tmcmc {mixstock} | R Documentation |
Runs a Gibbs sampler MCMC starting with 95% contribution from each source, then combines the chains
tmcmc(data, n.iter=20000, rseed=1001, n.burnin=floor(n.iter/2), n.thin=max(1, floor(n.chains * (n.iter - n.burnin)/1000)), verbose=FALSE, fprior=NULL, contrib.only=TRUE, rptiter=-1, outfile=NULL, lang="C",a=NULL,gr=FALSE) gibbs(sourcesamp, mixsamp, a = 1, startiter, maxiter, startfval = NULL, n.thin = 1, fprior = NULL, rptiter = -1)
data |
Data: a mixstock.data object |
n.iter |
Total length of each chain |
n.burnin |
Number of burn-in iterations |
n.thin |
thinning rate. Must be a positive integer. Set 'n.thin' > 1 to save memory and computation time if 'n.iter' is large. Default is 'max(1, floor(n.chains * (n.iter-n.burnin) / 1000))' which will only thin if there are at least 2000 simulations. |
rseed |
Random-number seed |
verbose |
Produce lots of output |
fprior |
Bayesian prior for source contributions |
contrib.only |
To save memory, store only information about contributions from each source and not about the estimated marker frequencies in each source |
rptiter |
How often to issue a progress report. Negative numbers mean no reports |
outfile |
file to use for output |
lang |
Run the chain in C or R (for debugging/testing purposes only)? |
a |
prior strength parameter |
gr |
calculate Gelman-Rubin convergence statistic? |
sourcesamp |
matrix of marker samples from sources |
mixsamp |
vector of marker samples from mixed stock |
startiter |
starting iteration |
maxiter |
max. number of iterations |
startfval |
starting source contributions |
Returns an object of type mixstock.est
Ben Bolker
Masuda and Pella
data(bolten98) b98c <- markfreq.condense(as.mixstock.data(bolten98)) t1 <- tmcmc(b98c); t1