sim.popn.secrgam: Generate a population using any density formula

In david-borchers/secrgam: Spatially Explicit Capture-Recapture with Generalised Additive Density Models

Description

description...

Usage

sim.popn.secrgam(fit = NULL, N = NULL, Dmodel = NULL, Dpars = NULL,
  Dlink = "log", mask = NULL)

Arguments

fit	a fitted secrgam or secr model
N	average abundance (converts density surface to relative density)
Dmodel	density model formula
Dpars	associated parameters for the density model
Dlink	link function (default is "log")
mask	mask object (as used by secrgam.fit and secr.fit)
log	if TRUE the log link is used, if FALSE the identity link is used

Details

details...

Value

value...

Examples

data(Boland.leopards1)
data(Boland.fits1)

# use a fitted model
popn = sim.popn.secrgam(fit1.a3) ; head(popn) ; dim(popn)
plot(fit1.a3, type = "density", asp = 1)
points(popn, cex = 0.5, pch = 19)

# use an inflated density surface from a fitted model
popn = sim.popn.secrgam(fit1.a3, N = 1000) ; head(popn) ; dim(popn)
plot(fit1.a3, type = "density", asp = 1)
points(popn, cex = 0.5, pch = 19)

## Not run: 
# use an alternative model
# e.g. construct a tensor product model to represent a bivariate normal
require(mvtnorm)
D = dmvnorm(Boland.mask1, apply(Boland.mask, 2, mean), 10e7 * diag(2)) * 10e6
N = sum(D) * attributes(Boland.mask)$area ; N # expected sample size
data = cbind(z = D, Boland.mask)
prep4image(data, key = FALSE, col = topo.colors(10), asp = 1)
model = D ~ te(x, y, k = 3)
fit = gam(model, gaussian(link = "log"), data, fx = TRUE)
prep4image(data.frame(x = data$x, y = data$y, z = fitted(fit)), key = FALSE, asp = 1)
popn = sim.popn.secrgam(Dmodel = model, Dpars = coef(fit), mask = Boland.mask1) ; head(popn) ; dim(popn)
points(popn, cex = 0.5, pch = 19)

## End(Not run)

david-borchers/secrgam documentation built on May 14, 2019, 9:30 p.m.