Closed FredericBlum closed 3 years ago
Comparison is very simple. Are they already in the lexibank dataset? Then we can add some code to a scripts/ folder that shows how to do that with B-Cubed scores (theory described in our paper in detail here: https://doi.org/10.1371/journal.pone.0170046, LingPy's functions are straightforward to calculate them).
Right now I have both CU's original cognacy judgements (many of them dubious) as well as automated cognacy assignment. With the help of Pache's thesis I should be able to identify the final COGID's