> http://www.michael-curran.com/gibbs.html
> and the two data sets are available at:
> http://www.geocities.jp/atsmatsumoto/ci.txt
> and
> http://www.geocities.jp/atsmatsumoto/callrate.txt
> 
> Note: I saved the first data set as ci.txt and the second as boj.txt and 
so 
> these are the file names that the code loads. If anyone uses the code 
and 
> the data and manages to get it to work, I would gratefully appreciate if 
> they could tell me what adjustments they made to the code in order to do 
> so.
> 
> Kind regards,
> 
> Michael
> -- 
> Michael Curran
> Candidate for the MPhil in Economics
> Cambridge University
> http://www.michael-curran.com/
> 
> 
> 
> [p??loha gibbsms-1.pdf odstran?na u?ivatelem Petr PIKAL/CTCAP] 
> ______________________________________________
> 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.