[Corpora-List] POS-tagger maintenance and improvement

Francis Tyers ftyers at prompsit.com
Wed Feb 25 13:04:38 CET 2009


El mié, 25-02-2009 a las 12:57 +0100, lars nygaard escribió:
> Adam,
>
> I'm happily using taggers from GrammarSoft
> (http://www.grammarsoft.com)/The VISL project (http://visl.sdu.dk). They
> provide high-precision, wide coverage Constraint Grammar taggers for
> Danish, Swedish, Norwegian, English, Spanish, French, Esperanto and
> Portuguese. Some of the systems are, naturally, more mature than others,
> but they are continually updated, and my suggestions for systematic
> improvements have been promptly implemented.
>
> The taggers are available for commercial use for a fee, and, I believe,
> free of charge for limited research use.

In so far as commercial efforts go, I also recommend the GrammarSoft stuff. You can try out the Danish one online here:

http://beta.visl.sdu.dk/visl/da/parsing/automatic/upload.php

If you are interested in using Constraint Grammar for your tagging needs, it is also worth noting that Norwegian[1] and Sámi[2] Constraint Grammars are available free of charge under the GPL from the Universities of Tromsø (for Sámi) and Olso (for Norwegian). Tromsø also has a development versions of some other taggers under the GPL.

Fran

1. http://omilia.uio.no:8050/cl/cgp/test.html 2. http://giellatekno.uit.no/cgi/d-sme.eng.html



More information about the Corpora mailing list