Stop Error in stratified areas-adjacency analyses

[Ziv Lieberman]

Hi all,

I'm trying to run a time-stratified analysis using areas-adjacency matrices only for each time bin. I am running into "stop error #2" whenever I try to run DEC, with LnL=NaN.

The error prompts that the adjacency matrices may be overly stringent. However, I lack the prior information to confidently use another approach (e.g., manual dispersal multipliers or distance methods). I have also tried the recommendation of not using the adjacency file but instead manually specifying nonadjacent areas for each time bin.

Is there a way to get around this issue, or do I really have to scrap the areas-adjacency approach? As I say, I'm not sure I can justify an approach that uses more specific kinds of matrices.

Please see attached my data files as well as the error message.

Thank you for any insight you can provide!

-Ziv

P.S. for what it's worth (probably nothing)!, I am able to run DEC under a strictly Bayesian framework using these same data and adjacencies (in revbayes).

[Nick Matzke]

Hi! -- adjacencies in RevBayes are probably used a bit differently?  I.e. "adjacency" can mean "I won't allow dispersal unless the areas are adjacent" or "I won't allow a range-state to exist in the model unless the areas are adjacent".  These are very different things.

The safest approach is to set up a manual states list, giving the exact list of ranges you want to allow in each time period.

(wikidot seems stressed, here is the archived link to that discussion)

https://web.archive.org/web/20231206053029/http://phylo.wikidot.com/example-biogeobears-scripts#manual_modify_states

Then use the manual_dispersal_multipliers to control which areas are allowed to have dispersal between them.

I will take a look at your files - that error is almost always someone making their data impossible, ie they have a tip range of ABC but their model disallows any transitions to get to ABC.

Cheers,

Nick

--
You received this message because you are subscribed to the Google Groups "BioGeoBEARS" group.
To unsubscribe from this group and stop receiving emails from it, send an email to biogeobears...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/biogeobears/01ed9fa7-e1e8-4ec5-9c64-269b64d90898n%40googlegroups.com.

[Ziv Lieberman]

Hi Nick,

Thanks for the rapid reply! I did also try the method at the link you provided - you will see it commented out in the .Rmd file - and got the same error. It is, of course, possible that I made a mistake in setting it up!

Cheers,

Ziv

[Nick Matzke]

Hi -- it looks to me like your area adjacency matrix excludes range AI, which is a range of the tip Ravavy_miafina_D0646.

When I added this, it worked...files attached... (note: this gives my manual ranges interpretation of your areas adjacency matrix).

Cheers,

Nick

To view this discussion visit https://groups.google.com/d/msgid/biogeobears/204ee6df-813e-4995-84eb-866d04522641n%40googlegroups.com.

[Ziv Lieberman]

Ah-ha! That would do it ;) Thank you for the very careful check of the inputs and for the explanation. I'll go ahead and make sure this runs for me too.

Thank you again! This was extremely helpful.

Ziv