Hi Joe, I'd say this issue is related to these particular data: there are four states so the ARD model has 12 parameters which is quite a lot considering that one state ('3') is overwhelming represented in the data. If you simulate data with two states, then the ARD model has only 2 parameters and the fit converges most of the times.
Best, Emmanuel ----- Le 29 Jan 21, à 5:49, Joseph Keating joe.keat...@bristol.ac.uk a écrit : > Dear R-sig-phylo mailing group > > I'm having some problems using the 'all rates different' (ARD) model of the > 'ace' function. The function seems to return different errors / warnings > depending on the specified method for computing the exponential of the rate > matrix. Most notably, its returning likelihoods for states that are >1 or <0. > I > don't get the same errors if using the 'SYM' or "ER" models. > > Please see my post on stackoverflow for an example. Link below: > https://stackoverflow.com/questions/65944177/nonsensical-likelihoods-returned-when-using-the-ace-function-and-the-all-rate > > If anyone could provide any help or advice, it would be hugely appreciated. > And > thanks to Ben Bolker for pointing me to this mailing group. > > Kind Regards > > Joe Keating > [https://cdn.sstatic.net/Sites/stackoverflow/Img/apple-touch-i...@2.png?v=73d79a89bded]<https://stackoverflow.com/questions/65944177/nonsensical-likelihoods-returned-when-using-the-ace-function-and-the-all-rate> > phylogeny - Nonsensical likelihoods returned when using the 'ace' function and > the 'all rates different' model in the R package: 'Ape' - Stack > Overflow<https://stackoverflow.com/questions/65944177/nonsensical-likelihoods-returned-when-using-the-ace-function-and-the-all-rate> > I'm having some problems using the 'all rates different' (ARD) model of the > 'ace' function. The function seems to return different errors / warnings > depending on the specified method for computing the exponential of the rate > matrix. > stackoverflow.com > > > > Dr Joseph Keating > Palaeobiology Research Group > University of Bristol > Bristol Life Sciences Building > 24 Tyndall Avenue > Bristol > BS8 1TQ > > [[alternative HTML version deleted]] > > _______________________________________________ > R-sig-phylo mailing list - R-sig-phylo@r-project.org > https://stat.ethz.ch/mailman/listinfo/r-sig-phylo > Searchable archive at http://www.mail-archive.com/r-sig-phylo@r-project.org/ _______________________________________________ R-sig-phylo mailing list - R-sig-phylo@r-project.org https://stat.ethz.ch/mailman/listinfo/r-sig-phylo Searchable archive at http://www.mail-archive.com/r-sig-phylo@r-project.org/