The NEGRA corpus consists of approximately 10,000 sentences of German newspaper text. The corpus is a type of treebank, but with a novel annotation scheme for discontinuous constituents. An example tree showing the visual format, the annotation format, and the Treebank equivalent, is available here. Annotate is a sophisticated tool which supports human-machine collaboration on the construction of syntactic trees.

http://www.coli.uni-saarland.de/projects/sfb378/negra-corpus/negra-corpus.html

Facebooktwittergoogle_pluslinkedin