Probabilistic Constituent Parser

Probabilistic (Constituent) Parser

action name	statParse
resources required	(Bikel-format) grammar & properties files
properties	StatParser.properties.fileName StatParser.grammar.fileName
annotations required	constit (for lexical items)
annotations added	constit

The statistical parser is an English parser trained on the Penn TreeBank. This annotator provides a Jet interface to the parser written by Dan Bikel at Penn, which is in turn based on the parser by Michael Collins.

statParse adds an annotation of the form <constit cat=category children=[child₁ child₂ ...]> for each non-terminal constituent in the parse tree. Here category is the non-terminal grammar category and child₁ child₂ ... are the annotations of the immediate constituent nodes.