Checking convergence

[Francesco Cicconardi]

Hi,

I run cafe5 in two steps, a first one to estimate both lambda and alpha on the dataset with most complete species and with the estimated values on the full dataset. How do I check if the final run converge?

Thanks for your help.

Best

F

____________________________________

Dr Francesco Cicconardi

School of Biological Sciences
University of Bristol
Email: F.Cicc...@bristol.ac.uk

Twitter: @FCicconardi

Google Scholar

Research Group: EBAB Lab

[Hahn, Matthew]

Convergence can only really be checked on the same dataset (and parameter settings), by seeing if the final likelihoods are the same. I don’t know of any other way.

Matt

--
You received this message because you are subscribed to the Google Groups "hahnlab-cafe" group.
To unsubscribe from this group and stop receiving emails from it, send an email to hahnlabcafe...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/hahnlabcafe/CAEe4u8UCu%2B5s2nYpZ6Sygt-CdFMOYqzRh%3DHRK7%3DBT%2BynzuoMFw%40mail.gmail.com.

[Francesco Cicconardi]

Hi Matt,

Right! Should I expect the exact same values?

Thanks

F

[Hahn, Matthew]

Pretty close, yeah. 

On Apr 20, 2023, at 5:58 AM, Francesco Cicconardi <fran...@gmail.com> wrote: