RMark (version 2.1.1)

rerun.mark: Runs a previous MARK model with new starting values

Description

Runs a previous MARK model with new starting values but without specifying the model parameter formulas. This function is most useful with mark.wrapper in which a list of models is analyzed and the set of formulas are not specified for each model.

Usage

rerun.mark(model, data, ddl, initial, output = TRUE,
    title = "", invisible = TRUE, adjust = TRUE,
    se = FALSE, filename = NULL, prefix = "mark",
    default.fixed = TRUE, silent = FALSE, retry = 0,
    realvcv = FALSE, external = FALSE)

Arguments

model

previously run MARK model

data

processed dataframe used with model

ddl

design data list used with model

initial

vector of initial values for beta parameters or previously run model object of similar structure

output

If TRUE produces summary of model input and model output

title

Optional title for the MARK analysis output

invisible

if TRUE, exectution of MARK.EXE is hidden from view

adjust

if TRUE, adjusts number of parameters (npar) to number of columns in design matrix, modifies AIC and records both

se

if TRUE, se and confidence intervals are shown in summary sent to screen

filename

base filename for files created by MARK.EXE. Files are named filename.*.

prefix

base filename prefix for files created by MARK.EXE; the files are named prefixnnn.*

default.fixed

if TRUE, real parameters for which the design data have been deleted are fixed to default values

silent

if TRUE, errors that are encountered are suppressed

retry

number of reanalyses to perform with new starting values when one or more parameters are singular

realvcv

if TRUE the vcv matrix of the real parameters is extracted and stored in the model results

external

if TRUE the mark object is saved externally rather than in the workspace; the filename is kept in its place

Value

• model: MARK model object with the base filename stored in output and the extracted results from the output file appended onto list; see mark for a detailed description of a mark object.

Details

This is a simple function that restarts an analysis with MARK typically using another model for initial values of the beta parameters. The processed dataframe (data) and design data list (ddl) must be specified but the model.parameters are extracted from model. initial values are not optional otherwise this would be no different than the original call to mark. More complete definitions of the arguments can be found in mark or run.mark.model or make.mark.model.

See Also

make.mark.model, run.models, extract.mark.output, adjust.parameter.count, mark, cleanup

Examples

#
# The following example will not run because the data are not included in the
# examples.  It illustrates the use of rerun.mark with mark.wrapper.  With this
# particular data set the POPAN models were having difficulty converging.  After
# running the set of models using mark.wrapper and looking at the results it
# was clear that in several instances the model did not converge.  This is easiest
# to discern by comparing nested models in the model.table.  If one model
# is nested within another,then the deviance of the model with more
# parameters should be as good or better than the smaller model.  If that
# is not the case then the model that converged can be used for initial
# values in a call to rerun.mark for the model that did not converge.
#

#do.nat=function()
#{
#Phi.ageclass=list(formula=~ageclass)
#Phi.dot=list(formula=~1)
#p.area=list(formula=~area)
#p.timebin.plus.area=list(formula=~timebin+area)
#p.timebin.x.area=list(formula=~-1+timebin:area)
#pent.ageclass=list(formula=~ageclass)
#pent.ageclass.plus.EN=list(formula=~ageclass+EN)
#pent.ageclass.plus.diffEN=list(formula=~ageclass+EN92+EN97+EN02)
#cml=create.model.list("POPAN")
#nat=mark.wrapper(cml,data=zc.proc,ddl=zc.ddl,
#  invisible=FALSE,initial=1,retry=2)
#return(nat)
#}
#nat=do.nat()
# model list
#            Phi                   p                      pent
#1  Phi.ageclass              p.area             pent.ageclass
#2  Phi.ageclass              p.area pent.ageclass.plus.diffEN
#3  Phi.ageclass              p.area     pent.ageclass.plus.EN
#4  Phi.ageclass p.timebin.plus.area             pent.ageclass
#5  Phi.ageclass p.timebin.plus.area pent.ageclass.plus.diffEN
#6  Phi.ageclass p.timebin.plus.area     pent.ageclass.plus.EN
#7  Phi.ageclass    p.timebin.x.area             pent.ageclass
#8  Phi.ageclass    p.timebin.x.area pent.ageclass.plus.diffEN
#9  Phi.ageclass    p.timebin.x.area     pent.ageclass.plus.EN
#10      Phi.dot              p.area             pent.ageclass
#11      Phi.dot              p.area pent.ageclass.plus.diffEN
#12      Phi.dot              p.area     pent.ageclass.plus.EN
#13      Phi.dot p.timebin.plus.area             pent.ageclass
#14      Phi.dot p.timebin.plus.area pent.ageclass.plus.diffEN
#15      Phi.dot p.timebin.plus.area     pent.ageclass.plus.EN
#16      Phi.dot    p.timebin.x.area             pent.ageclass
#17      Phi.dot    p.timebin.x.area pent.ageclass.plus.diffEN
#18      Phi.dot    p.timebin.x.area     pent.ageclass.plus.EN
#
# use model 9 as starting values for model 7
#nat[[7]]= rerun.mark(nat[[7]],data=zc.proc,ddl=zc.ddl,initial=nat[[9]])
# use model 3 as starting values for model 1
#nat[[1]]= rerun.mark(nat[[1]],data=zc.proc,ddl=zc.ddl,initial=nat[[3]])
# use model 14 as starting values for model 15
#nat[[15]]= rerun.mark(nat[[15]],data=zc.proc,ddl=zc.ddl,initial=nat[[14]])
# use model 5 as starting values for model 6
#nat[[6]]= rerun.mark(nat[[6]],data=zc.proc,ddl=zc.ddl,initial=nat[[5]])
# use model 10 as starting values for model 11
#nat[[11]]= rerun.mark(nat[[11]],data=zc.proc,ddl=zc.ddl,initial=nat[[10]])
# use model 10 as starting values for model 12
#nat[[12]]= rerun.mark(nat[[12]],data=zc.proc,ddl=zc.ddl,initial=nat[[10]])
# reconstruct model table with new results
#nat$model.table=model.table(nat[1:18])
# show new model table
#nat