[Corpora-List] spotting names of drugs

John F Sowa sowa
Tue Jan 8 17:42:26 CET 2013


On 1/8/2013 11:07 AM, Martin Krallinger wrote:
> There are several approaches to used for the identification of drug
> names and chemical compounds, some based on dictionaries, other based
> on rules or machine learning methods.

Also important is parsing. A perfect parse with all the semantic features identified is not necessary. But even an imperfect parse with as much semantics and coreference resolution as practical is an enormous help.

John Sowa



More information about the Corpora mailing list