Description Usage Arguments Details Value Examples
View source: R/MSOcc_mod_deprecated.R
This function fits the Bayesian multi-scale occupancy model described by Dorazio and Erickson (2017) using the polya-gamma data augmentation strategy described by Polson et al. (2012). Note that this documentation assumes there are M sites, J_i samples within each site, and K_{ij} replicates from each sample.
1 2 3 4 5 6 7 8 9 10 11 12 13 | msocc_mod(
wide_data,
site = list(model = ~1, cov_tbl),
sample = list(model = ~1, cov_tbl),
rep = list(model = ~1, cov_tbl),
priors = list(site = list(mu0 = 0, Sigma0 = 9), sample = list(mu0 = 0, Sigma0 = 9),
rep = list(mu0 = 0, Sigma0 = 9), a0 = 1, b0 = 1),
num.mcmc = 1000,
progress = T,
print = NULL,
seed = NULL,
beta_bin = T
)
|
wide_data |
object of class |
site |
object of class
|
sample |
object of class
|
rep |
object of class
|
priors |
object of class
|
num.mcmc |
number of MCMC samples |
progress |
should sampling progress be printed? |
print |
interval for printing; defaults to 5 percent of |
seed |
optional seed for reproducible samples |
beta_bin |
optional; should a beta-binomial sampler be used when possible? This option is considerably faster. |
This function fits the multi-scale occupancy model described by Dorazio and Erickson (2017). However, this function implements a fully Bayesian sampler based on the data augmentation strategy described by Polson et al. (2012)
object of class list
containing the following elements:
beta
an object of class matrix
of samples from the joint
posterior distribution of the regression coefficients at the site level
psi
an object of class numeric
of samples from the joint
posterior distribution of the site-level presence probability, psi. Note
that this is only returned if beta_bin
is TRUE
and
site$model = ~ 1
alpha
an object of class matrix
of
samples from the joint posterior distribution of the regression coefficients
at the sample level
theta
an object of class matrix
of
samples from the joint posterior distribution of the sample-level presence
probability, theta. Note that this is only returned if beta_bin
is
TRUE
and sample$model = ~ 1
or sample$model = ~ site
delta
an object of class matrix
of samples from the joint
posterior distribution of the regression coefficients at the replicate level
p
an object of class numeric
of sample from the posterior
distribution of the replicate level detection probability. NOte that this is
only returned if beta_bin
is TRUE
and rep$model = ~ 1
.
model.info
an object of class list containing the following
elements:
X
design matrix for site level predictors
W
design matrix for sample level predictors
V
design matrix for replicate level predictors
M
number of sites
J
vector of number of samples per site
K
vector of number of replicates per site-sample combination
z
matrix of posterior samples of latent site-presence z
z.vec
matrix of posterior samples of latent site-presence
stretched across samples
A
matrix of posterior samples of latent sample-presence A
Y
vector of binomial responses (aggregated at sample level)
y
vector of Bernoulli responses (stretched across site-sample
combination)
site_mod
model statement for site-level predictors
samp_mod
model statement for sample-level predictors
rep_mod
model statement for replicate-level predictors
num.mcmc
number of MCMC samples run
beta_bin
was beta-binomial sampler implemented if possible?
df
empty data.frame
of design in long format
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 | data(fung)
# prep data
fung.detect <- fung %>%
dplyr::select(1:4)
site.df <- fung %>%
dplyr::select(-sample, -pcr1, -pcr2) %>%
dplyr::distinct(site, .keep_all = TRUE) %>%
dplyr::arrange(site)
sample.df <- fung %>%
dplyr::select(-pcr1, -pcr2) %>%
dplyr::arrange(site, sample)
# fit intercept model at all three levels use beta-binomial sampler
fung_mod1 <- msocc_mod(wide_data = fung.detect, progress = T,
site = list(model = ~ 1, cov_tbl = site.df),
sample = list(model = ~ 1, cov_tbl = sample.df),
rep = list(model = ~ 1, cov_tbl = sample.df), # covariates aggregated at sample level
num.mcmc = 1000, beta_bin = T)
# model sample level occurence by frog density
fung_mod2 <- msocc_mod(wide_data = fung.detect, progress = T,
site = list(model = ~ 1, cov_tbl = site.df),
sample = list(model = ~ frogs, cov_tbl = sample.df),
rep = list(model = ~ 1, cov_tbl = sample.df),
num.mcmc = 1000, beta_bin = T)
|
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.