A robust Dependency Parser (including a morphological analyzer and a Part-of-Speech
Tagger) has been tested on a corpus of about 40.000 words. It analyzes
about 120 word per second.
The tests on the manually corrected treebank revealed an error rate (in terms of wrong attachments and wrong arc labels) of 16.7%. The parser also includes a preliminary treatment of traces, adopted for preserving the projectivity of the dependency trees.
Downloading the parser by following the link.