El ds 07 de 04 de 2012 a les 17:55 +0200, en/na Per Tunedal va escriure: > Bonjour Bernard, > > It seams very tedious to create the bilingual dictionary from scratch. > If there was on NO - ENG or DA - ENG I could easily build on that one. > > I realize that for a SV - FR I would have to start from scratch. > > Besides, there isn't even an editor for the dictionaries. You have to > code it "by hand". I WYSIWYG editor to just enter the information would > be handy. > > The statistical approach (MOSES) is tempting, as it offers an > opportunity to automatically build a translation system. But it demands > a VERY large corpus, and I doubt it would be manageable for me. > > My intention is to build a working system for a very limited domain, > using a small set of representative texts.
For the bilingual dictionary, you can take Moses + Europarl and generate/postedit a probabilistic bilingual dictionary from that. There are some instructions on the Wiki.[1] Deleting / editing entries is much faster work than creating them from scratch. I can do about 500 entries/day. Which gives you around 10 days of work to get a basic bilingual dictionary. There are probably also other sources for sv-fr. Regards, Fran 1. http://wiki.apertium.org/wiki/Extracting_bilingual_dictionaries_with_Giza%2B%2B ------------------------------------------------------------------------------ For Developers, A Lot Can Happen In A Second. Boundary is the first to Know...and Tell You. Monitor Your Applications in Ultra-Fine Resolution. Try it FREE! http://p.sf.net/sfu/Boundary-d2dvs2 _______________________________________________ Apertium-stuff mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/apertium-stuff
