openWordnet-PT
openWordnet-PT copied to clipboard
reuse propbank.br
The list of verbs is available at http://www.nilc.icmc.usp.br/portlex/
We have already done this. check the prototype experiments in http://wnpt.brlcloud.com/wn/prototypes/corpora#verbnet (Number of words in corpus: 5692. In OWN-PT: 1563. In suggestions: 1301. Missing: 2828.)
here you can see the missing verbnet verbs. @fcbr has to press his OWN-PT button again, as there are are many things with votes that have not been committed, it seems to me.
also I suggest removing the -se of the verbs in the list (2828), following our decision that every verb that takes a "se" should appear in the lexicon without it too.
are there other kinds of information (e.g subcategorization information) that we can extract from VerbNet.BR?
@vcvpaiva yes we have more info to extract from verbnet.br.
@arademaker did you check with @claudiafreitas if she's ok with the the theory in Duran, M. S.; Scarton, C. E.; Aluísio, S. M.; Ramisch, C. (2013) Identifying Pronominal Verbs: Towards Automatic Disambiguation of the Clitic “se” in Portuguese. NAACL 2013 - 9th Workshop on Multiword Expressions MWE 2013? there's a list in the page you mentioned before.
I completely forgot the prototype! @arademaker didn't we try to put some intern to make that list dynamic or something? I think we should try to maybe retake that work. Anyway, I'll push the button.