R help - Jan 2011 - MCMC object indexing

I have an mcmc object and I''m trying to plot the quantiles of the
variables - and not as a function of the iterations as in cumuplot.

I cannot seem to find the right combination of indexing to access the variables;
after which I''m sure I can plot all the statistics I could hope for.

Any hints for accessing the mcmc object would be appreciated.

=Dave

	[[alternative HTML version deleted]]

Robinson, David G <drobin <at> sandia.gov> writes:
> 
> I have an mcmc object and I'm trying to plot the quantiles of the 
> variables - and not as a function of the
> iterations as in cumuplot.
> 
> I cannot seem to find the right combination of indexing to 
> access the variables; after which I'm sure I can
> plot all the statistics I could hope for.
> 
> Any hints for accessing the mcmc object would be appreciated.
> 
> =Dave
  A reproducible example would be very useful.  In general
I find that I can just index an mcmc object as though it
were a matrix, e.g.:

z <- cbind(a=runif(20),b=runif(20),c=runif(20))
library(coda)
m <- mcmc(data=z)
t(apply(m,2,quantile,c(0.025,0.975)))

  If that sort of thing doesn't work for you, please
post an example of what doesn't work ...

  Ben Bolker

Stoker, Ben wrote:
> A reproducible example would be very useful.  In general
> I find that I can just index an mcmc object as though it
> were a matrix, e.g.:
> 
> z <- cbind(a=runif(20),b=runif(20),c=runif(20))
> library(coda)
> m <- mcmc(data=z)
> t(apply(m,2,quantile,c(0.025,0.975)))
> 
>   If that sort of thing doesn't work for you, please
> post an example of what doesn't work ...
> 
>   Ben Bolker
Ben, thanks for the suggestion and it got me going the right direction.
FWIW-I did get an error from your suggestion, but it led me to find that I
had an mcmc.list object and not an mcmc object.  Once I got past that, your
suggestion worked perfectly. Thanks!