DIVALIKE+J runs very slow compared to other models

[Daniel Eusebio Quintero Vallejo]

Hi Nick! I have run DEC, DIVALIKE (and their +J versions) and BAYAREALIKE for a phylogeny of 206 species. I have 16 areas and a max area num of 3 (so I have less than 700 states). Those analysys lasted 1 day (2 days for the +J versions), but in the case of BAYAREALIKE+J it has run for 5 days and has not finished yet. I don´t know why this is happening. I followed the instructions at http://phylo.wikidot.com/biogeobears (and I even raised the "min" nd "max" values for each parameter in case of a possible crash, because my computer is windows). Do you have an idea of what is happening?

Thnak you so much! 

Daniel. 

[Nick Matzke]

Hi -- that is peculiar, BAYAREALIKE+J is usually very similar to the other +J models.  It's somewhat possible that the BAYAREALIKE model gets very different ML parameter estimates, and when these are used as the starting points for BAYAREALIKE+J, it takes longer to get into the right zone of parameter space.  

You might consider running with the optimiser optimx instead of GenSA  -- the latter is slower & more thorough, but probably overkill for the 2/3 parameter models. 

My gut feeling is your BAYAREALIKE+J has got to be just about done.  But if you decide to kill it, just survey the screen printout and write down the parameter values producing the best lnLs, you can use these as starting points in a new search...

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 on the web visit https://groups.google.com/d/msgid/biogeobears/2da074f0-d92c-497f-95bf-5e9848e9f7dbn%40googlegroups.com.

[Daniel Eusebio Quintero Vallejo]

Thank you! If I decide to to kill it and start again with the parameters that produce the best InLs, do i have to modify this kind of line? I mean, instead of "dstart" should I use the value in the screen printout?

BioGeoBEARS_run_object$BioGeoBEARS_model_object@params_table["d","init"] = dstart

To view this discussion on the web visit https://groups.google.com/d/msgid/biogeobears/CAJdu7BA3AU7mqrw%3DQJ6QszRb57ek4dcRbx64g9xkRvFSBQ8mWA%40mail.gmail.com.

[Nick Matzke]

Hi -- yes, exactly...

Cheers!

Nick

To view this discussion on the web visit https://groups.google.com/d/msgid/biogeobears/CAPDeQA%2BH8AwTQNQ10FW%2BCHsRNM%3Dxsj4ma%3Dve6cN1dUw0o4s%2BhQ%40mail.gmail.com.