Paper: Semi-automatically Developing Chinese HPSG Grammar from the Penn Chinese Treebank for Deep Parsing

ACL ID C10-2162
Title Semi-automatically Developing Chinese HPSG Grammar from the Penn Chinese Treebank for Deep Parsing
Venue International Conference on Computational Linguistics
Session Poster Session
Year 2010
Authors

In this paper, we introduce our recent work on Chinese HPSG gramar development through trebank conversion. By manualy defining gramatical constraints and ano- tation rules, we convert the bracketing tres in the Pen Chinese Trebank (CTB) to be an HPSG trebank. Then, a large-scale lexi- con is automaticaly extracted from the HPSG trebank. Experimental results on the CTB 6.0 show that a HPSG lexicon was sucesfuly extracted with 97.24% acu- racy; furthermore, the obtained lexicon achieved 98.51% lexical coverage and 76.51% sentential coverage for unsen text, which are comparable to the state-of-the-art works for English.