runsims: Run mixed stock simulations

In bbolker/mixstock: Functions for Mixed Stock Analysis

Run mixed stock simulations

Description

Run multiple simulations of a mixed stock systems with specified marker frequencies and source contributions, running multiple estimations (bootstrap samples or MCMC chains) for each simulation

Usage

runsims(sim.n=10, mc.n=10, totsamp=200, which="all",
true.freq=matrix(c(0.65, 0.31, 0.01, 0.01, 0.01, 0.01,
               0.31,0.65,0.01,0.01,0.01,0.01),ncol=2),
       true.contrib=c(0.9, 0.1), est="MCMC", verbose=FALSE,
fuzz=0.001, nboot=1000, bootrpt=20, minmarks=3)

Arguments

sim.n	Number of simulations to run
mc.n	Number of bootstrap/MCMC chains to run for each simulation
totsamp	total sample size (to be distributed half in mixed stock and half, evenly, among sources)
which	which markers to use: "common", "rare", or "all"
true.freq	matrix of true marker frequencies in the sources (column=source, row=marker)
true.contrib	vector of true contributions of sources to the mixed stock
est	estimation method: "MCMC" (Markov Chain Monte Carlo), "cml" (conditional max. likelihood), or "uml" (unconditional max. likelihood)
verbose	produce lots of debugging output?
fuzz	"fuzz" parameters for (e.g.) keeping estimated values away from 0/1
nboot	Number of bootstrap samples/length of MCMC chain
bootrpt	Frequency for reporting on the progress of bootstrap code
minmarks	Minimum number of markers to allow for a simulation

Value

Array of the results from all simulations and bootstrap samples

Examples

## mild kluge to drop unneeded dimensions
  runsims(sim.n=1,mc.n=1,nboot=100,est="cml")[,,]
  

bbolker/mixstock documentation built on July 23, 2024, 12:18 p.m.