glmmADMB

Dear Thierry,



Many thanks for your help.

The animals do have an unique id number, so your expectations were correct
and unfortunately the same error message popped up.

With regards to your options, I am wondering if it is correct to insert
village as fixed effect or completely omit it, it feels quite
counterintuitive.

I have no experience with the other two options, but will try to dig into
it. If I encounter any additional questions, I will let you know.



Thank you again,

Best regards,

Veronique Dermauw.





*From:* Thierry Onkelinx [mailto:thierry.onkelinx at inbo.be]
*Sent:* dinsdag 17 november 2015 11:18

*To:* Veronique Dermauw
*Cc:* r-sig-mixed-models at r-project.org
*Subject:* Re: [R-sig-ME] glmmADMB



Dear Veronique,



This is helpful. It looks like the small number of villages is the
problem. If the animals stay within one village you could try if explicitly
nesting them helps (1|village/animal) instead of (1|village) + (1|animal).
It's a wild guess and I don't expect it to work. A crossed random effect
like (1|village) + (1|animal) with unique id's for each animal is an
implicit nested random effect. Hence (1|village/animal) should be the same
model as (1|village) + (1|animal). But if it does work, please let us know.



Other options:

- try to supply starting values

- add village to the fixed effects

- simply removing village from the model

- look for a Bayesian solution (e.g. INLA)



Best regards,


ir. Thierry Onkelinx
Instituut voor natuur- en bosonderzoek / Research Institute for Nature and
Forest
team Biometrie & Kwaliteitszorg / team Biometrics & Quality Assurance
Kliniekstraat 25
1070 Anderlecht
Belgium

To call in the statistician after the experiment is done may be no more
than asking him to perform a post-mortem examination: he may be able to say
what the experiment died of. ~ Sir Ronald Aylmer Fisher
The plural of anecdote is not data. ~ Roger Brinner
The combination of some data and an aching desire for an answer does not
ensure that a reasonable answer can be extracted from a given body of data.
~ John Tukey



2015-11-17 10:51 GMT+01:00 Veronique Dermauw <vdermauw at itg.be>:

Dear Thierry,



I don?t know if the following would already be helpful?

193 animals over 3 treatments and 6 villages (2 villages/treatment;
variable amount of animals/village: 29-35), with observations on 5
timepoints.

There is no error message when I remove ?(1|Animal)+(1|Village)? or only ?
+(1|Village)?.

So I am assuming there is something wrong with the random effect of
village, correct?
Based on my data, I assume that I need a model with the animal nested in
village (and now I am doubting this +(1|Animal)+(1|Village) is the right
way to attain this (even though I had found this approach for a nested term
online)).



Best regards,

Veronique.





*From:* Thierry Onkelinx [mailto:thierry.onkelinx at inbo.be]
*Sent:* dinsdag 17 november 2015 10:33


*To:* Veronique Dermauw
*Cc:* r-sig-mixed-models at r-project.org
*Subject:* Re: [R-sig-ME] glmmADMB



Dear Veronique,



We need more information. How many animals, villages, treatments,
timepoints, ...? How many observations per animal and per village? Can you
post the data or send it offline?



Best regards,


ir. Thierry Onkelinx
Instituut voor natuur- en bosonderzoek / Research Institute for Nature and
Forest
team Biometrie & Kwaliteitszorg / team Biometrics & Quality Assurance
Kliniekstraat 25
1070 Anderlecht
Belgium

To call in the statistician after the experiment is done may be no more
than asking him to perform a post-mortem examination: he may be able to say
what the experiment died of. ~ Sir Ronald Aylmer Fisher
The plural of anecdote is not data. ~ Roger Brinner
The combination of some data and an aching desire for an answer does not
ensure that a reasonable answer can be extracted from a given body of data.
~ John Tukey



2015-11-17 10:09 GMT+01:00 Veronique Dermauw <vdermauw at itg.be>:

Dear Thierry,



Thank you for your response and advice.

I?ve omitted missing values (94 out of 965), but I am getting another
error message, see below:

cambNA<-na.omit(camb)

mod1b<-glmmadmb(GIS~Treatment*Time+(1|Animal)+(1|Village),family="nbinom",data=cambNA)

Parameters were estimated, but standard errors were not: the most likely
problem is that the curvature at MLE was zero or negative

Error in glmmadmb(GIS ~ Treatment * Time + (1 | Animal) +  :

  The function maximizer failed (couldn't find parameter file)
Troubleshooting steps include (1) run with 'save.dir' set and inspect
output files; (2) change run parameters: see '?admbControl';(3) re-run with
debug=TRUE for more information on failure mode

In addition: Warning message:

running command 'C:\WINDOWS\system32\cmd.exe /c glmmadmb -maxfn 500 -maxph
5 -noinit -shess' had status 1



Thank you for your help,

Best regards,

Veronique Dermauw.



*From:* Thierry Onkelinx [mailto:thierry.onkelinx at inbo.be]
*Sent:* donderdag 12 november 2015 17:44
*To:* Veronique Dermauw
*Cc:* r-sig-mixed-models at r-project.org
*Subject:* Re: [R-sig-ME] glmmADMB



Dear Veronique,



It looks like your data contains missing values. Try to remove them. Your
syntax seems to be correct.



Note that Treatment + Time + Time * Treatment can be abbreviated to Time
* Treatment



Best regards,



Thierry


ir. Thierry Onkelinx
Instituut voor natuur- en bosonderzoek / Research Institute for Nature and
Forest
team Biometrie & Kwaliteitszorg / team Biometrics & Quality Assurance
Kliniekstraat 25
1070 Anderlecht
Belgium

To call in the statistician after the experiment is done may be no more
than asking him to perform a post-mortem examination: he may be able to say
what the experiment died of. ~ Sir Ronald Aylmer Fisher
The plural of anecdote is not data. ~ Roger Brinner
The combination of some data and an aching desire for an answer does not
ensure that a reasonable answer can be extracted from a given body of data.
~ John Tukey



2015-11-11 12:17 GMT+01:00 Veronique Dermauw <vdermauw at itg.be>:

Dear,

Recently I have been working on a dataset which seems to require a
generalized linear mixed model with a negative binomial distribution.
I came across the glmmADMB package which seems ideal.
However, I am a bit stuck here (see codes and error message below).
The dataset I am working on, involves count variables (GIS) from animals
from different villages who were allocated a treatment (Treatment) and were
sampled at various Time point (hence Time and Time*Treatment).
I was wondering if you could assist me with this command?

Many thanks for your response,
Best regards,
Veronique Dermauw.


mod1b<-glmmadmb(GIS~Treatment+Time+Time*Treatment+(1|Animal)+(1|Village),family="nbinom",data=camb)
summary(mod1b)
Error in II[, ii] = II[, ii] + REmat$codes[[i]] :
  number of items to replace is not a multiple of replacement length
In addition: Warning messages:
1: In glmmadmb(GIS ~ Treatment + Time + Time * Treatment + (1 | Animal) +
:
  NAs removed in constructing fixed-effect model frame: you should
probably remove them manually, e.g. with na.omit()
2: In II[, ii] + REmat$codes[[i]] :
  longer object length is not a multiple of shorter object length


Veronique Dermauw
PhD, MSc Veterinary Medicine
Institute of Tropical Medicine
Nationalestraat 155
2000 Antwerp
+32(0)32476447


P Please consider the environment before printing this e-mail

Disclaimer: Http://www.itg.be/disclaimer

Directions to our location(s): http://g.co/maps/ua89b

        [[alternative HTML version deleted]]