We describe in this paper an approach for synchronous parsing of syntactic and semantic dependency structures that combines recent advances in the area to get a very high accuracy as well at the same time very good parsing times. The time for parsing, the time for training and the values of the memory footprint are to our knowledge the best results reported while the parsing accuracy are as high as the highest results reported in the 2008 shared task. The corpora used in the shared task are still different to the dependency structures of the Meaning-Text Theory. Therefore, we outline the adaption of the approach to the dependency structures of the Meaning-Text Theory
© 2001-2024 Fundación Dialnet · Todos los derechos reservados