search

lexibank / abvdoceanic

Creative Commons Attribution 4.0 International
5 stars 2 forks source link

Check to which degree synonymy and mutual coverage influence the phylogenetic analysis #27

Closed LinguList closed 2 years ago

LinguList commented 2 years ago

We have detected from #26 already that we have some cases of excess synonymy. It seems reasonable to repeat the tossing-coins analysis to see how much this influences the phylogenies (albeit that this check only looks at NJ trees).

LinguList commented 2 years ago

A similar analysis can be done with mutual coverage then.

SimonGreenhill commented 2 years ago

Suspect NJ will struggle with these data...

LinguList commented 2 years ago

Agree, but there's no systematic study investigating how much this would affect Bayesian analyses, or is there? But as we discussed elsewhere, this is probably better done in another context with a dedicated and small dataset for testing, so one can run sufficient analyses.