simest: Simulation test hmltm performance

In DistanceDevelopment/hsltm: Line transect estimation with hidden state availability models

Description

Simulates line transect survey with stochastic animal availability, estimates detection probability and related parameters, and reports bias, variance of estimators. Currently only for models with no covariates.

Usage

simest(simethod = "hmm", shfun, spars, ehfun = shfun,
  parstart = spars, survey.pars, shmm.pars, ehmm.pars = shmm.pars,
  animal = NULL, adat = NULL, control.opt, control.fit, nsim = 50,
  report.progress = TRUE, En = 100, doplots = FALSE,
  varest = FALSE, hmmpars.bs = NULL, print.n = FALSE,
  silent = FALSE, nx = 100)

Arguments

simethod	either "hmm" (for availability simulation with HMM) or "animals" (for availability simulation by resampling availability or depth dime series).
shfun	detection hazard function name for simulation.
spars	detection hazard function parameters for simulation.
ehfun	detection hazard function name for estimation.
parstart	detection hazard function parameter start vallues for estimation.
survey.pars	survey parameters, as for est.hmltm.
shmm.pars	hmm.pars list (as for est.hmltm) to be used in simulating.
ehmm.pars	hmm.pars list (as for est.hmltm) to be used in estimating.
animal	which animals' availability data to use (only need if simethod="animals").
adat	availability data (only need if simethod="animals"): list of m>=1 vectors containing availability binary time series or depths time series.
control.opt	optimization parameters passed to optim, as for est.hmltm.
control.fit	fit parameters, as for est.hmltm.
nsim	number of simulations to do.
report.progress	if TRUE, reports progress as does simulations.
En	Mean sample size for simulations.
doplots	if TRUE does plots of simulaton results.
varest	if TRUE, estimates variance (using Hessian matrix).
hmmpars.bs	output from hmmpars.boot; required if simethod="hmm" and the availability HMM parameters are not to be treated as fixed, in which case the availability HMM parameters for each simulation are obtained by sampling with replacement from hmmpars.bs. If hmmpars.bs is NULL, the availability HMM parameters are treated as fixed and equal to ehmm.pars.
print.n	if TRUE, prints sample size for each simulation.
silent	parameter of try, controlling error reporting.
nx	number of perpendicular distances intervals to use in evaluating detection probability, p(x).

Value

list with these elements:

• N: population size.

• esw: true effective strip half-width.

• p0: true p(0).

• p.: true mean detection probability.

• meann: mean sample size across simulations.

• biasNhat: bias of estimated N.

• biaseswhat: bias of estiamted esw.

• biasp0hat: bias of estimated p(0).

• biasphat: bias of estimated mean detection probability.

• biasinvphat: bias of estimated inverse mean detection probability.

• n: sample sizes.

• parest: parameter estimates from each simulation.

• p0hat: p(0) estimates from each simulation.

• phat: mean detection probability estimates from each simulation.

• invphat: inverse mean detection probability estimates from each simulation.

• Nhat: abunance estimates from each simulation.

• invpse: estimated standard error of inverse mean detection probability estimates from each simulation (if varest=TRUE).

• p0se: estimated standard error of p(0) estimates from each simulation (if varest=TRUE).

DistanceDevelopment/hsltm documentation built on June 21, 2019, 2:22 p.m.