Displaying 2 results from an estimated 2 matches for "modmcmc".
Did you mean:
mermcmc
2011 Aug 04
1
use of modMCMC
...ov.unscaled) : 'V' is not a square numeric matrix
In addition: Warning message:
In summary.modFit(Fit) : Cannot estimate covariance; system is singular
This is due becasue the Hessian matrix has all the entries equal to 0.
In these cases, on the help page of modFit, it is suggested to use modMCMC
to generate new sets of parameters. modMCMC performs a Markov Chain Monte
Carlo simulation.
I do not understand very well how modMCMC can be used in a context of
parameter estimation. Could someone help me in understanding the use of this
function and its utility for parameter fiting?
Thank you v...
2013 May 02
0
modMCMC runs in FME package
...d) <- c("time", "y")
obj <- function(x, parset = names(x)) {
pars[parset] <- x
tout <- seq(0, 50, by = 1)
out <- model(pars, tout)
return(modCost(obs = observed, model = out))
}
prior <- function(p)
return( sum(((p-c(0.06,18))/c(0.04,6))^2 ))
final <- modMCMC(p = c(k=0.06,v=18), f = obj, prior=prior,lower=c(0.0001,0.1),jump = NULL, niter = 1000,updatecov=100, wvar0 = 1,var0=NULL,burninlength = 50)
summary(final)
plot(final)
the values for "observed" were generated with k=0.05, v=20, which are the retunrs I expect from the modMCMC run. Please...