[R-sig-eco] Possible bug with dredge function

[Dunbar, Michael J.]

Fernando

You have transect in both the fixed and random part of the model. Will the model fit when transect is not in the fixed part? Will it fit when you just call lmer with the full model without dredge? Can you explain a bit more about you data and in particular how the transect variable is coded?

regards
Mike


________________________________________
From: r-sig-ecology-bounces at r-project.org [r-sig-ecology-bounces at r-project.org] On Behalf Of Pedro Lima Pequeno [pacolipe at gmail.com]
Sent: 13 December 2011 19:01
To: Fernando Schmidt
Cc: r-sig-ecology at r-project.org
Subject: Re: [R-sig-eco] Possible bug with dredge function

Hi, Fernando.

The warning message "In mer_finalize(ans) : gr cannot be computed at
initial par (65)" points to problems during the iterative process of
model fitting. This may be caused by the complexity of the global
model you're implying. Have you tried using dredge() with a simpler
global model (e.g. without interactions)?

2011/12/14, Fernando Schmidt <schmidt.fa at gmail.com>:
> Dear colleagues,
>
> I am working with information-theoretic based on the second-order AKaike’s
> Information criterion corrected for small samples sizes AICc.  I have run
> the mixed-effect models with “glmer” function in the lme4 package. Then, I
> used the dredge function from the MuMIn package to test models defined by
> all possible variables combinations and rank them by AICc-based model
> weight.
>
> Usually, the main results from dredge function are AICc, delta and weight.
> However, for one model seems to have some “bug” I just could get AICc.
> Delta and weight were missing:
>
>>
> mfullepi1<-glmer(sepi~lhet+ldth+transect+lhet:transect+(1|park/transect),family="poisson")
>
>> dredge(mfullepi1)
>
> Global model: glmer(formula = sepi ~ lhet + ldth + transect + lhet:transect
> +
>
>     (1 | park/transect), family = "poisson")
>
> ---
>
> Model selection table
>
>    (Int)  ldt      lht     trn lht:trn k      Dev.            AICc
>
> 5  1.1630                  +                   14     97.79        129.8
>
> 7  1.0000           0.5228 +            15     95. 62       130.2
>
> 6  0.9803 0.002208         +          15    96.32        130.9
>
> 8  0.8236 0.002160  0.5188 +    16   94.17       131.5
>
> 3  1.4120           0.5538                    4   124.30      132.7
>
> 4  1.2810 0.002287  0.5153           5   122.40     132.9
>
> 2  1.3920 0.002544                          4   125.00     133.3
>
> 1  1.5480                                             3   127.20     133.4
>
> 15 1.4660          -1.0140 +   +       26
>
> 16 1.2560 0.002198 -0.9244 +   +       27
>
> Random terms: 1 | park/transect
>
> Mensagens de aviso perdidas:
>
> 1: In mer_finalize(ans) : gr cannot be computed at initial par (65)
>
> 2: In mer_finalize(ans) : gr cannot be computed at initial par (65)
>
>
>
> Please, someone could explain me what is going on?
>
>
>
> Thanks a lot!
>
> All the best,
>
> Fernando
>
>
> --
> Fernando Augusto Schmidt
> Lab. de Ecologia de Comunidades
> PPG - Entomologia. Universidade Federal de Viçosa
> Viçosa, MG - Brazil 36570-000
> www.labecol.ufv.br
> +55 3182388810
>
>       [[alternative HTML version deleted]]
>
>


--
Pedro A. C. Lima Pequeno
Programa de Pós-graduação em Ecologia
Instituto Nacional de Pesquisas da Amazônia
Manaus, AM, Brasil

_______________________________________________
R-sig-ecology mailing list
R-sig-ecology at r-project.org
https://stat.ethz.ch/mailman/listinfo/r-sig-ecology
-- 
This message (and any attachments) is for the recipient only. NERC
is subject to the Freedom of Information Act 2000 and the contents
of this email and any reply you make may be disclosed by NERC unless
it is exempt from release under the Act. Any material supplied to
NERC may be stored in an electronic records management system.