Treep

David Chiang dchiang at cis.upenn.edu
Dan Bikel dbikel at cis.upenn.edu

Treep stands for Tree Processor. It takes as input a sequence of phrase-structure trees and modifies their labels according to a set of rules. Its envisioned purpose is as a front-end to the trainer of a statistical parser. Its rule notation is flexible enough to emulate the head/argument-finding rules of many recent parsers, including Collins' 1997/1999 model. See David Chiang and Daniel M. Bikel, "Recovering latent information in treebanks," Proceedings of COLING '02, 2002.

Download Treep (including example rule sets): 80k gzipped tar archive