1

I am trying to use MuMIn's model.avg function:

 m06gamma.avg<-model.avg(get.models(dredge(glmer(biomassa_consumo ~ Renda.R.nor + dist_casa_urba_centro_kmnor +
media_esfor_caça.anonor+ veg_prim_2.5km.nor+ + (1|nested),
nAGQ=0,family=Gamma(link="log"), data=killb)), cumsum(weight)<=.95))

But, the "relative variable importance" isn´t working: 

(conditional average) 
                            Estimate Std. Error Adjusted SE z value Pr(>|z|)    
(Intercept)                  5.57789    0.14531     0.14960  37.286   <2e-16 ***
dist_casa_urba_centro_kmnor  0.28667    0.25240     0.25726   1.114    0.265    
veg_prim_2.5km.nor          -0.31277    0.23201     0.23645   1.323    0.186    
Renda.R.nor                  0.12301    0.14458     0.14885   0.826    0.409    
media_dist_caça.kmnor       -0.04138    0.17932     0.18388   0.225    0.822    
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

>

How can I make model.avg to retourn the "relative variable importance" values?

I'm looking for the result like this example:

result example

Thanks in advance.

Ben Bolker
  • 211,554
  • 25
  • 370
  • 453
Milton de Paula
  • 125
  • 4
  • 1
    (1) it's almost always better to use the `data=` argument rather than `with()` when fitting models. Don't know if that is affecting anything here. (2) Could we please have a [mcve]? (3) For people who don't use this package, could you explain what you expect to see in the results that you're not seeing? – Ben Bolker May 06 '20 at 00:11
  • 1
    did you look at `?MuMIn::importance` ... ? – Ben Bolker May 06 '20 at 00:47
  • Dear Ben Bolker, thank you for the suggestions. I'm not sure that I'm able to do a reproductive example, I need more time to know how about that. I will follow your suggestion "?MuMIn::importance", and then if the my question was not solved, I will return with a reproductive example. – Milton de Paula May 06 '20 at 03:17
  • 1
    Dear Ben Bolker, after look the description of per-variable sum of model weights, I was able to fix my question using this code: sw(x). Thanks a lot! – Milton de Paula May 06 '20 at 03:33
  • If you have solved your own question, you should either post an answer *or* (if you think the answer will be unhelpful to future searchers) delete your question ... – Ben Bolker May 06 '20 at 14:10

0 Answers0