Creates an Analysis object

This method creates an Analysis objects which describes a one or more models to fit to the distance data. The simulation will fit each of these models to the data generated in the simulation and select the model with the minimum criteria value.

Usage

make.ds.analysis(
  dfmodel = list(~1),
  key = "hn",
  truncation = numeric(0),
  cutpoints = numeric(0),
  er.var = "R2",
  control.opts = list(),
  group.strata = data.frame(),
  criteria = "AIC"
)

Arguments

dfmodel

list of distance sampling model formula specifying the detection function (see ?Distance::ds for further details)

key

key function to use; "hn" gives half-normal (default) and "hr" gives hazard-rate.

truncation

absolute truncation distance in simulation units matching the region units.

cutpoints

supply a vector of cutpoints if you wish the simulation to perform binned analyses.

er.var

encounter rate variance estimator to use when abundance estimates are required. Defaults to "R2" for line transects and "P3" for point transects. See mrds::varn for more information / options.

control.opts

A list of control options: method - optimisation method,

group.strata

Dataframe with two columns ("design.id" and "analysis.id"). The former gives the strata names as defined in the design (i.e. the region object) the second specifies how they should be grouped (into less strata) for the analyses. See details for more information.

criteria

character model selection criteria (AIC, AICc, BIC)

Value

DS.Analysis-class object

Details

It is possible to group strata at the analysis stage using the group.strata argument. For example, for design purposes it may have been sensible to divide strata into substrata. This can help make more convex shapes and therefore zigzag designs more efficient or perhaps it helped to keep transects angled parallel to density gradients across the study area. Despite these (purely design relevant) substrata we may still wish to calculate estimates of density / abundance etc. for each stratum. The table below gives an example of the data.frame which can be used to do this. Imagine a study region with an onshore strata and an offshore strata. The onshore strata has been divided in two at the design stage to keep transects perpendicular to the coast. We now want to analyse this as just two strata the onshore and offshore.

design.id	analysis.id
———	———–
onshoreN	onshore
onshoreS	onshore
offshore	offshore

See also

ds make.simulation

Author

Laura Marshall

Examples

# Model selection considering both a half-normal and a hazard-rate model
# using AIC criteria and truncating 5% of the data
ds.analyses <- make.ds.analysis(dfmodel = ~1,
                                key = c("hn", "hr"),
                                truncation = 500,
                                criteria = "AIC")

# Model selection considering both a half-normal with no covariates and with size
# as a covariate using AIC criteria and truncating at 500
ds.analyses <- make.ds.analysis(dfmodel = list(~1, ~size),
                                key = "hn",
                                truncation = 500,
                                criteria = "AIC")

# Model selection considering both a half-normal with no covariates and with size
# as a covariate and a hazard rate, using AIC criteria and truncating at 500
ds.analyses <- make.ds.analysis(dfmodel = list(~1, ~size, ~1),
                                key = c("hn", "hn", "hr"),
                                truncation = 500,
                                criteria = "AIC")