ubms

Description

Unmarked Bayesian Models using Stan

Author(s)

Ken Kellner

Get Information for a Restricted Spatial Regression Model

Description

A call to RSR in the formula for a state process adds an autocorrelated spatial random effect to the model in the form of a Restricted Spatial Regression (RSR). For examples of RSRs applied to ecological data, see Johnson et al. (2013) and Broms et al. (2014). The function can also be used outside a formula to visualize the spatial relationships between sites in your data and choose an appropriate distance threshold below which two sites will be considered neighbors, and thus potentially correlated, for the RSR model. For more details, see the example vignette: vignette("spatial-models", package = "ubms")

Usage

RSR(x, y = NULL, threshold, moran_cut = NULL, plot_site = NULL)

Arguments

Value

Either a list of spatial matrices used for the RSR (only useful internally to ubms), or if plot_site is an integer, a ggplot object.

References

Broms KM, Johnson DS, Altwegg R, Conquest LL. 2014. Spatial occupancy models applied to atlas data show Southern Ground Hornbills strongly depend on protected areas. Ecological Applications 24: 363-374.

Johnson DS, Conn PB, Hooten MB, Ray JC, Pond BA. 2013. Spatial occupancy models for large data sets. Ecology 94: 801-808.

Examples

# Generate some coordinates
x <- runif(100,0,10)
y <- runif(100,0,10)
# Show neighbors of site 10 when threshold is 3 units
RSR(x, y, threshold=3, plot_site=10)

Extract a Submodel from a ubmsFit Model

Description

Extract a Submodel from a ubmsFit Model

Usage

## S4 method for signature 'ubmsFit,character,missing,missing'
x[i]

Arguments

Value

An object of class ubmsSubmodel.

Extract a ubmsSubmodel From a ubmsSubmodelList Object

Description

Extract a ubmsSubmodel From a ubmsSubmodelList Object

Usage

## S4 method for signature 'ubmsSubmodelList,character,missing,missing'
x[i]

Arguments

Value

An object of class ubmsSubmodel.

Extract Coefficient Values From a ubmsFit Model

Description

Extract Coefficient Values From a ubmsFit Model

Usage

## S4 method for signature 'ubmsFit'
coef(object, ...)

Arguments

Value

A vector of coefficient values for all submodels.

Extract Samples From a ubmsFit Model

Description

Extract samples from a ubmsFit model

Usage

## S4 method for signature 'ubmsFit'
extract(object, pars, permuted = TRUE, inc_warmup = FALSE, include = TRUE)

Arguments

Value

If permuted=TRUE, a list; if permuted=FALSE, an array.

Extract Pointwise Log-likelihood From Model

Description

Extracts the pointwise log-likelihood matrix or array from a model. This is useful as an input for functions in the loo package. If called on a ubmsFit object, the log-likelihood matrix or array is calculated using the posterior distributions of the parameters; the log-likelihood values are not actually saved inside the model object. If called on a stanfit object, loo::extract_log_lik is used. In this case, the log-likelihood must be saved as one of the output parameters in Stan.

Usage

extract_log_lik(object, parameter_name = "log_lik", merge_chains = TRUE)

Arguments

Value

A matrix (samples x sites) or array (samples x chains x sites)

Create a List of ubmsFit Models

Description

Create a list of ubmsFit models

Usage

## S4 method for signature 'ubmsFit'
fitList(...)

Arguments

Value

An object of class ubmsFitList containing the list of models

Extract Fitted Values

Description

Extract fitted values for a given submodel from a ubmsFit object. Fitted values are calculated separately for each submodel using the posterior predictive distribution of the latent state z, following Wright et al. (2019).

Usage

## S4 method for signature 'ubmsFit'
fitted(object, submodel, draws = NULL, ...)

Arguments

Value

A matrix of fitted values with dimension draws by observations. Note that calculation of fitted values for the detection submodel is conditional on z > 0, so fitted values for an observation in a posterior draw where z = 0 are assigned value NA (Wright et al. 2019).

References

Wright, W. J., Irvine, K. M., & Higgs, M. D. (2019). Identifying occupancy model inadequacies: can residuals separately assess detection and presence? Ecology, 100(6), e02703.

Extract y, the Response Variable, From a ubmsFit Model

Description

Extract y, the Response Variable, From a ubmsFit Model

Usage

## S4 method for signature 'ubmsFit'
getY(object)

Arguments

Value

A matrix containing the response variable y.

Get Model Runtime

Description

Get warmup and sampling time from a ubmsFit object

Usage

## S4 method for signature 'ubmsFit'
get_elapsed_time(object, ...)

Arguments

Value

A matrix with one row per chain and two columns, containing the warmup time and sampling time, respectively, in seconds

Get Stan Code From Model

Description

Get the Stan code used to run a model as a character string

Usage

## S4 method for signature 'ubmsFit'
get_stancode(object, ...)

Arguments

Details

Pass the result of get_stancode to cat to get the code in a more readable format. Note that the output in most cases is Stan code that can be used to fit several types of models, and not all Stan code will be used in all models.

Value

A character string with the model code

Check model goodness-of-fit

Description

Goodness-of-fit tests for ubmsFit models using posterior predictive checks

Usage

gof(object, draws = NULL, ...)

## S4 method for signature 'ubmsFitOccu'
gof(object, draws = NULL, quiet = FALSE, ...)

## S4 method for signature 'ubmsFitAbun'
gof(object, draws = NULL, quiet = FALSE, ...)

Arguments

Value

An object of class ubmsGOF containing statistics calculated from the posterior predictive distribution.

Functions

• gof(ubmsFitOccu): Applies the MacKenzie-Bailey chi-square goodness of fit test for ocupancy models (MacKenzie and Bailey 2004).

• gof(ubmsFitAbun): A goodness-of-fit test for N-mixture type models based on Pearson's chi-square.

References

MacKenzie, D. I., & Bailey, L. L. (2004). Assessing the fit of site-occupancy models. Journal of Agricultural, Biological, and Environmental Statistics, 9(3), 300-318.

K-fold Cross-validation of a ubmsFit Model

Description

Randomly partition data into K subsets of equal size (by site). Re-fit the model K times, each time leaving out one of the subsets. Calculate the log-likelihood for each of the sites that was left out. This function is an alternative to loo (leave-one-out cross validation).

Usage

## S4 method for signature 'ubmsFit'
kfold(x, K = 10, folds = NULL, quiet = FALSE, ...)

Arguments

Value

An object of class elpd_generic that is compatible with loo::loo_compare

Leave-one-out Cross Validation

Description

Leave-one-out Cross Validation

Usage

## S4 method for signature 'ubmsFit'
loo(x, ..., cores = getOption("mc.cores", 1))

Arguments

Value

A named list of class loo with estimates of the expected log predictive density and other parameters used for model comparison. See ?loo::loo for more information.

Model Selection For a List of ubmsFit Models

Description

Construct a model selection table from a ubmsFitList

Usage

## S4 method for signature 'ubmsFitList'
modSel(object, ...)

Arguments

Value

A data.frame of model fit information with one row per model in the input ubmsFitList. Models are ranked in descending order by expected log pointwise predictive density (elpd).

See Also

loo, loo_compare

Get Parameter Names From a ubmsFit Model

Description

Get Parameter Names From a ubmsFit Model

Usage

## S4 method for signature 'ubmsFit'
names(x)

Arguments

Value

A character vector of parameter names.

Get Names of Models in a ubmsFitList

Description

Get Names of Models in a ubmsFitList

Usage

## S4 method for signature 'ubmsFitList'
names(x)

Arguments

Value

A character vector of model names.

Get number of Posterior Samples Stored in a ubmsFit Model

Description

Get number of Posterior Samples Stored in a ubmsFit Model

Usage

## S4 method for signature 'ubmsFit'
nsamples(object, ...)

Arguments

Value

An integer representing the number of posterior samples

Plot Residuals For All Submodels in a ubmsFit Model

Description

Plot Residuals For All Submodels in a ubmsFit Model

Usage

## S4 method for signature 'ubmsFit,ANY'
plot(x, y, ...)

Arguments

Value

A plot object of class gtable with one panel per submodel.

Plot Marginal Effects of Covariates

Description

Generates marginal fixed effects plots of one or more covariates from a ubmsFit submodel. For each plot, the focal covariate is allowed to vary across its range (or possible discrete values, for a factor), while the other covariates are held at their medians or reference levels. Random effects are ignored.

Usage

## S4 method for signature 'ubmsFit'
plot_effects(
  object,
  submodel,
  covariate = NULL,
  level = 0.95,
  draws = 1000,
  smooth = NULL,
  ...
)

## S4 method for signature 'ubmsFit'
plot_marginal(
  object,
  submodel,
  covariate = NULL,
  level = 0.95,
  draws = 1000,
  smooth = NULL,
  ...
)

Arguments

Value

A ggplot if a single covariate is plotted, or an object of class grob if there are multiple covariates/panels

Plot Posterior Distributions

Description

Plot posterior distributions for selected parameters. Posteriors can be represented as density plots or histograms.

Usage

## S4 method for signature 'ubmsFit'
plot_posteriors(object, pars = NULL, density = FALSE, ...)

Arguments

Value

A ggplot

Plot Model Residuals

Description

Plot residuals for a submodel from a ubmsFit object, for multiple posterior draws. By default, residuals are plotted against fitted values. When the submodel has a binomial response (e.g., detection models), regular residual plots are not typically informative. Instead, the residuals and fitted values are divided into bins based on fitted value and the averages are plotted. For a count response (e.g., Poisson), Pearson residuals are calculated. To plot residuals against values of a particular covariate instead of the fitted values, supply the name of the covariate (as a string) to the covariate argument.

Usage

## S4 method for signature 'ubmsFit'
plot_residuals(
  object,
  submodel,
  covariate = NULL,
  draws = 9,
  nbins = NULL,
  ...
)

Arguments

Value

A ggplot of residuals vs. fitted values or covariate values, with one panel per posterior draw. For binned residual plots, the shaded area represents plus/minus two standard deviations around the mean residual. If the model is true, we would expect about 95 fall within this area.

See Also

residuals

Plot A Map of the State Parameter Based on a Spatial ubms Model

Description

Plot A Map of the State Parameter Based on a Spatial ubms Model

Usage

plot_spatial(object, param = c("state", "eta"), sites = TRUE, cell_size = NULL)

Arguments

Posterior Distribution of the Linear Predictor

Description

Extract posterior draws of the linear predictor for a ubmsFit submodel, possibly transformed by the inverse-link function.

Usage

## S4 method for signature 'ubmsFit'
posterior_linpred(
  object,
  transform = FALSE,
  submodel,
  newdata = NULL,
  draws = NULL,
  re.form = NULL,
  ...
)

Arguments

Value

A matrix of simulations from the posterior predictive distribution of the linear predictor. The dimensions are draws by number of linear predictor values (e.g., number of sites or number of observations).

Draw from the posterior predictive distribution

Description

Draw from the posterior predictive distribution after fitting a model. You can draw from the posterior of the observed outcome y or the latent unobserved state z.

Usage

## S4 method for signature 'ubmsFit'
posterior_predict(
  object,
  param = c("y", "z"),
  draws = NULL,
  re.form = NULL,
  ...
)

Arguments

Value

A matrix of simulations from the posterior predictive distribution. If param = "z", the dimensions are draws by number of sites (or sites x primary periods in site-major order for dynamic models). If param = "y", the dimensions are draws by sites x observations (or sites x primary periods x observations for dynamic models).

Predict parameter values from a fitted model

Description

This method generates predicted parameter values for the original dataset or a new dataset using the posterior distribution. Standard deviation and a customizable uncertainty interval are also calculated.

Usage

## S4 method for signature 'ubmsFit'
predict(
  object,
  submodel,
  newdata = NULL,
  transform = TRUE,
  re.form = NULL,
  level = 0.95,
  ...
)

Arguments

Value

If newdata was a data frame: A data frame with one row per prediction and four columns: 1) Predicted point estimates (posterior means), 2) Standard deviation of the posterior, 3-4) Lower and upper bounds of the specified uncertainty interval

For parameters with more than one dimension, the rows are in site-major order, or site-year-observation for dynamic models.

If newdata was a SpatRaster/RasterStack, returns a SpatRaster/RasterStack with four layers corresponding to the four columns above with the same projection as the original SpatRaster/RasterStack.

See Also

posterior_linpred, posterior_interval

Prior distributions

Description

Specify prior distributions and associated parameters for use in ubms models.

Usage

normal(location = 0, scale = 2.5, autoscale = TRUE)

uniform(lower = -5, upper = 5)

student_t(df = 1, location = 0, scale = 2.5, autoscale = TRUE)

logistic(location = 0, scale = 1)

cauchy(location = 0, scale = 2.5, autoscale = TRUE)

gamma(shape = 1, rate = 1)

laplace(location = 0, scale = 2.5, autoscale = TRUE)

Arguments

Value

A list containing prior settings used internally by ubms.

Examples

normal()

Projected Occupancy Trajectories

Description

Generate posterior draws of occupancy probability for all sites and primary periods, i.e. the projected trajectory (Weir et al. 2009).

Usage

projected(object, ...)

## S4 method for signature 'ubmsFitColext'
projected(object, draws = NULL, re.form = NULL, ...)

Arguments

Value

A matrix of occupancy values from the posterior predictive distribution. The dimensions are draws by number of sites x primary periods in site-major order.

References

Weir LA, Fiske IJ, Royle J. 2009. Trends in Anuran Occupancy from Northeastern States of the North American Amphibian Monitoring Program. Herpetological Conservation and Biology. 4: 389-402.

See Also

stan_colext

Extract Random Effects

Description

Extract random effects from a ubmsFit model. Note that this function works like ranef for merMod objects from lme4, not like ranef for unmarkedFit objects. To get functionality similar to that of unmarkedFit, use posterior_predict.

Usage

## S4 method for signature 'ubmsFit'
ranef(object, submodel, summary = FALSE, add_mean = TRUE, ...)

Arguments

Details

Note: by default this function adds the overall intercept or slope to the (mean-0) random effect to get the complete random intercept or slope. In this way the output is more like the output of lme4::coef and not lme4::ranef. You can turn this off and return just the mean-0 random effect by setting argument add_mean = FALSE.

If you run ranef on a submodel with a spatial random effect, the function will return estimates of parameter eta.

Value

If summary=FALSE, a list of random effect values; if TRUE, a data frame with columns for random effect mean, SD, and 95

See Also

ranef, posterior_predict

Extract Model Residuals

Description

Extract residuals for a given submodel from a ubmsFit object. Residuals are calculated separately for each submodel using the posterior predictive distribution of the latent state z, following Wright et al. (2019).

Usage

## S4 method for signature 'ubmsFit'
residuals(object, submodel, draws = NULL, ...)

Arguments

Value

A matrix of residual values with dimension draws by observations. Note that calculation of residuals for the detection submodel is conditional on z > 0, so residuals for an observation in a posterior draw where z = 0 are assigned value NA (Wright et al. 2019).

References

Wright, W. J., Irvine, K. M., & Higgs, M. D. (2019). Identifying occupancy model inadequacies: can residuals separately assess detection and presence? Ecology, 100(6), e02703.

Fit the MacKenzie et al. (2003) Dynamic Occupancy Model

Description

This function fits the dynamic occupancy model of MacKenzie et al. (2003).

Usage

stan_colext(
  psiformula = ~1,
  gammaformula = ~1,
  epsilonformula = ~1,
  pformula = ~1,
  data,
  prior_intercept_psi = logistic(0, 1),
  prior_coef_psi = logistic(0, 1),
  prior_intercept_gamma = logistic(0, 1),
  prior_coef_gamma = logistic(0, 1),
  prior_intercept_eps = logistic(0, 1),
  prior_coef_eps = logistic(0, 1),
  prior_intercept_det = logistic(0, 1),
  prior_coef_det = logistic(0, 1),
  prior_sigma = gamma(1, 1),
  log_lik = TRUE,
  ...
)

Arguments

Value

ubmsFitColext object describing the model fit.

References

MacKenzie DI, Nicholas JD, Hines JE, Knutson MG, Franklin AB. 2003. Ecology 84: 2200-2207.

See Also

colext, unmarkedMultFrame

Examples

data(frogs)
umf <- formatMult(masspcru)
umf@y[umf@y > 1] <- 1 #convert counts to presence/absence
umf <- umf[1:100,] #Use only 100 sites

fit_frog <- stan_colext(~1, ~1, ~1, ~1, umf, chains=3, iter=300)

Fit the Royle et al. (2004) Distance Sampling Model

Description

This function fits the hierarchical distance sampling model of Royle et al. (2004) to line or point transect data recorded in discerete distance intervals.

Usage

stan_distsamp(
  formula,
  data,
  keyfun = c("halfnorm", "exp", "hazard"),
  output = c("density", "abund"),
  unitsOut = c("ha", "kmsq"),
  prior_intercept_state = normal(0, 5),
  prior_coef_state = normal(0, 2.5),
  prior_intercept_det = normal(0, 5),
  prior_coef_det = normal(0, 2.5),
  prior_intercept_scale = normal(0, 2.5),
  prior_sigma = gamma(1, 1),
  ...
)

Arguments

Value

ubmsFitDistsamp object describing the model fit.

Warning

Use of the hazard-rate key function ("hazard") typically requires a large sample size in order to get good parameter estimates. If you have a relatively small number of points/transects (<100), you should be cautious with the resulting models. Check your results against estimates from unmarked, which doesn't require as much data to get good estimates of the hazard-rate shape and scale parameters.

Note

Values of 'dist.breaks' in the 'unmarkedFrameDS' should be as small as possible (<10) to facilitate convergence. Consider converting 'unitsIn' from meters to kilometers, for example. See example below.

References

Royle, J. A., Dawson, D. K., & Bates, S. (2004). Modeling abundance effects in distance sampling. Ecology 85: 1591-1597.

See Also

distsamp, unmarkedFrameDS

Examples

data(issj)
#Note use of km instead of m for distance breaks
jayUMF <- unmarkedFrameDS(y=as.matrix(issj[,1:3]),
                          siteCovs=issj[,c("elevation","forest")],
                          dist.breaks=c(0,0.1,0.2,0.3),
                          unitsIn="km", survey="point")

fm_jay <- stan_distsamp(~1~scale(elevation), jayUMF, chains=3, iter=300)

Fit the Multinomial-Poisson Mixture Model

Description

This function fits the multinomial-Poisson mixture model, useful for data collected via survey methods such as removal or double observer sampling.

Usage

stan_multinomPois(
  formula,
  data,
  prior_intercept_state = normal(0, 5),
  prior_coef_state = normal(0, 2.5),
  prior_intercept_det = logistic(0, 1),
  prior_coef_det = logistic(0, 1),
  prior_sigma = gamma(1, 1),
  log_lik = TRUE,
  ...
)

Arguments

Value

ubmsFitMultinomPois object describing the model fit.

See Also

multinomPois, unmarkedFrameMPois

Examples

data(ovendata)
ovenFrame <- unmarkedFrameMPois(ovendata.list$data,
                                siteCovs=ovendata.list$covariates,
                                type="removal")

oven_fit <- stan_multinomPois(~1~scale(ufc), ovenFrame, chains=3, iter=300)

Fit the MacKenzie et al. (2002) Occupancy Model

Description

This function fits the single season occupancy model of MacKenzie et al. (2002).

Usage

stan_occu(
  formula,
  data,
  prior_intercept_state = logistic(0, 1),
  prior_coef_state = logistic(0, 1),
  prior_intercept_det = logistic(0, 1),
  prior_coef_det = logistic(0, 1),
  prior_sigma = gamma(1, 1),
  log_lik = TRUE,
  ...
)

Arguments

Value

ubmsFitOccu object describing the model fit.

References

MacKenzie DI, Nichols JD, Lachman GB, Droege S, Royle JA, Langtimm CA. 2002. Estimating site occupancy rates when detection probabilities are less than one. Ecology 83: 2248-2255.

See Also

occu, unmarkedFrameOccu

Examples

data(frogs)
pferUMF <- unmarkedFrameOccu(pfer.bin)

#Add some covariates
siteCovs(pferUMF) <- data.frame(cov1=rnorm(numSites(pferUMF)))

#Fit model
(fm <- stan_occu(~1~cov1, pferUMF, chains=3, iter=300))

Fit the Occupancy Model of Royle and Nichols (2003)

Description

Fit the occupancy model of Royle and Nichols (2003), which relates probability of detection of the species to the number of individuals available for detection at each site.

Usage

stan_occuRN(
  formula,
  data,
  K = 20,
  prior_intercept_state = normal(0, 5),
  prior_coef_state = normal(0, 2.5),
  prior_intercept_det = logistic(0, 1),
  prior_coef_det = logistic(0, 1),
  prior_sigma = gamma(1, 1),
  log_lik = TRUE,
  ...
)

Arguments

Value

ubmsFitOccuRN object describing the model fit.

References

Royle JA, Nichols JD. 2003. Estimating abundance from repeated presence-absence data or point counts. Ecology 84: 777-790.

See Also

occuRN, unmarkedFrameOccu

Examples

data(birds)
woodthrushUMF <- unmarkedFrameOccu(woodthrush.bin)
#Add a site covariate
siteCovs(woodthrushUMF) <- data.frame(cov1=rnorm(numSites(woodthrushUMF)))

(fm_wood <- stan_occuRN(~1~cov1, woodthrushUMF, chains=3, iter=300))

Fit Time-to-detection Occupancy Models

Description

Fit time-to-detection occupancy models of Garrard et al. (2008, 2013). Time-to-detection can be modeled with either an exponential or Weibull distribution.

Usage

stan_occuTTD(
  psiformula = ~1,
  gammaformula = ~1,
  epsilonformula = ~1,
  detformula = ~1,
  data,
  ttdDist = c("exp", "weibull"),
  linkPsi = c("logit"),
  prior_intercept_state = logistic(0, 1),
  prior_coef_state = logistic(0, 1),
  prior_intercept_det = normal(0, 5),
  prior_coef_det = normal(0, 2.5),
  prior_intercept_shape = normal(0, 2.5),
  prior_sigma = gamma(1, 1),
  log_lik = TRUE,
  ...
)

Arguments

Value

ubmsFitOccuTTD object describing the model fit.

References

Garrard, G.E., Bekessy, S.A., McCarthy, M.A. and Wintle, B.A. 2008. When have we looked hard enough? A novel method for setting minimum survey effort protocols for flora surveys. Austral Ecology 33: 986-998.

Garrard, G.E., McCarthy, M.A., Williams, N.S., Bekessy, S.A. and Wintle, B.A. 2013. A general model of detectability using species traits. Methods in Ecology and Evolution 4: 45-52.

Kery, Marc, and J. Andrew Royle. 2016. Applied Hierarchical Modeling in Ecology, Volume 1. Academic Press.

See Also

occuTTD, unmarkedFrameOccuTTD

Examples

#Simulate data
N <- 500; J <- 1
scovs <- data.frame(elev=c(scale(runif(N, 0,100))),
                    forest=runif(N,0,1),
                    wind=runif(N,0,1))
beta_psi <- c(-0.69, 0.71, -0.5)
psi <- plogis(cbind(1, scovs$elev, scovs$forest) %*% beta_psi)
z <- rbinom(N, 1, psi)

Tmax <- 10 #Same survey length for all observations
beta_lam <- c(-2, -0.2, 0.7)
rate <- exp(cbind(1, scovs$elev, scovs$wind) %*% beta_lam)
ttd <- rexp(N, rate)
ttd[z==0] <- Tmax #Censor at unoccupied sites
ttd[ttd>Tmax] <- Tmax #Censor when ttd was greater than survey length

#Build unmarkedFrame
umf <- unmarkedFrameOccuTTD(y=ttd, surveyLength=Tmax, siteCovs=scovs)

#Fit model
(fit <- stan_occuTTD(psiformula=~elev+forest, detformula=~elev+wind,
                     data=umf, chains=3, iter=300))

Fit the N-mixture model of Royle (2004)

Description

This function fits the single season N-mixture model of Royle et al. (2004).

Usage

stan_pcount(
  formula,
  data,
  K = NULL,
  mixture = "P",
  prior_intercept_state = normal(0, 5),
  prior_coef_state = normal(0, 2.5),
  prior_intercept_det = logistic(0, 1),
  prior_coef_det = logistic(0, 1),
  prior_sigma = gamma(1, 1),
  log_lik = TRUE,
  ...
)

Arguments

Value

ubmsFitPcount object describing the model fit.

References

Royle JA. 2004. N-mixture models for estimating populaiton size from spatially replicated counts. Biometrics 60: 105-108.

See Also

pcount, unmarkedFramePCount

Examples

data(mallard)
mallardUMF <- unmarkedFramePCount(mallard.y, siteCovs=mallard.site)

(fm_mallard <- stan_pcount(~1~elev+forest, mallardUMF, K=30,
                           chains=3, iter=300))

Extract Summary Statistics from a ubmsFit Model

Description

Extract Summary Statistics from a ubmsFit Model

Usage

## S4 method for signature 'ubmsFit'
summary(object, submodel, ...)

Arguments

Value

An object of class data.frame containing summary statistics for posterior distributions of parameters from the chosen submodel.

Markov Chain Traceplots

Description

Draws traceplots for chains from a ubmsFit object

Usage

## S4 method for signature 'ubmsFit'
traceplot(object, ...)

Arguments

Value

A ggplot object.

Turnover Probability

Description

Generate posterior draws of turnover probability from dynamic occupancy models. Turnover is calculated for each site and each primary period after the first.

Usage

turnover(object, ...)

## S4 method for signature 'ubmsFitColext'
turnover(object, draws, re.form = NULL, ...)

Arguments

Value

A matrix of turnover values from the posterior predictive distribution. The dimensions are draws by number of sites x (primary periods - 1) in site-major order.

See Also

stan_colext

Extractors for ubmsFitList objects Extract parts of ubmsFitList objects.

Description

Extractors for ubmsFitList objects Extract parts of ubmsFitList objects.

Usage

## S4 method for signature 'ubmsFitList'
x$name

## S4 method for signature 'ubmsFitList,numeric,missing'
x[[i]]

## S4 method for signature 'ubmsFitList,numeric,missing,missing'
x[i]

Arguments

Value

A ubmsFit object or list of such objects.

Widely Applicable Information Criterion (WAIC)

Description

Widely Applicable Information Criterion (WAIC)

Usage

## S4 method for signature 'ubmsFit'
waic(x, ...)

Arguments

Value

An object of class waic containing an estimate of WAIC and other parameters useful for model comparison. See ?loo::waic for more information.