bbgdm: bbgdm objects
In skiptoniam/bbgdm: Bayesian Bootstrap Generalised Dissimilarity Models.

Description Usage Arguments Value References Examples

creates a bbgdm model, comprising multiple gdms. bbgdm models are core of bbgdm, different parameterisation can be achieve similar to a glm, see bbgdm.fit for more details.

bbgdm(form, sp.dat, env.dat, family = "binomial", link = "logit",
  dism_metric = "number_non_shared", nboot = 100, spline_type = "ispline",
  spline_df = 2, spline_knots = 1, geo = FALSE, geo.type = "euclidean",
  coord.names = c("X", "Y"), optim.meth = "nlmnib", est.var = FALSE,
  trace = FALSE, prior = FALSE, control = bbgdm.control())

## S3 method for class 'bbgdm'
print(x, ...)

## S3 method for class 'bbgdm'
plot(x, ...)

## S3 method for class 'bbgdm'
predict(x, data, neighbourhood = NULL, outer = FALSE,
  uncertainty = TRUE, ...)

`form`	formula for bbgdm model
`sp.dat`	presence absence matrix, sp as columns sites as rows.
`env.dat`	environmental or spatial covariates at each site.
`family`	a description of the error distribution and link function to be used in the model. Currently "binomial" suppported. This can be a character string naming a family function, a family function or the result of a call to a family function.
`link`	a character string that assigns the link function to apply within the binomial model. Default is 'logit', but 'negexp' and other binomial link functions can be called.
`dism_metric`	dissimilarity metric to calculate for model. "bray_curtis" or "number_non_shared" currently avaliable.
`nboot`	number of Bayesian Bootstraps to run, this is used to estimate variance around GDM models. Default is 100 iterations.
`spline_type`	type of spline to use in GDM model. Default is monotonic isplines. Options are: "ispline" or "bspline".
`spline_df`	Number of spline degrees of freedom.
`spline_knots`	Number of spline knots.
`geo`	logical If true geographic distance is calculated if
`geo.type`	type of geographic distance to estimate, can call 'euclidean' and 'greater_circle'.
`coord.names`	character.vector names of coordinates, default is c("X","Y")
`optim.meth`	optimisation method options avaliable are 'optim' and 'nlmnib'
`est.var`	logical if true estimated parameter variance using optimiser.
`trace`	logical print extra optimisation outputs
`prior`	numeric vector of starting values for intercept and splines
`control`	control options for gdm calls bbgdm.control as default.
`x`	an object of class `bbgdm`, constructed by the function `bbgdm` giving a fitted bbgdm
`...`	for `bbgdm()`: one or more `plot()`, `print()` and `predict()`: further arguments passed to or from other methods
`data`	raster stack of the same covariates used to fit model object for the region you wish to predict too.
`neighbourhood`	int default is three, number of neighbouring cells to estimate mean dissimilarity.
`outer`	logical default is FALSE, if TRUE only calculates the outer edge of neighbourhoods area.
`uncertainty`	logical if TRUE predict will return a list with two rasters the mean estimate and the uncertainty (defined as the coefficent of variation)

a bbgdm model object

raster of mean turnover estimated based on neighbourhood distance.

Woolley, S. N., Foster, S. D., O'Hara, T. D., Wintle, B. A., & Dunstan, P. K. (2017). Characterising uncertainty in generalised dissimilarity models. Methods in Ecology and Evolution.

## Not run: 
sp.dat <- matrix(rbinom(2000,1,.6),200,10)# presence absence matrix
env.dat <- simulate_covariates(sp.dat,2)
form <- ~ 1 + covar_1 + covar_2
test.bbgdm <- bbgdm(form,sp.dat, env.dat,family="binomial",dism_metric="number_non_shared",
                    nboot=10, geo=FALSE,optim.meth='nlmnib',,control=bbgdm.control(cores=3))

## End(Not run)
#print model summary
## Not run: 
print(test.bbgdm)
## End(Not run)

#plot bbgdm fit
 ## Not run: 
plot(test.bbgdm)
## End(Not run)