www.phidot.org

In MARK under the huggins parameterization the derived abundance estimates are partitioned based on the group structure specified when the data were imported. If the user specified four groups on import - males2001, males2002, females2001, females2002 - then mark will derive an estimate of abundance +/- SE for these groups no matter what model the user specifies (dot, time varying, etc.). But what if the user wants a combined estimate WITH a SE of males and females in 2001 and the same for 2002? Sure, delta method, but RMark offers a much more attractive option, for individuals with average brain cells, since the group structure can be quickly manipulated (re-process.data).

I've been trying to do this and have run into an error. In the next set of code, I ran the dot model with the same group structure (combined age and year, 14 groups) as my MARK file,

Code: Select all

# convert
LYC=convert.inp("C:/Users/abreton.CNRDOM/Documents/CSU_Project/MARK_Analyses/LYC_2004_2008/EH_File_LYC_2004_10_Bins.inp",
group.df=data.frame(age.yr=c("SA04","A04","SA05","A05","SA06","A06","SA07","A07","SA08","A08","SA09","A09","SA10","A10"),
aclass=rep(c("SA","A"),7),
syear=c(rep("2004",2),rep("2005",2),rep("2006",2),rep("2007",2),rep("2008",2),rep("2009",2),rep("2010",2))),
covariates=c("L1","L2","L3","L4","L5","L6","L7","L8","L9","L10","L11","may1Bin","RMiBin"),use.comments=FALSE)

# process
LYC.huggins=process.data(LYC,model="Huggins",begin.time=1,mixtures=1,
 time.intervals=c(1,1,1,1,1,1,1,1,1,1), groups=("age.yr"))

# make design data
LYC.ddl=make.design.data(LYC.huggins)

# specify parameters to fix
p.fix.indices=c(83:88,6:11,93:99,16:22,107:110,30:33,119:121,42:44,129:132,52:55,154,77)
c.fix.indices=c(229:234,159:164,238:244,168:174,251:254,181:184,262:264,192:194,271:274,201:204,275:277,205:207,285:286,
 294,215:216,224)
p.fix.count=rep(0,length(p.fix.indices))
c.fix.count=rep(0,length(c.fix.indices))

# specify pim
p.dot.fix.shared=list(formula=~1,share=TRUE,fixed=list(index=c(p.fix.indices,c.fix.indices),value=c(p.fix.count,c.fix.count)))

# fit dot model
initial=-3.0337564
LYC.huggins.M0=mark(LYC.huggins,LYC.ddl, model="dot",
 model.parameters=list(p=p.dot.fix.shared),output=FALSE, simplify=FALSE, initial=initial)

..and the results match MARK perfectly. Here are elements from the dm,

Code: Select all

> LYC.huggins.M0$design.matrix[1:14,]
 1  2  3  4  5  6  7  8  9 10 11 12 13 14
 1  1  1  1  1  1  1  1  1  1  1  1  1  1
Levels: 1
> LYC.huggins.M0$design.matrix[155:168,]
155 156 157 158 159 160 161 162 163 164 165 166 167 168
  1   1   1   1   1   1   1   1   1   1   1   1   1   1
Levels: 1

Next, I changed the group specified in process.data from age.yr to syear (sample year). This resulted in an error. Here is the process.data with the alternative group structure,

Code: Select all

LYC.huggins=process.data(LYC,model="Huggins",begin.time=1,mixtures=1,
 time.intervals=c(1,1,1,1,1,1,1,1,1,1), groups=("syear"))

The error,

Error in extract.mark.output(out, model, adjust, realvcv) :
MARK did not run properly. If error message was not shown, re-run MARK with invisible=FALSE
********Following model failed to run : p(~1)c() **********

...and elements of the design matrix,

Code: Select all

> LYC.huggins.M0$results$beta
NULL
> LYC.huggins.M0$design.matrix[1:14,]
 1  2  3  4  5  6  7  8  9 10 11 12 13 14
 1  1  1  1  1  1  1  1  1  1  1  1  1  1
Levels: 1
> LYC.huggins.M0$design.matrix[155:168,]
<NA> <NA> <NA> <NA> <NA> <NA> <NA> <NA> <NA> <NA> <NA> <NA> <NA> <NA>
<NA> <NA> <NA> <NA> <NA> <NA> <NA> <NA> <NA> <NA> <NA> <NA> <NA> <NA>

Thanks for having a look.

andre

Andre-

Did you use simplify=FALSE on that run? I didn't see the run command that failed. If so, please try with default of simplify=TRUE. There is no reason to not simplify and I've been meaning to remove that argument because I've not been checking to see if it works since I implemented the simplification process.
If that doesn't do it, please look in the mark.out file that the run create and send the error that it reports.

This is not likely a problem with process.data but in building/running model with simplify=F.

--jeff

Previously, I experienced an error when I specified fix parameters in combination with (the default) simplify=TRUE. When I instead set the mark function argument to simplify=FALSE I had no problems as long as I didn't change the group structure. Nonetheless, I tried your suggestion with the dot model. Here is the code when I DON'T FIX any parameters,

Code: Select all

# convert
LYC=convert.inp("C:/Users/abreton.CNRDOM/Documents/CSU_Project/MARK_Analyses/LYC_2004_2008/EH_File_LYC_2004_10_Bins.inp",
group.df=data.frame(age.yr=c("SA04","A04","SA05","A05","SA06","A06","SA07","A07","SA08","A08","SA09","A09","SA10","A10"),
aclass=rep(c("SA","A"),7),
syear=c(rep("2004",2),rep("2005",2),rep("2006",2),rep("2007",2),rep("2008",2),rep("2009",2),rep("2010",2))),
covariates=c("L1","L2","L3","L4","L5","L6","L7","L8","L9","L10","L11","may1Bin","RMiBin"),use.comments=FALSE)

# process
LYC.huggins=process.data(LYC,model="Huggins",begin.time=1,mixtures=1,
 time.intervals=c(1,1,1,1,1,1,1,1,1,1), groups=("age.yr"))

# make design data
LYC.ddl=make.design.data(LYC.huggins)

# specify pim
p.dot.shared=list(formula=~1,share=TRUE)

# fit dot model
initial=-3.0337564
LYC.huggins.M0=mark(LYC.huggins,LYC.ddl, model="dot",
 model.parameters=list(p=p.dot.shared),output=FALSE,initial=initial)

This code runs in RMark without any errors and the results match what I get when I specify the same model/run directly in MARK.

Here's what happens when I specify parameters to be FIXED in the PIM specifications. First the code,

Code: Select all

# specify parameters to fix
p.fix.indices=c(83:88,6:11,93:99,16:22,107:110,30:33,119:121,42:44,129:132,52:55,154,77)
c.fix.indices=c(229:234,159:164,238:244,168:174,251:254,181:184,262:264,192:194,271:274,201:204,275:277,205:207,285:286,
 294,215:216,224)
p.fix.count=rep(0,length(p.fix.indices))
c.fix.count=rep(0,length(c.fix.indices))

# specify pim
p.dot.fix.shared=list(formula=~1,share=TRUE,fixed=list(index=c(p.fix.indices,c.fix.indices),value=c(p.fix.count,c.fix.count)))

# fit dot model
initial=-3.0337564
LYC.huggins.M0=mark(LYC.huggins,LYC.ddl, model="dot",
 model.parameters=list(p= p.dot.fix.shared),output=FALSE,initial=initial)

Here's the error reported in the R console,

[code]Error in extract.mark.output(out, model, adjust, realvcv, vcvfile) :
MARK did not run properly. If error message was not shown, re-run MARK with invisible=FALSE
********Following model failed to run : p(~1)c() **********

And here's the error from the bottom of the MARK...out file,
ERROR -- Error with integer read of COVARIATES

A fix for this error may/may not solve the problem I was having when I started this post. Regardless, I though it made sense to start here...let me know if you need any more details.

*** Update: Jeff resolved this problem on 10-10-11. ***