El dj 05 de 04 de 2012 a les 09:02 +0200, en/na Per Tunedal va escriure:
> Hi Keld,
> I haven't decided yet on the corpus to work on. My initial idea was to
> do some web crawling after texts in the two languages involved. I
> believe there is a tool available. Maybe at Uppsala University? Anyone
> who has any idea on how to proceed?
> 
> I thought there was a pair SV - EN to work on, thus I intended to
> quickly set up a minor working system (maybe limited to a restraint
> domain). Now, when there isn't any such pair, nor any DA - EN pair, I
> will have to think it over once more. My main interest is French, a
> language completely different from Swedish. If I would work on the pair
> SV - FR instead, I will have to consider using an other tool than
> Apertium. I will have a look at Matxin and MOSES (statistical).
> 
> Is it possible to use dictionaries and other resources from Apertium
> with Matxin?

Yes, but I wouldn't use Matxin unless you're working with Basque, the
generation componenent makes some Basque-specific assumptions and afaik
there is still no way to add nodes.

If you want to work with RBMT on a "deep" level, your other options are:

* Apertium (lexical processing) + Constraint Grammar (for
reordering/insertion)
* Some combination of the LOGON system
* MOLTO (but this assumes working with a restricted domain)

I think you could knock up an sv-fr or an sv-en example system in a few
weeks. There are plenty of resources for both languages.

What do you intend to use the system for ? 

Regards,

Fran


------------------------------------------------------------------------------
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

Reply via email to