Latent occupancy z_i ~ Bernoulli(psi_i) with logit psi = X_psi beta_psi,
observed through J replicate visits emitting a detection state y in {0,1,2}:
at an occupied site a visit detects with probability p11 and the detection is
certain (state 2) with probability b else ambiguous (state 1); at an
unoccupied site a visit yields a false-positive ambiguous detection (state 1)
with probability p10. Returns an N x J integer matrix in {0, 1, 2}
suitable for tobs() with fp_occu().
Usage
simulate_fp_occu(
N = 300,
J = 5,
n_occ_covs = 1,
beta_psi = NULL,
p11 = 0.6,
p10 = 0.05,
b = 0.5,
seed = NULL
)Arguments
- N
Number of sites (default 300).
- J
Number of visits (default 5).
- n_occ_covs
Number of occupancy covariates (default 1).
- beta_psi
Occupancy coefficients (logit). Default
c(qlogis(0.5), runif(n_occ_covs, -0.6, 0.6)).- p11, p10, b
True detection, false-positive, and certain-classification probabilities (scalars; defaults 0.6, 0.05, 0.5).
- seed
Optional random seed.
Value
A list with y (N x J state matrix), data (occupancy covariates),
and truth (coefficients, per-site psi, scalar p11/p10/b, latent z).
Examples
sim <- simulate_fp_occu(N = 50, J = 4, seed = 1)
table(sim$y)