occupancy: fit occupancy-detection models
In jdyen/occupancy: Bayesian Occupancy Detection Models in R

Description Usage Arguments Details Value Examples

occupancy is a function to fit occupancy-detection models in JAGS from within R. Models are specified with a formula interface and are supported by methods to summarise, visualise, validate, and predict from fitted models.

occupancy lets you fit occupancy-detection models in JAGS without having to prepare JAGS code. Models are specified with a formula interface and can incorporate random effects. Models are supported by several methods to summarise, visualise, validate, and predict from fitted models.

1 2	occupancy(formula_occ, formula_detect, site_id, survey_id, data, jags_settings = list())

`formula_occ`	model formula for occupancy component of the model. A two-sided formula with the response variable on the left of the `~` operator and the predictor variables on the right. Multiple predictor variables are separated by `+` operators. Random effects are included with vertical bars, using the notation `(1 \| group)` to specify a random intercept for the variable `group`. More complex random structures (e.g., random slopes) are not supported.
`formula_detect`	model formula for detection component of the model. A one-sided formula with predictor variables on the right, formatted as for `formula_occ`.
`site_id`	the name of the column in `data` in which site identifiers are recorded.
`survey_id`	the name of the column in `data` in which survey identifiers are recorded.
`data`	a `data.frame` in long format containing data on all variables. Required variables include the response (detection-nondetection data), site and survey identifiers (see `site_id` and `survey_id`, above), and predictor variables. Column names must match the names used in `formula_occ` and `formula_detect`.
`jags_settings`	optional list of MCMC settings. Any or all items can be altered if needed. Options are: `n_iter` the total number of MCMC iterations (including burn-in) `n_burnin` the number of MCMC iterations to discard as a burn-in `n_chains` the number of MCMC chains `n_thin` thinning rate of MCMC samples `parallel` logical, should chains be run in parallel? `modules` JAGS modules to load `params` character vector of parameters to store `seed` seed used to initialise MCMC chains

This function fits an occupancy-detection model in JAGS from two formulas: formula_occ and formula_detect. Occupancy-detection models separate the two processes of being present at a site and being detect given presence at a site. This requires data from repeated visits (surveys) to sites.

The occupancy component of the model (presence at a site) is defined at the site level. The detection component of the model (detections given presence) is defined at the survey level. The model assumes that associations between occupancy/detection and predictor variables are linear on a logit scale.

occupancy_model - a list object that can be analysed using functions described in methods.

## Not run: 

# fit a model to simulated data
mod <- occupancy(response ~ occ_predictor1 + occ_predictor2 + 
                    (1 | occ_random1) + (1 | occ_random2),
                  ~ detect_predictor1 + detect_predictor2 + 
                   (1 | detect_random1),
               site_id = "site",
               survey_id = "survey",
               data = occupancy_data,
               jags_settings = list(n_iter = 1000, n_burnin = 500, n_thin = 2))
               
# plot the model coefficients
par(mfrow = c(2, 1))
plot(mod)

# extract the model coefficients
coef(mod)

# check model fit
calculate_metrics(mod)
     

## End(Not run)

## Not run: 
# a simple occupancy-detection model for artificial data

# build and sample

# plot coefficients

# validate


## End(Not run)