RMark (version 2.1.1)

model.table: Create table of MARK model selection results

Description

Constructs a table of model selection results for MARK analyses. The table includes the formulas, model name, number of parameters, deviance, AICc, DeltaAICc, model weight and residual deviance. If chat>1 QAICc, QDeltaAICc and QDeviance are used instead.

Usage

model.table(model.list = NULL, type = NULL, sort = TRUE, adjust = TRUE,
  ignore = TRUE, pf = 1, use.lnl = FALSE, use.AIC = FALSE,
  model.name = TRUE)

Arguments

model.list

a vector of model names or a list created by the function collect.models which has each model object and at the end a model.table ; If nothing is specified then any mark object in the

type

type of model (eg "CJS")

sort

if true sorts models by criterion

adjust

if TRUE adjusts # of parameters to # of cols in design matrix

ignore

if TRUE collects all models and ignores that they are from different models

pf

parent frame value; default=1 so it looks in calling frame of model.table; used in other functions with pf=2 when functions are nested two-deep

use.lnl

display -2lnl instead of deviance

use.AIC

use AIC instead of AICc

model.name

if TRUE uses the model.name in each mark object which uses formula notation. If FALSE it uses the R names for the model obtained from collect.model.names or names assigned to marklist elements

Value

• result.table - dataframe containing summary of models
• model.name
  name of fitted model
• parameter.name - an entry for each parameter
  formula for parameter
• npar
  number of estimated parameters
• AICc or QAICc
  AICc value or QAICc if chat>1
• DeltaAICc or DeltaQAICc
  difference between AICc or QAICc value from model with smallest value
• weight
  model weight based on exp(-.5*DeltaAICc) or exp(-.5*QDeltaAICc)
• Deviance or QDeviance
  residual deviance from saturated model
• chat
  overdispersion constant if not 1

Details

This function is used by collect.models to construct a table of model selection results with the models that it collects; however it can be called directly to construct the table.

See Also

collect.model.names, collect.models

Examples

# This example is excluded from testing to reduce package check time
data(dipper)
run.dipper=function()
{
#
# Process data
#
dipper.processed=process.data(dipper,groups=("sex"))
#
# Create default design data
#
dipper.ddl=make.design.data(dipper.processed)
#
# Add Flood covariates for Phi and p that have different values
#
dipper.ddl$Phi$Flood=0
dipper.ddl$Phi$Flood[dipper.ddl$Phi$time==2 | dipper.ddl$Phi$time==3]=1
dipper.ddl$p$Flood=0
dipper.ddl$p$Flood[dipper.ddl$p$time==3]=1
#
#  Define range of models for Phi
#
Phi.dot=list(formula=~1)
Phi.time=list(formula=~time)
Phi.sex=list(formula=~sex)
Phi.sextime=list(formula=~sex+time)
Phi.sex.time=list(formula=~sex*time)
Phi.Flood=list(formula=~Flood)
#
#  Define range of models for p
#
p.dot=list(formula=~1)
p.time=list(formula=~time)
p.sex=list(formula=~sex)
p.sextime=list(formula=~sex+time)
p.sex.time=list(formula=~sex*time)
p.Flood=list(formula=~Flood)
#
# Return model table and list of models
#
cml=create.model.list("CJS")
return(mark.wrapper(cml,data=dipper.processed,ddl=dipper.ddl))
}

dipper.results=run.dipper()
dipper.results
dipper.results$model.table=model.table(dipper.results,model.name=FALSE)
dipper.results
#
# Compute matrices of model weights, number of parameters and Delta AICc values
#
model.weight.matrix=tapply(dipper.results$model.table$weight,
 list(dipper.results$model.table$Phi,dipper.results$model.table$p),mean)
model.npar.matrix=tapply(dipper.results$model.table$npar,
 list(dipper.results$model.table$Phi,dipper.results$model.table$p),mean)
model.DeltaAICc.matrix=tapply(dipper.results$model.table$DeltaAICc,
 list(dipper.results$model.table$p,dipper.results$model.table$Phi),mean)
#
# Output DeltaAICc as a tab-delimited text file that can be read into Excel
# (to do that directly use RODBC or xlsreadwrite package for R)
#
write.table(model.DeltaAICc.matrix,"DipperDeltaAICc.txt",sep="\t")