Carlos Rodriguez wrote:
> Hi,
>
> I was wondering if anyone could point me to domain corpora with the
> following characteristics:
>
> 1.- Written texts (ASCII, xml, txt,pdf, no need to be tagged) from
> specialized or technical domains.
If 1 million tokens is ok, you can try the MuchMore corpus of medical
texts (German/English):
http://muchmore.dfki.de/resources1.htm
Cheers,
Paul Buitelaar
DFKI - Language Technology Lab
Saarbruecken, Germany
> 2.- Open source, or reasonably priced, that can be downloaded to be
> processed (web-accesible through proprietary interfaces won't cut it).
> 3.- If possible, with machine-readable or electronic lexicons or
> dictionaries available for the domain represented by the corpora.
>
> I am thinking about experimenting with techniques for lexical
> acquisition.
>
> Thanks and best to all,
>
>
> Carlos Rodríguez
>
>
This archive was generated by hypermail 2b29 : Wed Mar 30 2005 - 12:48:05 MET DST