Maria Claudia de Freitas,  Milena Uzeda-Garrão,  Claudia Oliveira,  Cícero Nogueira dos Santos,  Maria Cândida Silveira. 

This paper describes the construction of the main linguistic resource associated with the task of training an automatic NP extractor: an annotated corpus. The annotation is specifically designed to be used in the training phase of a rule learning algorithm: Transformation Based error-driven Learning (TBL). The NP model used in the annotation has been developed to overcome performance related problems encountered when the target language changed from English to Portuguese.

http://www.lbd.dcc.ufmg.br/colecoes/til/2005/0017.pdf

Processando consulta.