Some models not appearing after running model selection function

[Anna V]

Hello,

I am a graduate student studying veiled chameleon ecology in southwest Florida and I am attempting to use ctmm’s to calculate home range sizes. I apologize if this question has been answered before in previous conversations, but I am quite new to ctmms and this has been a large obstacle in allowing me to continue my project forward.

I have noticed for some (most) of the lizards in my study (n = 10, tracked for 100 days as frequently as twice a day and as infrequently as once every four days, but usually once every other day), a few of the models do not appear when running the “summary(fitted.mods.[lizardID])” function. In the majority of cases, all of the anisotropic models would show up, but some of the other models, like OUf, OU or IID, would not. Then in a few lizards, some of the anisotropic models would not appear as well, only the OUF anisotropic, OU anisotropic, OUF, and OUf anisotropic models. I attached screenshots of what this looks like when run and a table of which models appear and which do not for each lizard. It seems like it might be related to differences in the manual fit of the variogram; at first I wasn’t saving my fitted GUESS objects for each individual, and models seemed to appear and disappear each time I refit.

Does anyone have any insight on why this may be and how to make it more consistent from animal to animal?

Thank you,

Anna Velten

[Andrew Durso]

We've continued to try troubleshooting this but so far have not found the solution. Anybody out there have any ideas?

[Jake Straub]

Disclaimer:  I'm no expert in CTMM but have used it and specifically the model selection approach.   I refer you to the following found on page 20 of the ctmm movement package documentation.   I bolded the section that I believe is most pertinent to you. 

"Model selection in ctmm.select proceeds in two phases. If there are a large number of parameters that must be fit numerically (such as when error is modeled), then the target model (argument CTMM) is worked toward by first fitting simpler, compatible models. The second phase proceeds by attempting to simplify the autocorrelation model and complexify the deterministic (trend) model until the information criterion fails to improve. The intent of working in these directions is to improve numerical convergence and avoid fitting trends to autocorrelation. Note that simpler models in a nested hierarchy will only be attempted if they appear credible, which can be adjusted with the level argument. level=1 will, therefore, always attempt a simpler model.

--
You received this message because you are subscribed to the Google Groups "ctmm R user group" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ctmm-user+...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/ctmm-user/c9a4089e-44dd-4226-8004-26b7aa1aa29dn%40googlegroups.com.